NMR Measurement of Bitumen at Different Temperatures

1
NMR Measurement of Bitumen at Different
Temperatures

• Elton Yang, George J. Hirasaki

Rice Consortium Rice University, Houston,
TX March 26th, 2008
2
Introduction

• Bitumen is characterized by its high viscosity
  and density, which is a major obstacle to both
  its recovery and property investigations.
• Due to the loss of T2 information shorter than
  echo spacing (TE) and interference of water
  signal, estimation of heavy oil saturation and
  viscosity from NMR T2 measurements is problematic.

3
Sample

• Athabasca Bitumen sample (bitumen water
  clay), contained in a cylinder glass tube.
• Tube I.D. 4.66 cm
• Sample height 3.70 cm
• Sample volume 63.1 mL

4
CPMG Measurement for Bitumen at 30 oC
5
Approach to Compensation of Lost T2 Information

• Determine initial magnetization M0 from FID.
• Supplement M0 into the regular CPMG data and do
  the interpretation.

6
Mo and T2 Can be Estimated from FID
7
Interpretation of Supplemented CPMG Data by Using
Standard Multi-exponential Model
8
Approach to Solution

• Distinguish oil and water response by using
  contrast in relaxation.
• Assume relaxation time of emulsified water is
  correct as interpreted in standard way.
• Assume ?f of bitumen is equal to difference
  between M0 and ?f of water.
• Assume lognormal distribution for bitumen and
  redo the interpretation. Here, unknowns are log
  mean T2 and standard deviation.

9
Fitting Supplemented CPMG Data at 30 oC

• CPMG data is supplemented with a specified Mo
  from FID.
• Assume lognormal distribution for bitumen T2
  distribution.

10
Corrected T2 of Bitumen from Lognormal
Distribution Model has Little Dependence on TE
11
T2 Estimated from CPMG is Echo Spacing Dependent.
This can be Fixed by Specifying Mo from FID and
Assuming Lognormal Distribution for Bitumen
12
Apply this method for the estimation of T2 at
different temperatures.
13
Experimental Method

• Sample is placed in thermal water bath with
  interested temperature for gt 4 hours before NMR
  measurement.
• Sample tube is wrapped with a 4-layer paper
  insulation during NMR measurement. The
  temperature of magnetic field is kept at 30 oC.
• FID is performed before each CPMG. The total time
  for making one measurement (FID CPMG) lt 1 min.
• Measurement is repeated for at least 3 times to
  ensure the reliability of experimental data.
• The interested temperature range is 8 90 oC,
  which is adequate for Canadian bitumen logging.

14
Temperature Change of Athabasca Bitumen Sample in
90 oC Case
Fig. (a) Heating Process It only took about 90
min. to heat the bitumen sample from room
temperature to 90 oC.
Fig. (b) NMR Measurement Process For the 90 oC
case, the sample temperature deviation was 1.6
within 1 min.
15
Change of Mo with Temperature is Governed by
Curies Law
(1)
Curies Law
Given a sample system, each parameter on the
right side of above equation is constant except
for the temperature, thus equation (1) becomes
(2)
16
Estimate Mo from FID at 8 60 oC

• Apparent Mo extrapolated from FID increases with
  temperature.
• Extrapolated Mo reaches a constant when
  temperature gt 40 oC (indicated by arrow).

• FID signals probably attenuate much faster within
  first 80 msec dead time. Loss of FID information.

17
FID of Bitumen Sample at 8 and 20 oC, Obtained by
Using 20 MHz Bruker Spectrometer with a Dead Time
of 50 msec
18
Estimate Mo from FID at 60 90 oC

• Estimated M0 of bitumen at T 60 oC starts
  decreasing as temperature increases

19
Mo of Bitumen Sample at 8 90 oC

• Difference between 60 oC-based prediction and 90
  oC-based prediction is 3.2 .
• Correct M0 by using Curies Law on the basis of
  60 oC value.

20
Fitting CPMG Signal at Different T, Constrained
by Curies Law Corrected M0 and Assuming
Lognormal Distribution for Bitumen
21
Interpretation of CPMG, Supplemented with Curies
Law Corrected M0
22
Interpretation of CPMG, Supplemented with
Apparent M0 without Curies Law Correction
23
Calculate Bitumen HI and Water Saturation
Definition
Here, assume the difference of sample volume
within our interested temperature range is
negligible, thus
? f b/bw sum of f of bitumen part in the
mixture sample M0,w initial
magnetization of pure water as standard Vb/bw
volume of bitumen part in the mixture
sample Vt total volume of mixture
sample, assumed to equal to volume of pure water
as standard Tstandard standard
temperature, K Tinterest
temperature of interest, K
24
Estimated Water Saturation of Bitumen Sample
25
T2 Distribution of Bitumen Sample at 60, 70, 80,
90 oC
26
Bitumen HI has Incorrect Dependence on
Temperature unless Constrained by Proper Mo from
Curies Law
0.82
27
Viscosity of Bitumen Estimated from Current
Correlation has Significant Discrepancy from
Experimental Value
28
Intramolecular Dipole-Dipole Interactions

• Normalize relaxation time, viscosity and
  viscosity/temperature ratio with respect to 2 MHz

• Rotational correlation time (McConnell, 1980)

• Normalized relaxation times by intramolecular
  dipole-dipole interaction for spherical molecule

Here
29
T1,LM T2,LM Normalized with 2MHz Larmor
Frequency
30
Conclusions

• The echo spacing restriction of regular CPMG
  measurement on highly viscous bitumen can be
  overcome by specifying the M0 in CPMG raw data
  and assuming lognormal distribution for bitumen
  during the interpretation.
• Apparent Mo of bitumen at low temperatures has
  incorrect dependence on temperature due to the
  loss of FID signal within initial decay period.
  This can be fixed by using Curies Law.
• Given proper Mo and cut-off between oil and
  water, the HI and Sw can be evaluated by using
  the new method.
• The T1 and T2 of Athabasca bitumen follow the
  trend of previous literature data. The existing
  T2 vs. viscosity correlations, which are good for
  the oil with relatively low viscosity, are not
  suitable for the samples with extremely high
  viscosity like Canadian bitumen.

31
Acknowledgements

• The financial support of Consortium on
  Processes in Porous Media at Rice University and
  DOE is gratefully acknowledged.
• Dr. Harold Vinegar is acknowledged for his
  important advices.
• Dr. Zvi Taicher is acknowledged for the use of
  his 20 MHz Bruker spectrometer.