Core Analysis By Low-Field NMR'

Christian Straley and Dan Rossini Scli I t i n berger- Doll Researck Ceriter Harold Vinegar and Pierre Tutunjian Shell Development Co. Chris Morriss Sckliimberger Wireline arid Testing

Abstract: NMR core measurements have been made customarily in high magnetic fields and have emphasized T I for pore-size evaluation or have used chemical shift spectroscopy for quantitative differentiation of oil and water. However, new NMR logging tools use T? measurements made at low field strengths. and these measurements should be supported by core analysis. This paper illustrates that low-field NMR T, 9 nieasurenients made on water-saturated rock cores can be used to nieasure total porosity and extract T2 distributions. The T2 distributions are shown to be closely related to pore-size distributions. Our experiments demonstrate that low-field nieasurements are quantitative, even for cases where high-field measurements fail. T2 distributions from water-saturated plugs are used to estimate producible fluids for sandstones and carbonates and to estimate clay bound water and matrix permeability. Although chemical shift spectroscopy is not feasible at low fields, a method is demonstrated on native state diatomites that permits the measurement of oil, water, and gas saturation. In the case presented, it is also possible to determine in-situ oil viscosIty.

Low-field NMR measures three properties: the equilibrium nuclear magnetization, M,,, and the two principal NMR relaxation times, T I (longitudinal relaxation time) and T2(transverse relaxation time). M, is proportional to the fluid-filled porosity, and T I and T2 have been correlated with petrophysical properties such as pore size, producible fluid, and permeability. Although T I measurements are more common in the literature, they are more time consuming to measure than T, measurements. Hence, the new pulsed NMR logging tools preferentially measure T2 for faster logging speeds and increased wellsite efficiency. Traditionally, quantitative analysis by NMR has used high-field rather than low-field measurements. The use of high field, H , increases the available signal as indicated by its relationship to the nuclear magnetization,
Mn=XriH,
(1)

INTRODUCTION

Hydrogen ('H) nuclear magnetic resonance (NMR) is a rapid, nondestructive measurement that can provide a wealth of infonnation about the fluids in oil- and watersaturated rock. This paper illustrates many applications of low-tield NMR measurements to core analysis. Low-field NMR has been found to be more accurate in measuring water and oil volumes than high-field NMR for rocks that have large internal magnetic field gradients. Moreover, low-field measurements are of great interest since the new pulsed NMR logging tools (Miller et al., 1990; Kleinberg et al., 1992; and Morriss et a].. 1993) are able to generate fields only in the range from 120 to 550 gauss.

where X,, is the static nuclear susceptibility caused by the hydrogen. Higher magnetic fields result in a stronger signal; hence, greater signal-to-noise ratio and faster measurements. At higher field, larger chemical shifts permit differentiation of oil and water signals (Vinegar et al., 1991). For fluids in porous media, however, the increased field that improves the signal also causes higher internal gradients, G, that degrade the measurements (Brown and Fantazzinie, 1993). When a rock is in an applied field, the gradients arise from magnetic field inhomogeneity, AH, caused by the susceptibility contrast between rock grains and pore fluids,

M--AX,H,

(2)

Presented previously at the 1994 International Symposium, Society of Core Analysts, SCA9404. Stavanger. Norway.
84

where Axe is the difference in susceptibilities (resulting primarily from electronic contributions). In rocks, the internal gradients resulting from field inhomogeneity are a complicated function of microgeometry but are on the order of AH divided by the grain size. G can frequently exceed
March-April 1997

The Log Analyst

Core Analysis by Low-Field NMR

100 gauss/cm at 20 kilogauss (85 MHz 'H Larmor frequency). The same rocks would have gradients of only 2.5 gauss/cm at 500 gauss (2.1 MHz 'H Larmor frequency). To understand why large internal magnetic field gradients can cause complications it is necessary to consider the three T2 relaxation mechanisms in rocks. These are due to molecular motion in fluids, surface relaxivity at the pore wall, and molecular diffusion in the internal gradients. The first mechanism, caused by local motions such as molecular tumbling, is equally effective whether or not the fluid is in a rock. This is called bulk liquid relaxation. The bulk liquid T2 decreases as fluid viscosity, r ] , increases (Bloembergen et al., 1948):

T
GB---

11

(3)

where T is the absolute temperature. The second mechanism is the relaxation of 'H nuclei when they closely approach paramagnetic ions such as iron or manganese, which reside on grain surfaces. Fluid molecules diffuse to the surface, have a probability of being relaxed there, and then diffuse away. For water in an individual pore the relaxation time constant is (Cohen and Mendelson. 1982)

T,,

=1(9)?
P2

diffusional relaxation is considered a nuisance. In rocks with potentially high internal gradients (e.g., very shaly sands) or measurements made at high field, the extrapolation to find A ( 0 ) can be inaccurate and the relaxation curves for the bulk fluid or surface relaxation can be completely obscured. Fortunately, the relaxation from the gradients can be minimized by shortening z as seen in Equation (6). However, because of spectrometer hardware constraints at high field it is not always possible to achieve the required interecho spacing, and for many rocks it is better to work at low magnetic field strengths (where G is also smaller). For the purposes of this paper, low field means a CPMG sequence with Hz = 0.1 gauss-s. Consider the extreme case of water (D = 2 X cm'/s) in an internal gradient of 10 gausskm (from a 500 Gauss applied field). If Hz = 0.1 gauss-s is maintained, gradient diffusional relaxation will be negligible for T2 values as high as that of bulk water. When gradient diffusional relaxation is negligible, TI and T2contain the same petrophysical information, and the ratio of T I to T2has been shown to be approximately 1.6 for a collection of sandstones and carbonates (Kleinberg et al., 1993). NMR transverse relaxation data can be expressed as a sum of exponentials
A(t)=

(4)

c
1

A(0);e-"T~' .

(7)

Equation (7) can be inverted into a T2 relaxation time distribution, where the T2; are a preselected basis set of The third mechanism is a result of the diffusion of molerelaxation time constants and the A(O), are the signal amplicules in magnetic field gradients. Using the Carr-PurcellMeiboom-Gill (CPMG) pulse sequence, the relaxation time tudes associated with each time constant by fitting (Gallegos and Smith, 1988; Fordham et al., 1994; and Freedman, resulting from diffusion (2'2D) is 1994). 3 Single-component fluids are characterized by a single (5) T2D = y2G2Dr2 ' For example, bulk water has a T2 of 3 s at room value of T2. where y is the 'H gyromagnetic ratio, D is the diffusion temperature. Multicomponent fluids such as crude oils coefficient, and z is the Carr-Purcell pulse spacing. The often display a distribution of relaxation times (Morris et CPMG sequence produces a closely spaced train of spin al., 1994). In water-saturated rocks, the bulk water relaxation is echoes whose decay is analyzed to determine T2.The interecho spacing, 22, can be varied and is an important often negligible, and the distribution of T2 arises from the experimental parameter (Kleinberg et al., 1993 and distribution of surface-to-volume ratio of the pores, as Kenyon, 1992). shown in Equation (4) (Kenyon, 1992). Because T2depends The time dependence of the signal, A(t), from all three linearly upon pore size, the T2distribution corresponds to a processes and measured at the peaks of the echoes (Klein- pore-size distribution with the largest pores having the berg et al., 1993 and Kleinberg and Horsfield, 1990), is longest relaxation times. When oil and water coexist in the pore space, the general principle is that the relaxation of fluid in contact with grain surfaces is enhanced by those surfaces. See Equation (4). A(O), the signal at time = 0, is proportional to Mn and Conversely, fluid that does not wet the surface relaxes at its therefore fluid-filled porosity. For quantitative analysis bulk rate, as relaxation does not occur at the oil-water A(O), T2B and T2smeasurements are of interest; the gradient interface. Because oil and water have similar 'H density, the
p2 is the transverse surface relaxivity.
March-April 1997

where (S/V) is the surface-to-volume ratio of the pore and

The Log Analyst

85

Straley, et al.

2c
15

a
.r(

=!

10

! i 0
a

$
I
1

0
0
5 10 15

20

10

20

30

Buoyancy porosity, p.u.

Comparison of NMR porosity measured at three field strengths to buoyancy porosity. o i s 85-MHz data; + is 2-MHz data: * is 1-MHz data.
Figure 1:

Buoyancy porosity, p.u.

Figure 3: comparison of 2-MHz NMR porosity with buoyancy porosity for 192 sandstones.

T2

distribution is a linear summation of the individual contributions from the oil and water.
RESULTS AND DISCUSSION

(3

8
! 3
2

-0.2

B
0 .

Porosity

+
e;:

; -0.4
-0.6

iE

-0.8
0 .

Evidence of internal gradients and their dependence on magnetic field can be seen in porosity measurements on a suite of shaly sands and shales from the Shell Johnson City test well. Figure 1 compares porosity measurements made at 8 5 , 2, and 1 MHz with Carr-Purcell pulse spacing (z) of 0.25, 0.08, and 0.25 msec, respectively. These values o f t were the shortest that could be achieved with the instruments. The unknown porosity, $Imhown, is determined by comparing the signal. A( O)u,,klown from Equation (6), to the values for a standard sample:
I

-11
0

II
4
The 85-MHz data were collected on a GE CSI-2T spectrometer, the 2-MHz data on a custom-built Schlumberger spectrometer, and the I-MHz data on the NUMAR CoreSpec 1000.1 clearly, the porosity from the 85-MHz

CEC,rneq/lOOg or Wt% Fe+fi

Figure 2: T h e relative error in 85-MHz N M R porosity crossplotted with CEC and paramagnetic impurities. The open circles indicate CEC values and the solid circles are weight percent Fe plus Mn.

'Mark of NUMAR Corp.

86 The Log Analyst

March-April 1997

Core Analysis by Low-Field N M R

70

60 +
+

4 +

*+
-

50 -

3
+

40

Buoyancy porosity, p.u.

Figure 4:

Comparison of 2-MHz NMR porosity with buoyancy porosity for 18 high-porosity diatomites.

Figure 5: Comparison of 2-MHz NMR porosity with buoyancy porosity for 71 carbonates.

significantly underestimated the gravimetric porosity for all samples, whereas the porosities from the 1-MHz measurement are in good agreement even though the same value of z was used. As expected, because of the low field and very short z values, the 2-MHz porosities are also in good agreement with gravimetric values. The fractional error in the 85-MHz porosity is strongly correlated with the weight of paramagnetic impurities (iron and manganese) and similarly with the cation exchange capacity (CEC) values (Figure 2). This is expected since iron and manganese, the most common paramagnetic impurities and largest contributors to grain susceptibility, are often associated with clay minerals. Clay minerals have large surface area and high magnetic susceptibilities leading to large internal gradients and extremely short values of

T2.
At low field, porosity can be accurately measured for both sandstones and carbonates. This is illustrated in Figures 3, 4, and 5, where 2-MHz NMR porosities are compared with gravimetric porosities for 28 1 samples. The sandstone suite has 2 10 samples, including 18 diatomitc plugs with porosities ranging from 45 to 65 p.u. The regression has a slope of 0.98, intercept of 0.5 P.u., and standard deviation of 0.8 p.u. A significant source of experimental error is the nonuniform sample size. There are 71 carbonate samples. The regression has a slope of 0.97, intercept of 1.4 PA., and standard deviation of 0.8 p.u. The higher intercept is due to two factors. First,
March-April 1997

Time constant, sec

Figure 6: T2 (dotted) and TI distributions (solid) for four sandstones measured at 2 MHz. The TI curves have been shifted by a factor of 2/3.

The Log Analyst

87

Stralcy. et al.

Loglo time constant, sec

r '

N M R FH;I (92 msec); p.u.

1

-3
I

-2
1

-1
I

-0

Figure 9: The volume of fluid expelled by centrifuging at 100

Loglo pressure, psi

Figure 7: Differential mercury injection curves (solid) with the Tz distributions (dotted) for t h e samples shown in the previous figure.

psi versus air plotted against t h e NMR free-fluid porosity calculated using 92-ms cutoff for a suite of carbonates. The dashed line is 1 : 1 .

about one-third of the samples were known to be incom-

30 -

I
0

, , '

I I

metric measurements. Second, many of the samples have surface vugs that were not completely dried prior to the NMR measurements.
PORE SIZE DISTFUBUTIQNS

,4

200

,
0

00

d0

+,*0+

3'

10 -

# # +

$ 4

+ + +

T, distributions from low-field CPMG data are similar to T I distributions and to pore-size distributions from mercury porosinietry measurements (Kleinberg et al., 1993; Morriss et al., 1993; Morriss et al., 1994). T , and T, distributions for four sandstone samples from different formations are displayed in Figure 6 to illustrate the similarity. The T I distributions have been shifted to shorter times using a factor of 2/3. This shift is required to roughly compensate for the different surface relaxivity for T , and T, processes. Figure 7 compares T, distributions and differential mer-

N M R FFP (33 msec), p.u.

Figure 8: The volume of fluid expelled by centrifuging at 100 psi versus air plotted against the NMR free-fluid porosity

T2P = 5.33 s-psi. Similar results have been reported for other sandstones and diatomite samples (Morriss et al., 1993; Morriss et al., 1994).
FREE-FLUID POROSITY

calculated using a 33-ms cutoff for a suite of sandstones. The dashed line is 1 : 1 .
XU

T, distributions can be used to estimate the volume of

March-April 1997

The Log Analyst

Core Analysis by Low-Field N M R

104

10-3

10-2

lo-'

Cutoff on T2 distribution, sec

Time constant, sec

Figure 10: Avertical cutoff on a 7 ' 2 distribution separating the

Figure 11: The minimum in the average deviation, at 3 ms,

distribution into the long T2 porosity (shaded) and the short T2 porosity. The solid curve i s the distribution for a fully saturated sandstone. The dashed curve shows the same sample after centrifuging at 100 psi versus air. The dotted vertical line is located at a cutoff of 33 ms. The long T2 porosity found in this manner is called the free-fluid porosity.

indicates the best agreement between clay-bound water volume from membrane potential measurements and the T2 distribution porosity for the smallest pores.
CLAY-BOUND WATER

producible fluid in a manner similar to that described for TI distributions (Straley et al., 1991). The NMR estimate of producible porosity is usually referred to as the free-fluid porosity, or free-fluid index. This technique is based on an expectation that the producible fluids reside in the larger pore spaces, whereas the bound fluids are held in the smaller pores. Hence, a T2 cutoff (i.e., a pore-size cutoff) may be applied to the T2 distribution that divides the total NMR porosity into boundand free-fluid porosity. The value of the T2 cutoff is determined by the best agreement between the free-fluid porosity and the volume of fluid that is centrifuged from the sample at 100 psi air-brine equivalent capillary pressure. Good agreement is obtained with cutoff values of 33 ms and 92 ms for sandstones and carbonates, respectively. Results are shown in Figures 8 and 9 for 86 sandstone and 25 carbonate samples. T2 distributions obtained for samples after centrifuging PERMEABILITY show that the bound, or irreducible, fluids are held in the small pores that have short T2 values. For example, Figure k = T <(b4 and similar expressions have been used to 10 compares T2distributions for a sandstone sample before estimate matrix permeability in sandstones (LaTorraca et and after centrifugation under air. After centrifuging, the al., 1993). Applying this equation with porosity as a decimal fraction and TI in milliseconds gives permeability estilongest T, values correspond roughly to the 33-ms cutoff.
March-April 1997 The Log Analyst

Estimation of clay-bound water volume from the T, distribution is based on an assumption that clay-bound water has the shortest T2 values in the distribution; clays have high surface area and small pores. Hence, the pore volume with T2 values shorter than a certain cutoff can be correlated with clay-bound water estimates from electrical or other measurements. Cation exchange capacity/pore volume (Q,,) was determined from electrical membrane potential measurements (Yuan and Diederix, 1987) for 45 oilfield sandstones from North America, South America, and Europe. Clay-bound water volumes were then calculated from the Q,,=values using the Hill-Klein-Shirley equation (Hill et al., 1979). Comparing these bound-water volumes with the cumulative porosity along the distribution, the best agreement with Q v , measurements occurs at a cutoff of 3 ms (Figure 11). This results in a minimum error of 0.75 p.u. A crossplot of the NMR estimate versus clay bound water is shown in Figure 12. These NMR measurements were made at 2.1 MHz using z = 80ys.

89

Straley, et al.

1 0

E
5
J

I-.

0
I

1 0

N M R k estimate, md
Flgure 13: The measured permeability for three sandstone suites versus the permeability estimated by ~ . ~ T $ L M $ ~ .

NMR clay bound water, p.u.

Flgure 12: The clay-bound water volume from membrane potential measurements compared with the cumulative porosity up to 3 ms on the T2 distributions.

103

1 0 2

+
0

+& +P:+k %+
+

1 0 '
a
A

a E
2 4 4 .

+f+&

+
+

E 100

+ % + + +$ + 0 @a+
@ +

+++

"+++*+

++++pa

*
e

10"
L

0 0

00.

I-

1 0 1

'

10-3 10-3
I
I

100

103

0
+ I

10

15 Porosity, p.u.

20

25

k estimate, md
Flgure 15: Crossplot of the NMR permeability estimate and the measured permeabilities for 23 Clearfork dolomites. This estimate is from the short-T2 portion of the distribution that yielded the smallest error (see Figure 16).

Flgure 14: Permeability plotted against the total porosity for the three sandstone suites used in the previous plot.

mates in millidarcies. Analogously, permeability may be estimated using

( 9) The logarithmic mean T,, T2,,L is the value of T2 that

bisects the area under the Tz distribution curve and is given by

k = CT&L44 .

90

The Log Analyst

March-April 1997

Core Analysis by Low-Field NMR

lo'

10-3

I2 limit (sec)
Figure 16: Variation in error from log-log crossplots of the measured permeability versus the NMR estimate calculated using a limited extent of the distributions (as indicated on the x-axis).

Viscosity, cp
Figure 17: The measured oil viscosity is plotted against T ~ M LBelridge . samples are indicated by o and other oil samples are indicated by +. The fit (solid line), for the data L from 0.7 to 1,000 cp, is given by a power law; i.e. T ~ M = I. ~ I ~ O . ~ .

The NMR permeability estimate, using Equation (9) and a premultiplier ( C ) of 4.6, gives good agreement with air permeability for 1 10 sandstones from three reservoirs (Figure 13). Better agreement can be obtained by adjusting the premultiplier for each reservoir. The improvement using the NMR estimate rather than porosity alone is evident from a comparison of Figures 13 and 14. In vuggy carbonates, a straightforward application of Equation (9) results in more scatter because of a greater variation in pore-body to pore-throat ratio. Use of the T2 distribution allows discrimination of the separated vuggy porosity, which has a weak contribution to the permeability (Chang et al., 1994). Figure 15 shows the permeability computed from Equation (9) for 23 vuggy dolomite samples from the Clearfork formation using an upper T2 cutoff of 750 msec. That is, values of both porosity and T2ML are computed from that portion of the distribution below 750 msec. The dolomite samples are the same samples described in the free-fluid porosity section. By using a truncated T2 distribution rather than the entire distribution, the error in permeability estimation was reduced from a factor of 4.8 to a factor of 2.6, as shown in Figure 16.
~~

OIL VISCOSITY

T2 distributions were determined for 64 bulk oil samples and values of T ~ Mwere L determined from their distributions. Many of these crudes could not be characterized by a single time constant and T ~ Mis L used to approximate T2B. As expected from Equation (3), values of T,, vary inversely with oil viscosity (see crossplot, Figure 17). The best fit of the data from 0.7 to 1,000 cp is given by T?ML = i.21p.
OIL SATURATION

'Mark of Schlumberger
March-April 1997

Low-field NMR measurements were used to determine the oil saturation and oil viscosity for diatomite samples from the Belridge field, Kern County, California. The laboratory measurements were used to evaluate logs recorded on the well with an experimental logging tool, the CMR' Combinable Magnetic Resonance tool. The results of the log and laboratory measurements are described by Morriss et al. (1994). The diatomite samples were measured three times. The first measurements were made in their native state condition (i.e., prior to cleaning), at which time the pore fluids consisted of connate water, gas, and native crude oils. The second measurements were made after displacing the gas with brine by placing the samples under brine with a light
91

The Log Analyst

Straley, et al.

30 I

i?o-3

10-2

10-1

100

10

20

30

Time constant, sec

Figure 18: The plot shows 2-MHz Tz distributions for a

Dean-Stark oil volume, p.u.

diatomite sample. The difference between the native state Figure 19: Comparison of oil volume determined by NMR material (solid) and t h e fully saturated samples (dashed) with Dean-Stark extraction made subsequent to NMR analysis. shows gas saturation. The dash/dot curve is the distribution for oil measured after t h e Hz0 signal was eliminated by dilution with D20. tnethods, low-field NMR is capable of quantifying a wide range of petrophysical properties with one nondestructive vacuum. Hence. differences between T, distributions from measurement. the first and second measurements, shown in Figure 18. 0 NMR measurements using low magnetic fields give accurate porosity values for both sandstones and carbonindicate the volume of gas present at the time of the native ates. Low tields have Hz I 0.1 gauss-s. state measurement. 0 Low-feld T2 distributions show good agreetnent with T , The samples were then soaked in DzO. During this prodistributions and correspond to pore-size distributions cedure the D 2 0 diffuses into the sample and dilutes the from mercury porosimetry. concentration of H 2 0 to a negligible level. After repeated 0 Free-fluid porosities calculated using cutoffs of 33 and soaking. the remaining NMR signal originates entirely from 92 ms for sandstones and carbonates, respectively, prethe crude oil, since D?O does not contribute an NMR signal dict the producible fluid. at the operating frequency of the laboratory spectrometer. 0 In sandstones, clay-bound water can be estimated from These T, distributions (Figure 18) were found to be similar T2 distributions using a 3-ms cutoff. to T, distributions for the bulk oils. 0 Sandstone permeability can be estimated with Hence, surface effects are not apparent; this indicates 4.6T;ML@4. For some vuggy carbonates, the NMR perthat the diatomite samples are predominantly water wet. meability estimate is improved by excluding the long T, Because of the absence o f surface effects, oil viscosity can porosity that is associated with the vugs. be estimated froin T2hfL of the distribution using Figure 17. 0 Measurements on a large suite of crude oils and viscosity The NMR porosity after D,O diffusion is a direct measz , = 1.2/~'.~ where , T,hfL is standards established that T ure of the volume of oil i n the sample and was found to be in seconds and is in centipoise. 0 Using D 2 0 diffusion, NMR can measure So and S,,, in i n very good agreement with oil volumes determined from native state core. This method allows oil viscosity to be Dean-Stark extraction (Figure 19). estimated in situ for water-wet rocks.

CONCLUSIONS

Low-field NMR measurements can be obtained rapidly, in about 20 minutes, on both native state and cleaned samples. Although not as well established as other core testing
92

ACKNOWLEDGMENTS

We thank Phil Dryden at Schluniberger-Doll Research and Noel Patterson and John Ferris at Shell Development
March-April 1997

The Log Analyst

Core Analysis by Low-Field NMR

Co. for providing some of the experimental data. The NUMAR Corporation graciously provided Shell Development Co. with the spectrometer used to make the 1-MHz measurements.
NOMENCLATURE

A CEC

D
G H k Mil
S t

Q,

T T I T 2 T 2 , T2B T 2 s
V
z
Xe

x rr
rl Y

signal amplitude cation exchange capacity diffusion constant of water magnetic field gradient magnetic field strength permeability net magnetization resulting from hydrogen cation exchange capacity/pore volume surface area elapsed time absolute temperature NMR longitudinal relaxation-time constant NMR transverse relaxation-time constant NMR mean logarithmic T2 bulk fluid transverse relaxation-time constant surface-induced transverse relaxation-time constant volume Carr-Purcell pulse spacing, one-half the echo spacing in a CPMG electronic magnetic susceptibility nuclear magnetic susceptibility caused by hydrogen viscosity gyromagnetic ratio for hydrogen porosity
REFERENCES

Bloembergen, N., Purcell, E. M., and Pound, R. V., 1948, Relaxation effects in nuclear magnetic resonance absorption: Phy. Rev., v. 73, p. 679-7 12. Brown, R. J. S. and Fantazzini, P., 1993, Conditions for initial quasi-linear T2-l vs z for CPMG NMR with diffusion and susceptibility differences in porous media and tissues: Phy. Rev. B, v. 47, p. 14823-14834. Chang, D., Vinegar, H. J., Momss, C. E., and Straley, C., 1994, Effective porosity, producible fluid and permeability in carbonates &om NMR logging, paper A, in 35th Annual Logging Symposium Transactions: Society of Professional Well Log Analysts. Cohen, M. H. and Mendelson, K. S., 1982, Nuclear magnetic resonance and the internal geometry of sedimentary rocks: Journal ofApplied Physics, v. 53, p. 1,127-1,135. Fordham, E. J., Sezginer, A., and Hall, L. D., 1994, Imaging multiexponential relaxation in the (y,log,(TI)) plane: Application to clay filtration in rock cores. Submitted to J. Mag. Res. for publication. Freedman, R., Processing method and apparatus for processing
March-April 1997

spin echo inphase and quadrature amplitudes kom a pulsed nuclear magnetism tool and producing new output data to be recorded on an output record: US Patent 5,291,137, March 1, 1994. Gallegos, D. P. and Smith, D. M., 1988, A NMR technique for the analysis of pore structure: Determination of continuous pore size distributions: J. Col. I. Sci., v. 122, no. 1, p. 143-153. Hill, H. J., Shirley, 0. J., and Klein, G. E., 1979, Bound water in shaly sands - its relation to Q, and other formation properties (edited by M. H. Waxman and E. C. Thomas): The LogAna/yst, V. 20, p. 3-25. Kenyon, W. E., 1992, Nuclear magnetic resonance as apetrophysical measurement: Nuclear Geoph.ysics, v. 6, p. 153-1 7 1. Kleinberg, R. L. and Horsfield, M. A., 1990, Transverse relaxation processes in porous sedimentary rock: Journal of Magnetic Resonance, v. 88, p. 9-19. Kleinberg, R. L., Sezginer, A., Griffith, D. D., and Fukuhara, M., 1992. Novel NMR apparatus for investigating an external sample: Journal of Magnetic Resonance, v. 97, p. 466485. Kleinberg, R. L., Straley, C., Kenyon, W. E., Akkurt, R., and Farooqui, S. A., 1993, Nuclear magnetic resonance of rocks: T , vs. T2,SPE-26470, in Annual Technical Conference and Exhibition Proceedings: Society of Petroleum Engineers. Miller, M. N., Paltiel, Z., Gillen, M. E., Granot, J., and Bouton, J. C., 1990, Spin echo magnetic resonance logging: Porosity and free fluid index determination, SPE-20561, in Annual Technical Conference and Exhibition Proceedings: Society of Petroleum Engineers. Morriss, C. E., Freedman, R., Straley, C., Johnston, M., Vinegar, H. J., and Tutunjian, P. N., 1994, Hydrocarbon saturation and viscosity estimation kom NMR logging in the Belridge diatomite, paper C , in 35th Annual Logging Symposium Transactions: Society of Professional Well Log Analysts. Morriss, C. E., MacInnis, J., Freedman, R., Smaardyk, J., Straley, C., Kenyon, W. E., Vinegar, H. J., and Tutunjian, P. N., 1993, Field test of an experimental pulsed nuclear magnetism tool, paper GGG, in 34th Annual Logging Symposium Transactions: Society of Professional Well Log Analysts. Straley, C., Morriss, C. E., Kenyon, W. E., and Howard, J. J., 1991, NMR in partially saturated rocks: Laboratory insights on free fluid index and comparison with borehole logs, paper CC, in 32nd Annual Logging Symposium Transactions: Society of Professional Well Log Analysts. LaTorraca, G. A., Dunn, K. J., and Brown, R. J. S., 1993, Predicting permeability from nuclear magnetic resonance and electrical properties measurements, SPE-18272, in Annual Technical Conference and Exhibition Proceedings: Society of Petroleum Engineers. Vinegar, H. J., Tutunjian, P. N., Crabtree, P. T., Raffaldi, F. J., DiFoggio, R., and Edelstein, W. A., 1991, NMR spectroscopy of tight, gypsum-bearing carbonates: The Log Analyst, v. 32, p. 527-533. Yuan, H. H. and Diederix, K. M., 1987, The role of membrane potential measurements in shaly sand evaluation, paper GG, in 28th Annual Logging Symposium Transactions: Society of Professional Well Log Analysts.

The Log Analyst

93

Straley, et al.

ABOUT THE AUTHORS

Christian Straley has a BS degree in chemistry from Washington and Lee University and an MS degree from the University of Delaware with an emphasis in physical chemistry. He has worked at Schluniberger-Doll Research for 10 yr and is presently associated with the Interpretation Sciences Department. His primary research interest is the NMR properties of rocks and nuclear magnetism logging; he has also worked on the low-frequency electrical properties of fully and partially saturated rocks and on clay properties. Mr. Straley may be contacted at his e-mail: [email protected]. Dan J. Rossini is Manager, Health, Safety, Environment and Facilities for Schlumberger-Doll Research. Prior to this position he worked on NMR properties and CMR support at SDR. Dan holds a BS degree in mechanical engineering and is currently finishing work on a MS degree in environmental engineering. Harold J. Vinegar is a research advisor with Shell Exploration and Production Technology Company in Houston. Since joining the Shell Bellaire Research Center in 1976, his research interests have been in shaly sand evaluation, NMR logging, CT and NMR Imaging. enhanced recovery processes, and environmental reme-

diation. He holds MA and PhD degrees in physics from Harvard University and a BA from Columbia. He served as president and vice-president technology of the Society of Core Analysts. Harold is an associate editor for The Log rlnalysf and was coorganizer of the 1994 SPWLA Forum on Well Logging Applications of Nuclear Magnetic Resonance. He received the SPWLA Distinguished Technical Achievement Award in 1992.

Pierre Tutunjian is a senior research chemist at Shell Development Company. His research has been in the application of electron spin resonance and NMR spectroscopy and imaging techniques for products and exploration and production R&D. He holds a BA degree in chemistry from New York University and a PhD degree in physical chemistry from the Massachusetts Institute of Technology. Chris Morris graduated from the University of Aston, Birmingham in 1975, with a BSc degree in civil engineering. After joining Schluniberger in 1978, he worked in Canada as a field engineer, log analyst, sales engineer, product development engineer, and interpretation development manager. He is presently marketing development engineer for the Electrical Engineering Department, Schlumberger Well Services in Houston.

94

The Log Analyst

March-April 1997

Summary: Petrophysics in Three Dimensions Topical Conference

tailing with a clearer insight into the ordering role of geological architecture. Petrophysics was identified as the fulcrum of ANISOTROPY IS GOOD FOR YOU! this contemporary philosophy, its function being partly in The 1996 SPWLA Topical Conference on Petrophysics in response to advances in logging and drilling technology, with Three Dimensions, was held in Taos, New Mexico, during requirements for 3D interpretation of anisotropic formations. the period 13-17 October, 1996. The meeting was attended by Attendees therefore gained a contemporary insight into the 43 technical participants, 10 accompanying persons, and the direction of petrophysics as we approach the millennium. It was proposed that an SPWLA committee be established spirit of Kokopelli, the mischievous pied piper adopted by the to capture the current evolution of the subject and to formuPueblo Indians as their god of fertility. The stimulating technical debate and the high degree of late it into a modern syllabus to guide the teaching of formaresonant synergy has re-opened the question of the level at tion evaluation in universities. In that way, it was argued, the which the number of attendees should be set. Most attendees spirit of the meeting would live on by contributing to the felt that at this meeting the numbers were ideal for productive development of those in whom our future is vested. Many thanks to Vicki for all her organizational input and discussion of such a pioneering topic. to the attendees whose cross-fertilization generated such a In broad terms, the discussions dwelt on the concept of an expanded 3D reservoir framework based on seismic interpre- memorable and contributive technical event. Kokopelli tation, with progressively sharper spatial resolutions, dove- would have been proud of you.

Submitted by: Gary Beck and Paul Worthington

March-April 1997

The Log Analyst

95