SPE

Society of Petroleum Engineers

SPE 21279

Yield Points in Oilfield Cement Slurries

P.E. Clark, * L. Sundaram, and M. Balakrishnan, U. of Alabama
·SPE Member

Copyright 1990, Society of Petroleum Engineers, Inc.

This paper was prepared for presentation at the SPE Eastern Regional Meeting held in Columbus, Ohio, October 31-November 2, 1990.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect
any position of the Society of Petroleum Engineers. its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society
of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuoiJs acknowledgment
of where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836 U.S.A. Telex, 730989 SPEDAL.

Abstract yield stress the fluid acts as an elastic solid and above the
yield stress the fluid flows with a plastic viscosity (I!p)'
The determination of yield points in cement slurries is This is a statement of the von Mises crIterion for fluids
important in the overall description of slurry flow with a yield point. The equation (Bingham equation)
properties. Yield points'affect both the start-up pressure which governs the flow of a fluid with a yield stress is
after a temporary shut-down and void filling properties of given by
cements. Values for the yield point are normally obtained
by extrapolation of rotational or tube flow rheometer data. 't = 'ty + IIp y (1)
This technique is subject to both experimental and
analytical errors.
Recently, a number of papers have described a more
direct method of measuring yield points in slurries. A
where 't is the shear stress and is the shear rate. r
Unfortunately, for fluids equation 1 is an idealization
rotational viscometer equipped with a vane rather than a of the actual flow. In controlled shear rate experiments
bob can be used to accurately measure yield points. The the stress normally builds exponentially at small shear
yield surface with the vane test fixture is in the fluid rates to yield data which is similar to the data points
rather than at a fluid-solid interface as it is with a solid bob shown in Figure 1. Papanastasiou1 ,2 presents a
or a tube flow apparatus. Experiments with a vane fixture constitutive equation (2) which accurately describes the
using a shear rate controlled rheometer can be subject to behavior of a real system over a wide range of shear rates.
errors caused by wind-up of the torque spring. In this
study a controlled stress rheometer was used to
circumvent these problems. 't = 2II! + 't
y
[1 -lm ITIoll/l)]} D
Yield points for various oilfield cement slurries are \ 12 lIDIl/
2
(2)
presented. Data reflect the differences in additives, aging
time and water content. The vane technique can be used Where I! is the viscosity of the fluid after yielding, D is the
to produce accurate and reproducible measurements of rate of strain tensor, 't is the stress tensor, and lID is the
yield points in cements. Using a controlled stress . second invariant of the rate of strain tensor. This rather
instrument has a number of advantages in making these imposing equation can be reduced to
measurements.

Introduction
or
Yield stress ('ty ) of a fluid is defined as the minimum 't = Kf + 'ty (1 - e-mr) (3)
stress required to deform the fluid. In this paper 'ty will
also be referred to as the Bingham yield stress. Below the for viscometric, one dimensional flows. Equation 3 is
similar to the Herschell-Bulkley3 equation with a term (1 -
e-mf) that modifies the yield stress parameter. When n is
References and illustrations at end of paper. equal to one and m is large, this equation reduces to the
167
2 Yield Points in Oilfield Cement Slurries SPE 21279

advantages. This is in contrast with data presented by

form of Equation 1. Equation 3 can be used to fit the data Haimoni and Hannant9 on oilfield cement slurries. They
shown in Figure 1. The fit is shown by the solid line report that there is a large difference between the yield
through the data. values measured with a vane geometry and those
Because of the apparent deviation from the von Mises measured with a concentric cylinder fixture. In this study,
criterion, there is a great deal of controversy over the exact the authors measured the yield stresses from cement
nature of the yield stress. Barnes and Walters 4 argue that slurries which had been mixed in accordance to API
the yield stress is merely an artifact of the measurement procedures.
process. They found that in fluids that exhibit a yield Controlled stress measurements provide a somewhat
stress under normal measurement conditions, failed to different approach to the measurement of yield stresses
exhibit a yield stress under conditions where the stress than is proposed by the API8 standard procedure. There
was applied for long periods of time in a controlled stress are three different experimental techniques that can be
experiment. This led Hartnett and HuS to undertake a used to determine the yield stress with a controlled stress
series of falling ball experiments in a 2500 ppm solution of instrument. Creep measurements yield distinctly
Carbopol 960 that showed that the ball did not move different curves when run at initial stresses above and
under the influence of gravity for an extended period of below the yield stress. The solid line in Figure 3 is a
time. From their data, they conclude that yield points in representation of creep behavior of a fluid when the
. fluids exist and must be considered in engineering design. initial stress is below the yield stress. The curve is
In a response to the paper by Hartnett and Hu, Astarita6 characterized by a rapid build-up in the strain followed by
presents an interesting and humorous discourse on what a straight line portion with a zero slope. A test run at a
constitutes a reasonable length of time for an experiment stress significantly above the yield stress results in a curve
which purports to measure a yield point. Cheng7 has also similar to the dashed line in Figure 3. This curve shows a
commented on the time-dependence of the yield stress. continuing increase in the strain or strain rate after the
All of this leads to the conclusion that yield points can be rapid initial build-up. Tests run at stresses slightly above
time dependent and the magnitude of the yield point can the yield stress result in a curve similar to the dotted line
depend strongly upon the type of experimental in Figure 3. Creep tests can be used to bracket the yield
measurement. stress value.
The measurement of yield points is important in A second technique involves running a flow test. This
rheological characterization of oilfield cement slurries. test is analogous to a shear rate sweep except that the
Yield points are typically determined by extrapolation value of the applied stress is varied and the shear rate is
from rotational viscometer data 8 . There are a number of measured. Normally there is little or no motion (shear
factors that can bias this type of measurement and result rate) generated until the yield point is reached (Figure 4).
in yield points that may be lower than the true yield A step-stress test, which is a variation of the flow test with
point. 9 . Slip at the wall of the measurement fixtures is longer times between changing the stress than are
probably the largest contributor to incorrect yield point normally found a flow test, can also be used (see reference
data for slurries. The slip is caused by the formation of a 13). In this study we have employed controlled shear rate
thin layer of liquid near the wall which does not contain tests using the API8 standard procedure and both creep
the same concentration of particles as the bulk material. and flow tests using a controlled stress rheometer.
Yoshimura and Prud'homme 10 have published a method
of correcting for slip in data taken using the concentric Experimental
cylinder geometry, but it requires the use of two different
rotor-bob combinations. Since this is not specified in the Cement slurries were prepared according to API8
API8 procedures, the tests are almost never run using procedures except that the slurries were continuously
different rotor-bob combinations. Similar problems are stirred in a blender rather than in an atmospheric
observed in pipe or tube flow viscometers. consistometer. The cement was a Class H cement. All
Duzy and Bogerl l present a method of measuring yield slurries were prepared with tap water. Controlled shear
points using a technique common to soil mechanics. This rate experiments were run with a Bariod variable speed
technique uses a vane test fixture (Figure 2) rather than a viscometer. Measurements were taken at each shear rate
solid bob. The vane minimizes the surface contact after waiting 20 seconds 14. When tests were run with both
between the metal at the edge of the vane, and the slurry instruments every effort was made to keep the time
and the actual yield occurs at a fluid-fluid interface. Dzuy between initial mixing and the start of the test the same
and Boger12 compare the vane technique with yield stress for each instrument. Controlled stress measurements
values measured by capillary rheometry for a red mud were made with a Carri-Med CS 100 Controlled stress
slurry and they conclude that the vane technique provides Rheometer. The test fixture was a vane similar to that
accurate values of the true yield stress without the need shown in Figure 2. with a height (H) of 30.5 mm and a
for extrapolation. diameter (D) of 18.2 mm. Equation 4, given by
Yoshimura, et al. l3, compare results obtained with a Yoshimura, et a1.1 3, is used to calculate the yield stress
concentric cylinder geometry, parallel disk geometry at from the data.
two gap heights, and vane geometry using a controlled
stress instrument in a series of oil-in-water emulsions.
They conclude that all three techniques provide
comparable results and each technique offers some ~y ~~~Ir~ iJ
= + (4)

168
SPE 21279 P. E. Clark, L. Sundaram, and M. Balakrishnan 3

Where Tm is the measured torque, H is the height of the causes an apparent shear plane to form in the slurry18. In
vane and D is the diameter of the vane. several tests, the sudden failure of the fluid was
accompanied by a separation of a water layer on the top of
Results and Discussion the slurry. If this is the correct explanation, the measured
yield stresses may be lower than would be expected from
Rheological experiments on cements are notoriously other measurements.
difficult to reproduce15. Slip, particle-particle interactions, In order to determine the effect of concentration on the
chemical reactions, and nonhomogeneous flow fields16 rheology of cement slurries, tests were conducted with
are partly to blame for the measurement problems. 10% less water and 10% more water than specifie.d by the
Mixing and aging of the slurry can also affect the API test procedure. The results generated using the
measurements. It is not always easy to tell from plots of Bariod viscometer are shown in Figure 8. Data for the
the data whether to analyze the experimental results with correct amount of water are shown in Figures Sa and 5b.
a power law (pseudoplastic) or a Bingham plastic model. The Bingham yield stress from the best fit lines are 824
The API8 Spec 10 gives examples of the use of both dynes/cm2 for the 10% less water slurry, 327 dynes/cm2 for
models, however, neither case presented is particulary the 10% more water slurry and 509 dynes/ cm2 for the
strong and arguments can be made for using either fluid slurry with the correct amount of water.
model to fit both sets of data. Figure Sa presents data from Bentonite is a common additive for cement, and its
a neat class H cement. The dashed line represents the best addition can have a profound effect on the measured yield
fit line through the linear segment of the data plotted in stress. Data from the Bariod viscometer are shown in
linear space. If the slurry behaves as a Bingham plastic, Figure 9. Again, these data are similar to that shown in
the intercept of the dashed line (509 dynes/cm2 ) is the Figures Sa, 6, and 8. From extrapolation of the best fit line
Bingham yield point. A log-log plot of the data (Figure 5b) to the y-axis, the yield stress is estimated to be 591
indicates that a power law model might also fit the data. dynes/cm2 . A flow curve generated using the Carri-Med
During the experiment, the shear stress did not go to zero rheometer is shown in Figure 10. The yield stress from
as the shear rate approached zero indicating that the slurry the best fit line is 656 dynes/ cm 2 • Creep test data presented
exhibited a yield point. This observation is probably the in Figure 11 brackets the yield stress around 640
best indication of the presence of a yield stress. These data . dynes/cm2 . Differences between the yield stresses
emphasize the necessity of collecting a number of data measured by the flow and creep tests are insignificant.
points during any study of a cement slurry. The five or six The discrepancy between the data obtained with the
data points normally collected (shown in Figures Sa and Bariod viscometer and the Carri-Med rheometer is
5b) with a six-speed instrument are not adequate for a probably to differences in the test geometry or the
complete analysis. This is especially true since data taken construction materials used in the test fixtures.
at the two lowest shear rate values often contain a great Flow test data shown in Figure 4 emphasize the
deal of error. sensitivity of yield stress measurements to the time
In order to understand the flow of cement slurries in between mixing and measurement. Some of the
viscometric flow and thus obtain a better understanding differences between the data obtained with the Bariod
of cement flow in field applications, a study has been viscometer and the Carri-Med rheometer can be attributed
undertaken using both an oilfield standard instrument to the timing of the tests.
and a Carri-Med Controlled Stress Rheometer. Figure 6
presents data obtained in a test of a neat cement with 10% Conclusions
excess water using the Bariod variable speed viscometer.
While the linear plot looks much the same as that shown Data are presented to show that yield stresses for
in Figure 5, the log-log plot shows that there is a deviation cement slurries can be measured using both a
from linearity at the lower shear rates. This plot indicates conventional rotational viscometer and a controlled stress
that the data might best be described by a Bingham plastic rheometer. The controlled stress rheometer provides a
model with a Bingham yield stress of 324 dynes/cm 2 . A more direct measurement of the yield stress using flow
log-log plot of the viscosity versus shear rate indicates that tests since the yield stress must be exceeded before any
the non-linearity may just be the transition from the measurable motion occurs. Controlled stress
power law region to the lower Newtonian region for a measurements also indicate that most cement slurries
power law fluid. However, the curve appears to change exhibit a yield stress within the time frame of the
slope again at low shear rates indicating that a yield point experiments. This does not prove or disprove the thesis
may exist. The data presented in Figure 7 are from a creep of Barnes and Walters 4 . Due to chemical reactions taking
test on the same sample run with a applied stress of 75 place within cement slurries, these slurries can not be
dynes/cm 2 using the controlled stress rheometer equipped subjected to the long term rheological tests necessary to
with the vane fixture. If the test had been terminated after test their hypothesis.
20 to 30 seconds, the applied stress would have been The data presented here also indicate that it is
considered below the yield stress. However, at dangerous to use a log-log plot of shear rate-shear
approximately 50 seconds the vane begins to move stress data to discriminate between slurry data that can be
indicating that the yield stress has been exceeded. The analyzed with a Bingham plastic model and those that
sudden, catastrophic failure of the fluid is probably a result must be treated with a power law model. Plotting the log
of a stress induced inhomogeneity in the fluid15,17 that of the viscosity versus the log of the shear rate can

169
4 Yield Points in Oilfield Cement Slurries SPE 21279

sometimes be instructive, but this type of plot can not Measuring Yield Stresses," J. of Rheology, (1987) 31, No.8,
always be used to discriminate between the two types of 699-710.
behavior. The residual dial reading at zero shear rate after
a test is completed is also an indication of the presence of a 14. Barkat, 0., Shaughnessy, R, and Clark, P.E.: "A
yield point. However, the value of the yield stress Rheological Study of a Cement Slurry, Proceedings of the
provided by this measurement will be different from the Third International Symposium on Liquid-Solid Flows, p.
Bingham yield stress. It is probably best to treat all cement 115 (1988).
slurries with a Bingham plastic or Herschel-Bulkley
model. 15. Shaughnessy, R and Clark, P. E.: "Rheological
Behavior of Fresh Cement Pastes," Cement and Concrete
References Research, (1988)18, No.3, 327-341.

16. Cameron, J.R: "Viscometry of Nonhomogeneous

1. Papanastasiou, T.C.: "Flows of Materials with Flows and the Behavior of a Titanium-Crosslinked
Yield," J. Rheol. (1987) 31, No.5, 385-404. Hydroxypropyl Guar Gel in Couette Flow," J. Rheol.
(1989) 33, No.1, 15-46.
2. Ellwod, K.RJ., Georgiou, G.c., Papanastasiou, T.C., and
Wilks, J.O.: "Laminar Jets of Bingham-Plastic Liquids," J. 17. Barns, H.A and Carnali, J.O.: "The Vane-in-Cup as a
Rheol., (1990) 34, No.6, 787-812. Novel Rheometer Geometry for Shear Thinning and
Thixotropic Materials," J. Rheol. (1990) 34, No.6, 841-866.
3. Pilehvari, A and Clark, P. E.: "Rheology of Hydraulic
Fracturing Fluids: Wall Slip During Viscosity 18. Bird, RB., Dia, G.c., and Yarusso, B.J.: "The Rheology
Measurement," JPT, (1985) 37, 1840. and Flow of Viscoplastic Materials," Rev. Chern. Eng.
(1982) 11 No.1, 1-71.
4. Barnes, H.A. and Walters, K.: "The Yield Stress Myth,"
Rheologica Acta, (1985) 24, 323-325.

5. Hartnett, J. P. and Hu, R Y. Z.: "The Yield Stress-An

Engineering Reality," J. Rheology, (1989) 33 No.4, 671-679.

6. Astaria, G.: "Letter to the Editor: The Engineering

Reality of the Yield Stress," J. Rheol., (1990) 34, No.2, 275-
277.

7. Cheng, D. C-H.: "Yield Stress: A Time-dependent

Property and How to Measure It," Rheol. Acta (1986) 25
No. 5,542-544.

8. "Specification for Materials and Testing for Well

Cements," API Specification 10, Fifth Edition, July 1, 1990,
p.76-88.

9. Haimoni, A and Hannant, D.J.: "Developments in the

Shear Vane Test to Measure the Gel Strength of Oilwell
Cement Slurry," Advances in Cement Research, (1988) 1,
221-229.

10. Yoshimura, A. and Prud'homme, RK.: "Wall Slip

Corrections for Couette and Parallel Disk Viscometers," J.
of Rheology, (1988) 32, No.1, 53-67.

11. Dzuy, N.Q. and Boger, D.V.: "Yield Stress

Measurements for Concentrated Suspensions," J. of
Rheology, (1983) 27, No.4, 321-349.

12. Duzy, N.Q. and Boger, D.V.: "Direct Yield Stress

Measurement with the Vane Method," J. Rheol. (1985) 29,
No. 3,335-347.

13. Yoshimura, A.5., Prud'homme, RK., Princen, H.M.,

and Kiss, AD.: "A Comparison of Techniques for

170
 80 S~E 2 1279
I
~~~~..
..
~' .
60
I
M'§
......
Ul
OJ
I':
>,
~
Ul
Ul
40
OJ
.b
{fJ
k
cO
OJ
--,- 21
Top view of the vane
.<:;

dn V
{fJ

20J ! I.
-
Synthetic Data
Best Fit Line Using EquatIon 3
. - - Ideal Bingham Plastic
.
I ~O_
;.. Dt-":

Dt - cup diameter
D - vane diameter
o--¥ I I I I H - vene height
o 2 4
1
6 8 21 - vane depth below the surface
Shear Rate (5- ) 22 - distance between the vane and the bottom of the cup
Fig. 1-Representatlon of the stress bUildup in a controlled shear rate experiment. Fig. 2-Vane test fixture.

;j

300

000000000
"
OooocP°
,--' S 250
M

u
......
000000
........ Ul 0000
2 OJ

~ -- _ i OO
0000
-- -_..... .. ..-.- cP°
~
~.~.~
,;,;'
..._ .. .H_..... ···-··_··-·~·_····_· Ul
,00
Ul
",,;' ~ 200
/"~"'~_._-_
.. _-- en k
cO
OJ

:~# c7J
,'//
/1
,'1;-
150
, Stress below the yield point
,, Stress above the yield point
.. Stress slightly above the yield point
"
Time after mixing 26 minutes
Time after mixing 33 minutes

100

o 10 20 30 40 50 60 70
Time Shear Rate (5- 1)
Fig. 3-Creep tests on a fluid with a yield point. Fig. 4-Flow test on a class H cement with 10% excess water.
 Shear Rate (S·I)
SPE 2127 </
800
10 100
--- 4 5 6 7 a 4 5 6 7 B 4 5
1000
6 7 B

o
_1lI- _- - - 0 - - - - -0- - - - - -- 181 500 1000

0-----°-- Q 18l_---o ---- e---- o--- i8J

600
,,",,-" a -0- --" a.---- g -" --"- ,," 400 I8l- _-9----9----- ----
_-------'3---- - 00 0 0 :::Jcor:::PDJ
M
S U
......
Ul o
M
S U
......
._--.-- o0
o
0 0
(f)
::r
ro
~
>."I': o Ul
300 -j 0 o
~
400 o
"~ o
;(?
'1
ro
en
Ul
Ul ~ o en
(lD 100
~"
Ul 0
Ul 0:
....
181
o ~
(fJ o
o '8ro
""'
200J 0 en
..c:: iii ......
(fJ 1181 ()

6J" 0 ~
200 o Experimental data
. . " Least Squares line 0
Slope = 0.515 181 a Experimental data (linear plot)
Intercept = 509 100--1 . - - Best fit line
I!!I Points from the standard five speeds g Experimental data (standard five points)
& o Experimental data (log plot)

o --j "

100 200 300 400

I
500
o~ 0 100
I
200 300 400 500 600
10

Shear Rate (S·I) Shear Rate (S·I)

Fig. 5a-Neat class H cement measured with a rotational viscometer. Fig. 6-Neat class H cement with 10% excess water measured with a rotational viscometer.

;;j
1000 1.0

-oD~ 0.8
o P-g{)
~/

M'§ O .. .'
0····
......
0.6

I
0.13···-13"
•• .l8I •••• ~
~
Ul
.••• ··0 ....
""'
Ul

~ 181
..c::
(fJ

iii 0.4

6J"

0.2
o Experimental data
- - . Least Squares line
1m Points from the standard Ii ve speeds

1 0 O-'t,'---T.--.-rl-r.-.,-,-1"""---T.--r.-r.-r.,.-.."""---"T'---.'-T'-r-r.-r---'rT"'r. o.o~ I •
4 5 6 7 8 4 5 6 7 8 4 5 6 7 8
10 100 1000 10 20 30 40 50
Shear Rate (S·I) Time (s)
Fig. Sb-Neat class H cement measured with a rotational viscometer. Fig. 7-Creep test on a neat class H cement with 10% excess water.
 Shear Rate (S'I)
SPE 2 127 9 0
10 100 1000 0
4 5 6 7 , I 4 56 7 4 5 6 7
0
o 10% less water (linear plot)
__ 0-- __ 0
_----------
0 0
~1 0000 740 0
1200 - '" 10% less water (log·log plot) 0
---Bestfitline 0
• 10% more water (linear plot)
.& 10% more water (log-log plot)
0._--- 0 ----°-- 7
0
..... Best fit line 0

.---- -- _-_.--- _--~-~- ---;; ----

1000 720
0
N
S U 0
N'§ gl
ro
"-
<J) 0
o Q)

"-
:G
800 '"
.... K 0
o
(/)
::rro :s 700 o
K '" o
:s '"
o /::, /::,/::,/::,ti'M <J)

E o
~1 0:
<J)
<J)
600 /::,M (/)
o
~
Q)
lJ o /::, 000 ~
(/)
o /::, Q) o
~ "-
()
'" 6J 680 o
:.~.::~
Q)

6J 400 r§':
o o •••
..."'- _....
• .........•........•.........•
A A A f';..A..•........
• • :
A
: .. - .• '
'
~
o
o

o
/::,
o
200
I• /::,
/::,

'"
'" '
660
b I0 Experimental data I
10
o ~ '"
'" '" Hoo
I I I I I
100 200 300 400 500 o 10 20 30 40 50
Shear Rate (S·I) Shear Rate (S·l)
Fig. a-Composite or rotational viscometer tests on neat class H cement. Fig. 10-Flow test on a class H cement with 1% bentonite.

<;j

1.2
o .0_---- 0 ----- 0 -
800
- Initial stress 400 dynes/cm 2
- - Initial stress 398 dynes/cm 2
____________-_0 -6 _-- - -e- -- - _.0- -- - - 9- ------ -- -- 1.0

N
S 600u o
"-
:G o
o ...~ 0.8
K o
:s 1:!
<J)
<J)

o
o a...
8 400
en '"Q)

~ o 6J 0.6
,.<:;
(/)
Q)
o

200
!
0.4

o Experimental data

- - - Best £it line

o 0.2

o 100 200 300 400 500 10 20 30 40 50 60

Shear Rate (S·l) Time (seconds)
Fig. 9-Class H cement with 1% bentonite measured with a rotational viscometer. Fig. 11-Creep test on a class H cement with 1% bentonite.