size distribution is less clear.

At densities above about bond to compressive strength is higher for foamed

1.0 g/cm3, foams with broad bubble-size distributions have cements and increases with nitrogen concentration.
lower permeabilities. At densities below about 1.0 g/cm3, These data are presented in Table 7-19.
foams with a narrow bubble-size distribution have
lower permeabilities. Thickening time
Chekiri (1978) reported that perforation of foamed Among the tests performed on foamed cement, thicken-
cements with qualities above 40% tends to cause exces- ing time is the most difficult to perform and the least
sive fracturing. As a rule, the permeability also increases conclusive. To be valid, this test should be performed
dramatically when the quality exceeds 40%; however, under simulated downhole conditions, and the foam
this depends on the additives, type of foamer, and curing should be mixed in a manner comparable to what occurs
conditions (Aldrich and Mitchell, 1975; Smith et al., on location. Thus, ideally, the slurry should be prepared
1984). in a pressurized mixer and transferred under pressure to
the pressurized consistometer. The thickening time test
Mechanical properties involves measuring the evolution of slurry viscosity
Foamed cement has a lower Young’s modulus than (Appendix B). Because of the particular rheological
conventional cements (Deeg et al., 1999). To achieve a behavior of foam, the shear field in the consistometer is
lower Young’s modulus with conventional cements, one not uniform. A large part of the foam remains static,
must add large amounts of water, resulting in lower com- while the small amount that is sheared is destabilized.
pressive strength. With foamed cement, the impact on Calorimetry experiments performed under static
compressive strength is lower. Cements with lower conditions at atmospheric pressure showed that the
Young’s moduli are less susceptible to failure when foam quality does not influence the hydration kinetics
exposed to the common mechanical stresses associated (de Rozières and Ferrière, 1991) (Fig. 7-24). A calori-
with well operations. A detailed discussion of the metric thermogram is not equivalent to the thickening
mechanical properties of well cements is presented in time. These results only demonstrate that the cement
Chapter 8. hydration process is not affected by the presence of gas
in the system.
Shear bond Instead of testing foamed systems, a common proce-
Davies et al. (1981) reported that foamed cement can dure is to measure the thickening time of the base slurry
undergo a bulk expansion before setting. In some situa- containing the additives, surfactants, and stabilizers.
tions, this can result in improved bonding (Slaton, This method gives a reasonable estimate of the working
1981). This hypothesis is supported by indirect evidence time for the foamed slurry (Davies et al., 1981; McElfresh
from the improved bond logs obtained from wells and Boncan, 1982).
cemented with foamed cement and can be explained as
an effect of pressure maintenance by the compressed gas Fluid loss
in the cement. As the cement loses hydrostatic pressure Introducing a gas to a liquid medium significantly
during gelation, the gas pressure maintains tight contact reduces the rate at which the liquid will flow through
between the cement and the casing or formation. porous media (Anderson, 1975). De Rozières and
Smith et al. (1984) reported that foamed cement at Ferrière (1991) evaluated foamed cements, with and
7.9 lbm/gal [948 kg/m3] develops higher shear-bond without fluid-loss additives, and found that the fluid-loss
strength than a 12% bentonite cement at 12.7 lbm/gal rates were lower when gas was present (Fig. 7-25).
[1,520 kg/m3]. They also found that the ratio of shear

Table 7-19. Compressive Strength and Shear-Bond Strength of Conventional and Foamed Cements†, ‡
Composition Density Compressive Strength Shear Bond Strength Ratio of Shear to
(lbm/gal [g/cm3]) (psi [MPa]) (psi [MPa]) Compressive Strength (%)
Class G 15.8 [1.90] 4,200 [29.0] 403 [2.8] 9.6

Class G + 12% bentonite 12.7 [1.53] 772 [5.0] 70 [0.5] 9.5

40% gas 9.5 [1.14] 873 [6.0] 118 [0.8] 13.5

50% gas 7.9 [0.95] 571 [3.9] 97 [0.7] 17.1

† From Smith et al. (1984). Reprinted with permission of World Oil.
‡ Cured 24 hr at 80°F (27°C), then 24 hr at 176°F (80°C)

260 Well Cementing

 56 200
Base slurry
52 Foam quality = 18%
Foam quality = 27% 160
48 Foam quality = 50%
Foam quality = 71%
44 120
Temperature Fluid-loss
(°C) 40 volume
(mL) 80
36
40 Base slurry
32 Foam 20%
Foam 37%
28 Foam 50%
0
0 10 20 30 40 50 60 70 80
0 5 10 15 20 25 30
Time (hr)
Time (min)
Fig. 7-24. Effect of foam quality on setting kinetics of foamed
Fig. 7-25. Effect of foam quality on fluid-loss behavior of foamed
cements (from de Rozières and Ferrière, 1991). Reprinted with per-
cements (from de Rozières and Ferrière, 1991). Reprinted with per-
mission of SPE.
mission of SPE.

Thermal and electrical conductivity Table 7-20. Resistivity of Foamed and Conventional
Short et al. (1961) reported that foams have lower Cements†
thermal conductivity, because of the presence of gas Cement type Density Specific Resistivity‡
voids and the lower amount of solids. Nelson (1986) (lbm/gal [g/cm3]) (ohm-cm)
reported that the thermal conductivity of cement Conventional 15.7 [1.88] 1.25 × 104
systems is roughly proportional to slurry density, regard-
less of whether the cement was foamed. These data are Foamed 9.3 [1.11] 1.4 × 10 4
presented in Fig. 7-26. † From Smith et al. (1984). Reprinted with permission from World Oil.
Studies of the resistivity of foamed cement indicate ‡ ASTM D-257

that the electrical conductivity is similar to that of

conventional cements (Smith et al., 1984) (Table 7-20).

Cement density (lbm/gal)

0.9 5.8 6.7 7.5 8.3 9.1 10.0 10.8 11.6 12.5 13.3 14.1 15.0 15.8

0.8

0.7

Thermal 0.6
conductivity Conventional
0.5 low-density
⎛ BTU ⎞
⎜⎝ hr × ft × °F ⎟⎠ systems
0.4
Foamed cements and
0.3 microsphere systems

0.2

0.1
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
Cement density (g/cm3)

Fig. 7-26. Cement density/thermal conductivity relationship (from de Rozières and Ferrière, 1991). Reprinted with permission of SPE.

Chapter 7 Special Cement Systems 261

Rheology high fluid pressures and temperatures inside the casing
The rheological behavior of foams is unlike that of other during testing, perforating, hydraulic fracturing, or fluid
fluids (Chapter 4). The differences arise from many production (Bosma et al., 2000). Another stress condi-
factors. Foams are compressible fluids; they are hetero- tion results from exceedingly high pressures that occur
geneous and have variable properties under shear. Foams inside the cement sheath because of thermal expansion
are dynamically unstable, shear-history-dependent fluids of the interstitial fluid. A third condition involves
in which the bubble structure is continuously destroyed tectonic movement of the formation. When such stresses
and rebuilt. For these reasons, rotational viscometers are exerted on set cement in the wellbore, the set
with a fixed amount of sample are unsuitable (Heller and cement can fail in the form of radial circumferential
Kuntamukkula, 1987). The rotational shear disturbs the cracking of the cement matrix or by a breakdown of
bubble network, often resulting in foam collapse. the cement/casing or cement/formation bonds. Such
Continuous flow-tube viscometers are more suitable failures compromise zonal isolation and can lead to
for testing foams, despite the fact that foams are com- severe well problems. Thus, the well cementing industry
pressible, are non-Newtonian, and will never attain a has recognized the need for highly resilient and flexible
steady state (Reidenbach et al., 1986; Harris and well-cement compositions that can withstand the
Reidenbach, 1987; Mueller et al., 1990). Viscosity, den- stresses outlined above.
sity, and flow rate will vary continuously as the system As explained in Chapter 8, a detailed analysis of the
pressure changes along the tube. As a result, the mechanisms leading to failure of the cement sheath
equations to calculate shear stress and shear rate at showed that the risk of rupture is directly linked to the
the wall require corrections (Harris and Reidenbach, tensile strength of the set cement and is attenuated
1987). In addition, low-pressure viscosity measurements when the ratio of tensile strength to Young’s modulus is
of foamed cements may not be representative of field increased. Young’s modulus characterizes the flexibility
conditions. Such problems are minimized if tests are of a material. Thus, to increase the tensile strength to
performed at the pressures encountered in the well and Young’s modulus ratio, the set cement should have a low
at low differential pressure. Young’s modulus (Thiercelin et al., 1997).
While studying foamed fracturing fluids, Harris and
Reidenbach (1987) found that the bubbles tend to reach
a small, uniform size at high energy levels. At low energy
7-11.1 Slurry density reduction
levels, the bubbles coalesce and drain, forming large Conventional slurry density reduction, using extenders
nonuniform bubbles. Harris and Reidenbach also that accommodate additional water (e.g., sodium silicate
observed that the frictional pressure drop for foamed and bentonite), increases the flexibility of the set cement
fluids can increase twofold from the time a foam is (Table 7-21). Although the flexibility is increased, the
created until equilibrium is reached. effects on compressive strength and permeability are
To the best of the authors’ knowledge, no routine rhe- detrimental.
ological measurement is made on foamed cement slur-
ries. However, visual observations of foamed cements 7-11.2 Flexible particles
flowing in plastic tubes have shown that, at low flow
rates, the foams flow as a rigid plug, moving on a thin Another possibility to modify the elasticity of the set
film of water next to the wall (Princen, 1982). cement is to incorporate flexible particles in the slurry
design. In the construction industry, incorporating
ground particles from rubber tires in concrete improves
toughness, durability, and resilience (Eldin et al., 1993).
7-11 Flexible cements The rubber aggregates are used in highway construction
As discussed in Chapter 8, the mechanical properties for antishock properties, as a sound-absorbing material in
of well cements have become a topic of considerable antinoise walls, and in earthquake-resistant buildings.
interest. Today, in addition to the traditional unconfined- Adding ground rubber particles to well cements has
compressive-strength measurement, tensile strength, been practiced for more than 30 years as a method to
Young’s modulus, and Poisson’s ratio are frequently improve resistance to shocks during perforating (F.E.
considered during the cement-design process. The Hook, unpublished data, 1971). More recently, improve-
methods employed to measure these parameters are ments have been made with optimized rubber-particle-
discussed in detail in Appendix B. size distributions. Karimov et al. (1985) proposed using
During the life of a well, the set cement can fail particles in the 4/20 mesh range to improve impact
because of shear and compressional stresses. There are and bending strength. Brothers and de Blanc (1998) pro-
several stress conditions associated with cement-sheath posed adding ground-rubber-tire particles in the 10/20
failures. One such condition is the result of relatively mesh to 20/30 mesh size range. The set cements

262 Well Cementing

Table 7-21. Mechanical Properties of Conventionally Extended Set Cements†
Slurry Density Tensile Strength, TS Young’s Modulus, E TS/E Compressive Strength Water Permeability, k
(lbm/gal [kg/m3]) (psi [MPa]) (psi [MPa]) (× 1,000) (psi [MPa]) (mD)
15.8 [1,897] 605 [4.23] 944,000 [6,600] 0.64 51,230 [36.6] 0.001

14.0 [1,680] 479 [3.35] 538,000 [3,759] 0.90 3,270 [22.9] 0.008

12.0 [1,440] 86 [0.60] 72,100 [504] 1.19 450 [3.21] 0.138

† After Le Roy-Delage et al. (2000). Reprinted with permission of SPE.

exhibit improved elasticity, ductility, and expansion Table 7-22. Mechanical Properties of Set Cements
properties over a wide density range—14.3 to 19.0 lbm/ Containing Rubber Particles†
gal [1,720 to 2,280 kg/m3]. Le Roy-Delage et al. (2000) Slurry Tensile Young’s TS/E Compressive
described cement systems containing from 30% to 100% Density Strength, Modulus, E (× 1,000) Strength
BWOC rubber particles in the 40/60 mesh range. As (lbm/gal TS (psi [MPa]) (psi [MPa])
shown in Table 7-22, the resulting set cements are [kg/m3]) (psi [MPa])
more flexible. 12.0 [1,440] 93 [0.65] 73,900 [516.7] 1.26 412 [2.88]
EPS cement systems (Section 7-9) have been formu-
13.1 [1,570] 140 [0.99] 137,800 [963.6] 1.04 819 [5.73]
lated using flexible particles as one of the components
(Le Roy-Delage et al., 2000). The particles have the 13.6 [1,630] 207 [1.45] 189,000 [1,320] 1.11 1,130 [7.93]
following characteristics:
14.1 [1,690] 237 [1.66] 240,000 [1,678] 1.01 1,782 [12.46]
■ Particle size: ≤500 μm

■ Young’s modulus: <5,000 MPa, preferably <2,000 MPa 15.2 [1,820] 390 [2.73] 460,900 [3,223] 0.86 2,890 [20.21]
■ Poisson’s ratio: >0.3. 16.4 [1,970] 323 [2.26] 864,000 [6,042] 0.76 3,934 [27.51]
† After Le Roy-Delage et al. (2000). Reprinted with permission of SPE.
Thermoplastics, like polyamide, polypropylene, and
polyethylene, or polymers, like styrene divinylbenzene
or styrene butadiene, are compatible with these perfor-
mance criteria. Because the specific gravities of these
materials fall between 0.9 and 1.2, they can also reduce the
cement system density. Table 7-23 presents mechanical-
properties data from set cements containing various
amounts of flexible particles.

Table 7-23. Mechanical Properties of EPS Set Cements Containing Flexible Particles†
Flexible Particle Slurry Density Tensile Strength, Young’s Modulus, E TS/E Compressive Strength,
(volume %) (lbm/gal [kg/m3]) TS (psi [MPa]) (psi [MPa]) (× 1,000) CS (psi [MPa])
Styrene divinylbenzene (25) 14.0 [1,680] 365 [2.52] 521,400 [3,595] 0.72 4,860 [33.5]

Styrene divinylbenzene (30) 12.1 [1,450] 160 [1.10] 194,200 [1,339] 0.84 1,930 [13.3]

Polyamide (25) 14.0 [1,680] 406 [2.80] 374,200 [2,580] 1.09 4,050 [27.9]

Polypropylene (19) 14.0 [1,680] 329 [2.27] 426,600 [2,941] 0.78 3,130 [21.6]

Polypropylene (24) 13.7 [1,640] 381 [2.63] 438,000 [3,020] 0.88 3,810 [26.3]

Styrene butadiene (25) 14.2 [1,700] 299 [2.06] 302,400 [2,085] 1.00 2,100 [14.5]

Polyethylene (25) 13.6 [1,630] 306 [2.11] 299,800 [2,067] 1.04 3,320 [22.9]
† After Le Roy-Delage et al. (2000). Reprinted with permission of SPE.

Chapter 7 Special Cement Systems 263

7-11.3 Elastomeric composites 1,600
Neat slurry
Onan et al. (1993) described a composite material based 1,400 Microribbon slurry
on styrene-butadiene rubber latex. Various fillers, such
as carbon black, calcium carbonate, silica, and quick- 1,200
setting Portland cement can be added to adjust the
mechanical properties of the composite. This system has 1,000
been used to cement CO2-injection wells and multilat- Force 800
eral junctions (Xenaxis et al., 1997). Typical mechanical (N)
performance data are shown in Table 7-24. 600

400
7-11.4 Fibers
Adding fibers or ribbons to a cement matrix also 200
improves flexural strength. Nylon fibers have been used 0
for many years for this purpose (Chapter 3). More 0 0.2 0.4 0.6 0.8 1
recently, Le Roy-Delage et al. (2000) and Baret et al. Deflection (mm)
(2002) described the addition of metallic microribbons
to improve impact resistance, toughness, and tensile Fig. 7-27. Typical load-deflection curve comparing neat and micro-
strength. The principal applications of this system are ribbon cement systems.
kickoff plugs and multilateral junctions.
The performance of the metallic ribbon system is
illustrated in Fig. 7-27, which shows a load-deflection
curve recorded during a flexural test. The plot shows the
amount of force required to bend or deflect the set-
cement sample a given distance. The results show that
a neat slurry fails completely after being deflected
less than 0.1 mm. The microribbon slurry was able to
bear a load after a deflection of nearly 1 mm.

Table 7-24. Mechanical Properties of Elastomeric Composite Systems†

Elastomeric Composite Slurry Density Tensile Strength, Young’s Modulus, TS/E Compressive Strength, Air Permeability,
(%BWOC) (lbm/gal) TS (psi) E (psi) (× 1,000) CS (MPa) k (mD)
Neat 15.7 200 700 0.28 –‡ 0.383

67% thermatomic carbon 14.0 720 4,100 0.18 – 0.014

80% CaCO3 12.0 910 5,200 0.18 – 0.029

130% CaCO3 14.0 1,080 10,000 0.11 – 0.007

300% silica 12.1 110 5,300 0.02 105 0.012

150% silica flour, 14.0 210 8,200 0.03 245 0.001

25% quick-setting cement

67% quick-setting cement 14.6 180 9,100 0.02 505 –

77% quick-setting cement 14.0 470 23,000 0.02 985 –

113% quick-setting cement 14.0 215 200,000 <0.01 732 –

150% Class A cement 14.7 150 25,000 0.01 795 –

† After Onan et al. (1993). Reprinted with permission of SPE.
‡ Not measured

264 Well Cementing

7-12 Microfine cements 6
Microfine cements are composed of very small particles
(generally 4 to 15 μm). Therefore, the surface area is 5
very large (500 to greater than 1,000 m2/kg). The most 140°F
4 160°F
common microfine cements are very fine Portland
cements (Ewert et al., 1991; Bensted and Barnes, 2002); 116°F
however, slag cements may also be incorporated (Clarke Retarder 3
(% BWOC) 155°F
and McNally, 1993; Section 7-8). The advantage of such
cements is their improved ability to penetrate and flow 2
through tight spaces and porous media. The Portland 102°F
cement–based microfine cements generally behave sim- 1
ilarly to their conventional counterparts; however, owing
to their higher reactivity, additional gypsum or retarders 0
may be necessary to achieve predictable rheological per- 0 100 200 300 400
formance and setting. Thickening time (min)
The principal uses for microfine cements include the
following. Fig. 7-28. Thickening time of acid-soluble cement versus borate
retarder concentration. Reprinted with permission of SPE.
■ Squeeze cementing (Heathman and East, 1992)
(Chapter 14; Section 7-8.2.1)
■ Sealing casing leaks (Meek and Harris, 1991;
cements can be pumped similarly to conventional well
Lizak et al., 1992) (Chapter 14) cements. Once placed, the set cement can be readily
■ Cementing permafrost zones (Section 7-4) removed by acidizing with HCl.
■ Cementing shallow strings in deepwater wells In addition to solving lost circulation problems, acid-
soluble cements have been used as diverters to provide
More detailed discussions of the applications of micro- temporary zonal isolation and as kickoff plugs in weak
fine cements are found in the sections indicated above. formations (Chapter 14).

7-13 Acid-soluble cements 7-14 Chemically bonded phosphate

As discussed in Chapter 6, lost circulation during drilling ceramics
and cementing is a common problem. Most common
remedies, such as lost-circulation materials (e.g., flakes Chemically bonded phosphate ceramics (CBPCs) are
and fibers) and thixotropic cements, remain in the thief binders that fall between sintered ceramics (e.g., pot-
zone permanently. This is not a problem unless the thief tery, porcelain) and chemically bonded systems (e.g.,
zone is also the producing zone. Such situations include Portland cement). CBPCs are formed by acid-base reac-
gas-injection and gas-storage wells. Therefore, there is a tions between an acid phosphate (e.g., Mg, Ca, or Al) and
need for lost circulation materials that can easily be a metal oxide (e.g., MgO, CaO, or ZnO2) (Jeong and
removed after well construction is complete. Wagh, 2003).
Such a temporary system, based on magnesium oxy- Calcium aluminate–based CBPCs are being used
chloride (or Sorel) cement, was developed by Sweatman as cements for geothermal wells (Chapter 10).
and Scoggins (1990) and Vinson et al. (1992). Magnesium Magnesium potassium phosphates, originally devel-
oxychloride is made by mixing powdered magnesium oped to stabilize and encapsulate radioactive and haz-
oxide with a concentrated solution of magnesium chloride. ardous waste streams, have also found use as fast-setting
The principal binder phases are Mg3(OH)5Cl • 4H2O and well cements (Wagh and Brown, 1999) The cement
Mg2(OH)3Cl • 4H2O. The set-cement is very strong; how- is prepared by calcining MgO at 2,372°F [1,300°C].
ever, it readily dissolves when exposed to acid. The calcined MgO is then reacted with a KH2PO4 solu-
A modified Sorel cement for oilfield applications tion to form the binder.
consists of magnesium and calcium oxides, carbonates, MgO + KH2PO 4 + 5H2 O ⎯→
⎯ MgKPO 4 + 6H2 O
and sulfates. When added to seawater or a chloride brine,
magnesium oxychloride hydrates form. The thickening (7-20)
behavior can be modified by common retarders such as
borax (Fig. 7-28) and accelerators such as calcium chlo-
ride (Bensted and Barnes, 2002); therefore, such

Chapter 7 Special Cement Systems 265

 When the MgO is added to the KH2PO4 solution, the liquid cement premixes that can be stored in the liquid
resulting paste sets within 1 hr at ambient temperature. state almost indefinitely and activated at the wellsite
The system can be retarded by adding boric acid. The during the cement job. This technique can lead to
boric acid reacts with the acid phosphate and forms a tem- improvements in job efficiency and service quality in
porary coating of magnesium boron phosphate on the sur- both primary and remedial cementing operations (Rae
face of the MgO particles. Typical performance of a system and Johnston, 1996a and 1996b).
containing boric acid retarder is shown in Fig. 7-29. The storable slurries can be made from Portland
The performance is notable in that the ultimate com- cement, slag cement, or other blended cements, and
pressive strength is much higher than that of set can also contain other cement additives. Water is the
Portland cement. carrier fluid. A set retarder, such as a hydroxycarboxylic
acid, is necessary to prevent the setting of the slurry
during storage. To prevent settling of the cement solids
7-15 Storable cement slurries during storage, suspending agents such as polymers and
clays are added. At the wellsite, the storable slurry is
As described in Chapter 11, cement-slurry mixing in the
pumped to the cement mixer, where an activator is
well cementing industry essentially consists of combining
added to counteract the retarder and reinitiate cement
fine cement powder with mix water. The process
hydration. The preferred activator is sodium silicate.
involves cutting sacks of cement or transferring bulk
cement pneumatically to the mixing unit. Many parameters
must be controlled at the mixing unit, including slurry den-
sity, mixing energy, and mixing time. Density is arguably 7-16 Summary
the most difficult parameter to control. Key cement-slurry Table 7-25 summarizes the cement systems presented in
properties such as thickening time, rheology, free water, this chapter, indicating their principal uses, chemical
and compressive strength depend upon density. compositions, and mechanisms of action.
A cementing operation can be simplified by elimi-
nating the handling of dry powders. This concept was
realized by Rae and Johnston (1995), who introduced

160 13,000
140 12,000
120
11,000
100
Setting Compressive 10,000
time 80 strength
(min) (psi) 9,000
60
8,000
40
20 7,000

0 6,000
30 40 50 60 70 80 90 0 10 20 30 40 50
Temperature (°C) Time (hr)

Fig. 7-29. Setting characteristics of a magnesium potassium phosphate cement system (Wagh and Brown, 1999).

266 Well Cementing

Table 7-25. Cement System Summary
Cement System Principal Uses Chemical Compositions Mechanisms of Action
Thixotropic cements Lost circulation prevention Portland cements containing one
Slurry fallback prevention of the following additives
Gas migration prevention
Bentonite Increased gel strength

Calcium sulfate hemihydrate Formation of ettringite crystals

Aluminum sulfate/iron sulfate

Crosslinked cellulose polymer Increased gel strength

Expansive cements Improved cement/casing and Commercial expanding cements or Portland Formation of ettringite crystals
cement/formation bond cement containing calcium sulfate hemihydrate

Salt cements Internal pressure exerted by

crystallization of salt in pores

Portland cement containing aluminum powder Generation of hydrogen gas in situ

Portland cement containing Conversion of MgO (periclase) to

calcined magnesium oxide Mg(OH)2 (brucite)

Freeze-protected cements Cementing across permafrost zones Calcium aluminate cement Rapid strength development at
low temperatures

Gypsum-Portland cement blends Rapid strength development at low

temperatures; lower heat of hydration
than calcium aluminate

Ultrafine Portland cement High surface area increases hydration rate

Salt cement systems Cementing across salt zones Portland cements containing sodium Systems do not disturb salt-bearing
or sensitive formations chloride or potassium chloride at formations
concentrations up to saturation Systems do not disturb sensitive clays

Latex-modified cement Improved cement/casing and Portland cements containing one

systems cement/formation bond of the following latexes
Improved fluid-loss control
Gas migration prevention Polyvinylidene chloride Film formation on bonding surfaces

Polyvinyl acetate Latex particles plug small pores in

Styrene-butadiene copolymer cement matrix

Cements for corrosive Cementing chemical waste Epoxy-based cement systems Chemically inert to strong acids and
environments disposal wells Elastomeric composites bases

Cementing CO2 injection wells Pozzolanic cement systems Reduced cement-matrix permeability,
BFS systems improved chemical resistance

BFS systems Alternative to or supplement for BFS + activator (e.g., sodium hydroxide Formation of C-S-H phase, with extensive
Portland cement or Portland cement) incorporation of Al, Mg, Fe, and sulfate
Conversion of drilling fluid to cement BFS + activator mixed with drilling fluid in structure

EPS cement systems Systems with improved Cement blends with multimodal Particle-size distribution minimizes mix-water
placement and set-cement particle-size distributions concentration required to prepare pumpable
properties over a wide slurry- slurry; set cement is less permeable than
density range conventional systems

Ultralow-density cement Cementing across formations Cements containing glass or ceramic Low slurry density reduces hydrostatic
systems with low fracture gradients or microspheres pressure in wellbore and prevents
that are vuggy or cavernous Foamed cements formation breakdown

Flexible cement systems Improved resistance to stresses Cements containing flexible particles Flexible particles decrease Young’s modulus
induced by perforating, hydraulic and increase Poisson’s ratio
fracturing, and tectonic movement

Cements containing nylon or metallic Fibers act as reinforcement and improve

fibers flexural strength and toughness

Elastomeric composites Polymer matrix filler is flexible

Microfine cement systems Squeeze cementing, sealing casing leaks Portland or BFS cements with small particle Smaller particles more readily enter cracks
size (4 to 15 μm) and high surface area and formation pores
(500 to 1,000 m2/kg)

Acid-soluble cement Temporary solution for lost circulation Magnesium oxychloride (Sorel) cements Principal binder phase is readily removed
systems by contact with a strong acid (e.g., HCl)

Chemically bonded Fast-setting cements that develop high Magnesium potassium phosphate Acid-base reaction between an acid
phosphate ceramics compressive strength phosphate and a metal oxide

Storable cement slurries Eliminates handling of dry powders Portland, BFS or blended cements During cement job, concentrated slurry
during cementing operations slurried in aqueous solution containing is diluted with mix water containing
a strong cement retarder activator (e.g., sodium silicate)

Chapter 7 Special Cement Systems 267

7-17 Acronym list
API American Petroleum Institute
BFS Blast-furnace slag
BHCT Bottomhole circulating temperature
BHST Bottomhole static temperature
BWOW By weight of water
CBPC Chemically bonded phosphate ceramics
CS Compressive strength
EPS Engineered particle size
ISO International Organization for
Standardization
PVF Packing volume fraction
SG Specific gravity
SVF Solid volume fraction
TS Tensile strength

268 Well Cementing

8-1 Introduction
Until recently, the well cementing industry focused on
one principal mechanical property of set cement—
Mechanical Properties
of Well Cements
Marc Thiercelin—Schlumberger 8
field stress, but also from any change of downhole tem-
perature or pore pressure. Cement tensile strength,
elasticity, and ductility are mechanical parameters that
unconfined uniaxial compressive strength—to qualify a may be more important to long-term durability than an
cement design. The uniaxial compressive strength is arbitrary measure of cube compressive strength
determined by a cube-crushing test (Appendix B) and is (Thiercelin et al., 1997; Bosma et al., 1999; di Lullo and
used to estimate the ability of set cement to support Rae, 2000; Ravi et al., 2002).
casing and survive the perforation process. When com- Knowledge of the appropriate cement mechanical
bined with water- or air-permeability measurements properties allows one to characterize the cement defor-
(Appendix B), one can also estimate the cement’s abil- mation under applied downhole stresses and to predict
ity to provide zonal isolation and resist attack from for- whether the cement will be able to survive the stresses.
mation fluids. In this chapter, basic rock-mechanics concepts are pre-
Indirect ultrasonic techniques using tools like the sented in the context of well cements, followed by a dis-
ultrasonic cement analyzer (UCA) (Appendix B) allow cussion of how these concepts can be applied to design
one to monitor the evolution of strength with time. With cement systems that are appropriate for the anticipated
the UCA, one can determine the curing time required downhole environment.
for a cement system to attain a given compressive
strength.
In recent years, the well cementing industry has 8-2 Basic concepts
begun to concentrate on set cements’ ability to provide The basic concepts of stress and strain are presented in
zonal isolation throughout the lifetime of the well. This this section. These concepts have been presented previ-
was triggered by the observation that, even in situations ously in the context of rock mechanics analysis
in which the cement was properly placed and provided (Thiercelin and Roegiers, 2000). For further details, the
a good initial hydraulic seal, zonal isolation disappeared reader is referred to the classical works by Love (1927),
with time (Goodwin and Crook, 1992; Jackson and Timoshenko and Goodier (1970), and Muskhelishvili
Murphey, 1993). Zonal isolation loss has been attributed (1977).
to several causes. Stress and strain have distinct meanings to the engi-
■ Gas-migration problems that were not initially neer. For example, a normal stress, σ, can be viewed as
detected a normal force (to a surface) per unit area while linear
■ Loss of cement-bond log response with time
strain, ε, is the resulting deformation.
■ Fracturing the wrong zone during a stimulation
treatment 8-2.1 Stress
■ Extreme downhole temperature or pressure changes Consider a randomly oriented plane of area Δ A, cen-
■ Chemical attack tered on a Point P within a body. When a resultant force
Δ –F acts on the body (Fig. 8-1), the stress vector σ
– at
■ Pressure migration to shallower zones
that point is defined as
Laboratory and modeling studies have shown that the
principal cause of cement-sheath damage is the stresses ⎛ ΔF⎞
σ = lim ⎜ ⎟. (8-1)
induced by varying downhole conditions (Parcevaux and Δ A→0 ⎝ Δ A ⎠
Sault, 1984; Jutten et al., 1989; Goodwin and Crook,
1992; Jackson and Murphey, 1993; Bol et al., 1997; Therefore, this quantity is expressed as a force per
Thiercelin et al., 1997; Bosma et al., 1999). The stresses unit area. In classical solid mechanics, the tensile stress
arise from not only the wellbore pressure or the rock far- is taken to be positive by convention. However, in rock

Well Cementing ■ Chapter 8 Mechanical Properties of Well Cements 269

 ΔF
y
σ
n
σn n
τ
τxy

θ
σx

ΔA τxy x
σy
Fig. 8-1. Stress vector definition.

Fig. 8-2. State of stress in two dimensions.

mechanics, compressive stress is taken to be positive
because geologic forces are usually compressive in
nature. Because either convention can be used for the shear-stress components vanish. These are called
mechanical studies of well cements, it is important to the principal planes. The normal stresses associated
clearly state the convention being used. In this chapter, with these planes are called the principal stresses. In
tensile stresses are taken to be positive. two dimensions, expressions for the principal stresses
The resultant stress, σ, can be decomposed into a can be found by setting τ = 0 in Eq. 8-3 or, because they
normal component (σ n) and a shear component (τ). are the minimum and maximum values of the normal
The shear component tends to “shear” the material in stresses, by taking the derivative of Eq. 8-2 with respect
the plane Δ A. The normal component of stress acts to the angle θ and setting it equal to zero. In either case,
perpendicular to Δ A, and the shear component acts in one obtains the following expression for the value of θ at
the plane of Δ A. which the shear stress vanishes
An infinite number of planes can be drawn through a
given point, varying the values of σ n and τ. The stress 1 ⎛ 2τ xy ⎞
θ = arctan ⎜ ⎟. (8-4)
condition, therefore, depends on the inclination of the 2 ⎝ σy −σx ⎠
plane. Consequently, for a complete description of a
stress, one must not only specify its magnitude and The two principal stress components, σ1 and σ2, are
direction, but also the direction of the surface upon
1/ 2

( ) ( ) ( )
which it acts. 1 ⎡ 2 1 2⎤
In a two-dimensional situation, if σx, σy , and τxy are σ 1 = σ x + σ y + ⎢ τ xy + σ x − σ y ⎥ (8-5a)
known (Fig. 8-2), the stress state on any plane oriented 2 ⎣ 4 ⎦
at an angle θ from σx can be expressed as follows:
and
( 2
)( )(
σ n = σ x cos θ − 2τ xy sin θ + σ y sin θ 2
) (8-2) 1/ 2

( ⎡
) ( ) ( )
1 2 1 2⎤
σ 2 = σ x + σ y − ⎢ τ xy + σ x − σ y ⎥ (8-5b)
and
2 ⎣ 4 ⎦
where θ is given by Eq. 8-4.
⎡1
( )⎤
(
τ = ⎢ σ y − σ x sin 2θ ⎥ + τ xy cos 2θ .
⎣2 ⎦
) (8-3) If one generalizes this concept to three dimensions,
it can be shown that six independent components of the
stress (three normal and three shear components) are
These expressions are obtained by writing equilib- needed to define the stress unambiguously. The stress
rium equations of forces along the σ n and τ directions, vector for any direction of Δ A can generally be found
respectively (see Fig. 8-2). Note also that the moment by writing equilibrium-of-force equations in various
equilibrium implies that τxy is equal to τyx. There are directions. There are three principal planes for which
always two perpendicular orientations of ΔA for which the shear-stress components vanish and three principal

270 Well Cementing

stresses exist. Quantities that show this kind of behavior borhood of a point. For example, compatibility equations
are known as second-order tensors, and the state of ensure that the strained body remains continuous and
stress in a body is often referred to as the stress tensor. that no cracks or material overlaps will occur.

8-2.2 Strain
8-3. Cement behavior
When a body is subjected to a stress field, the relative
positions of points within it are altered. The body A cement sample, like any material, deforms when sub-
deforms. If the new positions of the points do not corre- jected to stress. Determining a relationship between stress
spond to a translation and/or a rotation (i.e., by rigid-body and strain is an important aspect of solid mechanics. This
motion), the body is strained. This strain along an arbi- relationship is called the constitutive equation of the
trary direction can be decomposed into two components: material under consideration, and various theories have
been developed to describe it in a simplified way. The sim-
■ an elongation, defined as
plest one is the theory of elasticity, which assumes a
unique relationship between stress and strain (and that
L* − L the behavior is reversible). This theory is usually sufficient
ε = lim (8-6)
L→0 L to analyze cement failure in tension or in compression at
ambient conditions. Other theories, such as the theory of
■ and a shear strain, defined as elastoplasticity, have been developed to take into account
nonreversible behaviors that are observed in materials
γ = tan Ψ ,( ) (8-7) before failure. Significant nonreversible behavior is
observed in cements subjected to confining pressure.
where Ψ is the change of angle between the two direc-
tions that, before straining, were perpendicular (Fig. 8-3).
Consequently, strain (being either a ratio of lengths 8-3.1 Stress-strain curve
or a change of angle) is dimensionless. If one assumes Fig. 8-4 presents a typical stress-strain relationship for
that the stresses are positive in traction, a positive lon- cement. The test is carried out under constant confining
gitudinal strain, ε, corresponds to an increase in pressure and constant axial strain rate (Appendix B).
length. Just as in the case of stresses, principal strains The sample is protected from the confining fluid by an
can be defined as longitudinal strain components acting impermeable flexible jacket. Measurements include the
on planes in which the shear strains have vanished. axial stress, the axial strain, and the radial strain. When
The analogy between stress and strain analyses is a confining pressure is applied to the sample, the origin
not completely valid; equilibrium equations must be of the stress-strain plots is usually translated to remove
satisfied by the stresses and compatibility equations by the influence of the hydrostatic loading on the stress
the strains. These relationships place some restrictions and strain (i.e., the axial stress is actually the differ-
on the local variation of stress and strain in the neigh- ential σa – pcon), where σa is axial stress and pcon is the
confining pressure.

O x
F

P A Strength
C
B
P′ Axial stress
pcon (compression)
A′ E

y A

Ψ Radial strain O Axial strain

B (expansion) (shortening)

Fig. 8-4. Typical stress strain-curve during compression. The axial

B′ stress is computed from the load. The symbol pcon denotes the confin-
ing pressure and is maintained constant during the test. Compressive
Fig. 8-3. Shear strain component. The points P, A, and B are adjacent stresses are taken as positive in this figure for practicality.
points. P’, A’, and B’ represent new positions after a displacement.

Chapter 8 Mechanical Properties of Well Cements 271

 During the initial stages of loading, from Points O when a compressive (negative) force, F, is applied to its
to A, the cement becomes stiffer. This nonlinear regime ends (Fig. 8-5). According to the definitions given in the
is probably caused by the closing of pores, although previous section, the axial stress applied to the sample is
experimental artifacts, such as the misalignment of the
sample surfaces with the loading platen of the machine, F 4F
σa = = , (8-8)
could also produce such an effect. A π d2
As the load increases, the stress/strain curve
becomes linear (from A to B). This is the portion of the where A is the surface area of the end of the sample. The
stress-strain curve in which the behavior of the cement axial strain is:
is nearly elastic. If unloading occurs in this region, the
strain returns nearly to zero, often along a different L* − L
εa = , (8-9)
path. When the strain follows different paths during L
loading and unloading, it is called hysteresis and indi-
cates that some energy is dissipated during this cycle. where tension is assumed to be positive. Linear elastic-
When the cement specimen is loaded beyond Point B, ity assumes a linear and unique relationship between
the yield point, the stress-strain behavior becomes stress and strain, in which strain returns to zero when
nonlinear and large deformations will eventually occur. the material is unloaded. In the case of a uniaxial test,
If the sample is unloaded in this region, permanent this means that:
strains at zero stress are observed. The Point C is the
σa = E × εa. (8-10)
maximum load that the cement can sustain under a
given confining pressure. Cement failure (i.e., when the The coefficient of proportionality, E, is the Young’s
sample loses its integrity) occurs around this point. modulus. The higher the value of the Young’s modulus,
After this, the cement might start losing its strength the less the sample will shorten for a given load. Some
because of the development of cracks. This is called the examples of Young’s moduli are given in Table 8-1.
postfailure region. Some cements, especially those with
high porosity and/or high elasticity, may not exhibit a
maximum peak stress at high confining pressure, but
A F
will continue to carry increasing stress (i.e., continue
to harden) because of compaction. Under such test
conditions, these cements do not fail.
Another interesting cement characteristic is the vol- dL
umetric strain, defined as the volume change divided by
the original specimen volume. For the triaxial test, the L
volumetric strain is εa + 2 εr , where εa is the axial strain L*
and εr the radial strain. Nonelastic volumetric strain can
show an increase of volume (called dilatancy when
cracks are created) or a decrease in volume because of
F
compaction.
The three important regions are therefore: Fig. 8-5. Sample deformation in the axial direction.
■ the elastic region, in which deformation is reversible

■ the plastic region, in which permanent deformation

occurs Table 8-1. Young’s Moduli of Various Materials
■ the postfailure region, in which the cement has lost Material Elastic Modulus (psi [MPa])
its integrity. Aluminum 10 × 10 6 [70,000]
An important stress value is the cement strength, Copper 16 × 10 6 [110,000]
which is the peak strength measured during the test.
These various regions and critical stresses are presented Steel 30 × 10 6 [210,000]
in more detail below. Window glass 10 × 10 6 [70,000]

Oilwell cement 0.14–1.4 × 10 6 [1,000–10,000]

8-3.2 Linear elasticity
Polyethylene 14–200 × 10 3 [100–1,400]
To introduce the theory of linear elasticity, consider a
cylindrical sample of initial length L and diameter d. Rubber 0.6–11 × 10 3 [4–80]
The sample will shorten along the loading direction

272 Well Cementing

 When the specimen is compressed in one direction, it These stress-strain relations can be generalized to
will not only shorten along the loading direction, but the full three-dimensional space by
also expand in the lateral directions (Fig. 8-6). This
σx ν
effect is quantified by the introduction of another con-
stant, the Poisson’s ratio, ν, defined as the ratio of the
εx = − σ +σz ,
E E y
( )
radial strain to the axial one: σy ν
ε
εy = − σ +σz ,
E E x
( )
ν=− r , (8-11)
εa and
σz ν
with
∗ ∗
εz = −
E E x
(
σ +σy . )
d −d r −r
εr = = , (8-12) 1 1 1
d r γ xy = τ xy ; γ yz = τ yz ; γ xz = τ xz , (8-13)
G G G
where r is the sample radius.
with the shear modulus, G, given by
E
G= . (8-14)
r
F
r* 2 (1 + ν )
In isotropic linear elasticity, only two elastic con-
stants are independent. For example, and as seen above,
the shear modulus G can be written as a function of E
and ν. Similarly, the bulk modulus, which is the coeffi-
L cient of proportionality between a pressure applied to
L* the sample and the volumetric strain, is
E
K= . (8-15)
F
3 (1 − 2ν )
Fig. 8-6. Sample deformation in the radial direction. Elasticity theory can be extended to nonlinear and
anisotropic materials. A nonlinear elastic material does
not have a linear relationship between stress and strain,
The negative sign is included to ensure that the but recovers all strain when unloaded. An anisotropic
Poisson’s ratio is a positive quantity in the most common material has properties that vary in different directions.
cases (i.e., when the axial strain and radial strain are of
different signs). A Poisson’s ratio of 0.5 means that the 8-3.3 Thermoelasticity
material is incompressible (rubber for example).
Conventional cements have a Poisson’s ratio of approxi- In a downhole environment, set cement will be subjected
mately 0.15. A foamed rubber or a sponge shows a very to temperature changes. For example, cement liberates
low value of Poisson’s ratio. heat during hydration, and the temperature will change as
In experiments to determine the Young’s modulus the set cement equilibrates to the surrounding formation
and Poisson’s ratio of cements, the sample must be sat- (i.e., bottomhole circulating temperature to bottomhole
urated with water, and the strain rate must be chosen static temperature).
appropriately. These precautions prevent the generation One of the most dramatic examples is the wellbore-
of capillary pressure and excessive pore pressures, temperature increase that occurs during steam injection
which would change the sample response. The measure- or geothermal well production, in which temperature
ment can be made at various pressures to determine the fluctuations can exceed 350°F [194°C]. If the wellbore
influence of confining pressure on the elastic response. temperature increases, heat will diffuse by conduction
through the casing then through the cement sheath and
the formation rock, and the temperature will progres-
sively increase in the wellbore region (Fig. 8-7).

Chapter 8 Mechanical Properties of Well Cements 273

 250 To calculate the temperature change with time, one
uses the conduction equation:
200

150 ∂2 T ∂2 T ∂2 T ρC ∂T
Temperature + + = × , (8-18)
(°C) ∂x 2
∂y2
∂z 2 K ∂t
100

50
where K is the thermal conductivity, ρ the density, and
C the specific heat. A temperature gradient will develop
0 and, as shown in Fig. 8-7, a heat front will move with
60 65 70 75 80 85 90 time.
Distance from well axis (mm)

Fig. 8-7. Temperature profile in the cement sheath at different 8-3.4 Poroelasticity
times, following a 200°C temperature increase in the wellbore: after The pore pressure can be expected to change in the
1 min. (blue), 10 min. (red), and 100 min (green).
cement sheath during the life of the well because, just
after the hydration and the resulting water consump-
tion, the cement pore pressure drops to a very low value
Large temperature changes can affect the mechanical (Chapter 9). Once the cement has set, fluid from the
properties of set cement through chemical transforma- formation will flow into the set cement to equilibrate
tions (see Chapter 10); however, the thermal dilation of the pressure, and the set-cement pore pressure will
the steel, cement, and rock can have a greater influence. increase. A rapid increase of wellbore pressure and
All of these materials expand as the temperature temperature will also lead to a pore-pressure increase.
increases and contract as the temperature decreases. Pore fluids in cement play an important role because
Damage may occur owing to nonuniform heating. A they support a portion of the total applied stress. Thus,
portion of the material that is being heated might be only the remainder of the total stress, the effective stress
prevented from expanding, while an unheated portion component, is carried by the cement matrix (Fig. 8-8).
might be subjected to forced expansion. This generates In 1923, Van Terzaghi first introduced the effective-
stresses in the set cement that may lead to failure or stress concept for one-dimensional consolidation of
debonding. To determine the thermoelastic behavior, soils, and proposed the following relationship:
such stresses must be added to those generated by
elasticity. Set cement, rocks, and tubulars will follow the σ ′ = σ + p pore , (8-19)
same behavior, although they will differ quantitatively.
The deformation caused by thermal effects in where σ is the total applied stress, σ′ is the effective
absence of stress is given by stress governing the consolidation of the material,
and ppore is the pore pressure (note that if compressive
ε = β T, (8-16) stresses are taken to be positive, the equation is
σ′ = σ – ppore). Biot (1941, 1956) later proposed a con-
where β is the coefficient of linear thermal expansion and sistent theory to account for the coupled diffusion and
Τ is the temperature. Therefore, for the stress-strain rela- deformation processes that are observed in elastic mate-
tionship: rials. For time-independent processes, this poroelastic
material behavior is similar to that of an elastic solid
σx ν
εx −βT = − σ +σz ,
E E y
( ) when the stresses in Eq. 8-14 are replaced by effective
stresses such that
σy ν
εy −βT = − σ +σz ,
E E x
( ) σ ′ = σ + αppore . (8-20)

σ This poroelastic constant, α, varies between 0 and 1,

ν
ε z −βT = z − σx + σy ,
E E
( ) and describes the “efficiency” of the fluid pressure
in counteracting the total applied stress. Its value
and depends on the pore geometry and the physical proper-
ties of the solid.
1 Another similarity is that fluids diffuse through set
γ rz = τ . (8-17) cement in a manner similar to temperature. Assuming
G rz

274 Well Cementing

 ■ An increase of pore pressure induces cement dilation.
■ A compression of the cement produces a pore pres-
sure increase if the fluid is prevented from escaping
the porous network.
ppore When the fluid is free to move, pore-pressure diffu-
ppore
sion introduces a time-dependent character to the
mechanical response of the set cement. The set cement
ppore will react differently according to the loading rate and
the ability of the cement matrix to accommodate the
resulting pore-pressure fluctuation.
ppore σ Hence, two limiting behaviors must be introduced:
ppore drained and undrained responses. One limiting case is
realized when a load is instantaneously applied to a
ppore porous cement. In this case the excess fluid pressure has
no time to diffuse; the medium will react as if it were
undrained. Undrained values of Young’s modulus and
ppore ppore Poisson’s ratio must be used in this situation. To determine
undrained values, one performs a compression test as
shown in Fig. 8-9a, in which the pore fluid is not allowed
to escape from the sample. On the other extreme, if the
pressurization rate is sufficiently slow and the diffusion
has ample time to drain excess pressure areas, the
Fig. 8-8. The cement matrix supports only a portion of the total cement will be less stiff, or more compliant. The stiffen-
stress, σ; the other portion is supported by the pore pressure
inside the cement.
ing effect during undrained deformation is greater if the
pore is filled with a relatively incompressible liquid,
rather than a relatively compressible gas.
To determine drained elastic values (which are the
that the set cement is incompressible and the pore fluid intrinsic elastic properties of the sample), one performs
is slightly compressible, the diffusion equation is a compression test as shown on Fig. 8-9b, in which the
pore fluid is allowed to escape from the sample and the
∂2 p pore ∂2 p pore ∂2 p pore φμ c ∂ p pore loading rate is low enough to prevent the generation of
+ + = × , (8-21)
∂x 2
∂y 2
∂z 2 k ∂t excess pore pressure in the sample.

with ppore the pore pressure, φ the porosity, μ the

viscosity of the fluid, k the permeability, and c the F F
fluid compressibility.
Permeability measurements use this diffusion
pgauge
equation (Appendix B). In this case one assumes that
the fluid is incompressible.

kA( Δ p ) pore pcon pcon

q= , (8-22)
μL

where q is the flow rate, A the sample cross section, and

L the sample length.
In reality, however, solid materials are compressible.
This compressibility, which changes the porosity of the a b
solid when stress and/or pore pressure is applied, means Fig. 8-9. In test a, the fluid cannot leave the sample and the pore
that any stress change induces a pore-pressure change. pressure is measured during the test. The test is undrained and the
One cannot dissociate the pore-pressure changes from deformations measure undrained elastic constants. In test b, the
the rock deformation. Two basic mechanisms highlight fluid can escape the sample and the pore pressure is maintained
this coupled behavior (Detournay and Cheng, 1993). constant (atmospheric in this case). The test is drained and the
deformations are used to measure drained elastic constants.

Chapter 8 Mechanical Properties of Well Cements 275

8-3.5 Plastic behavior is associated with a stress increment is the sum of an
Most set cements exhibit nonreversible deformations elastic component and a nonelastic component:
after unloading, or at least a nonunique relationship dε = dε el + dε pl , (8-23)
between stress and strain. This means that set cements
are not perfectly elastic materials, and a number of the- where dε is the total-strain increment, dε el is the
ories have been developed to model such behavior. They elastic-strain increment, and dεpl is the plastic-strain
include the theories of plasticity, damage mechanics, increment. Unlike the elastic-strain component, the
and time-dependent analysis (creep). As an example, plastic-strain component cannot be recovered during
the theory of elastoplasticity will be briefly described. unloading. To predict the plastic-strain increment one
Fig. 8-10 shows the stress-strain relationship of an needs a yield criterion that indicates whether plastic
ideal cylindrical elastoplastic sample. From O to A, deformation occurs, a flow rule that describes how the
the relation between stress and strain is linear, and the plastic strain develops, and a hardening law.
slope of the curve is the Young’s modulus, E. The The yield criterion is a relationship between the
stress-strain relationship does not change if the sample stresses that is used to define the conditions under
is unloaded in this region. This is the region in which which plastic deformation occurs. In three dimensions,
the theory of elasticity applies. Beyond Point A, the this is represented by a yield function that is a function
slope of the curve decreases. Moreover, if one unloads of the state of stress and a hardening parameter:
the sample in this region, say at Point B, the unloading
portion does not follow the same path as the loading
portion but is perfectly linear with a slope E. At zero
( )
f σ 1 , σ 2 , σ 3 , M = 0. (8-24)
stress, part of the deformation has not been recovered. The hardening parameter, M, determines the evo-
This unrecovered deformation represents the plastic- lution of the yield curve with the amount of plastic
strain component in the theory of elastoplasticity. The deformation of the material. For further details on
Point A is actually the initial yield stress of the cement. elastoplasticity, the reader is referred to Hill (1951) and
During reloading, the sample behaves as a perfectly Chen and Han (1988).
elastic solid up to the Point B, which is the new yield
stress. The increase of the yield stress with the increase
of plastic strain is called strain hardening; a decreasing 8-3.6 Creep of cement
yield stress with plastic strain is also possible, and is Creep is a general term that covers the time-dependent
called strain softening. A perfectly plastic material deformations of the set-cement matrix (not to be con-
is a material with no strain hardening or softening. As fused with the time-dependent deformation caused by
shown in this example, the yield stress is a function of pore fluid flow or temperature diffusion). All set
the loading history. cements present some creep behavior. Creep has been
In elastoplasticity, part of the strain is predicted thoroughly studied for concrete to prevent the failure
by the theory of elasticity. Any strain increment that of a concrete structure after a given amount of time,
but has been poorly studied in oilwell cements even at
ambient conditions.
As with plasticity, deformation caused by creep can
be added to the instantaneous elastic deformation
B ε = ε el + ε c ( t ), (8-25)
A
where εc(t) is the time-dependent deformation. This
deformation can be permanent (e.g., viscous defor-
Axial stress mation) or reversible (time-dependent elasticity).
Generally speaking, at low loads the deformation will be
time-dependent elastic, while at higher loads, viscous
deformation will dominate.
In general, the material creeps to reduce high shear
εpl εel
stresses. This process can be unstable, leading to mater-
O Axial strain ial failure. A typical example is the flow of salt zones. To
avoid high shear stresses at the wellbore wall, they tend
Fig. 8-10. Graphical representation of plastic deformation. to fill in and close the wellbore. Once the wellbore is
closed, the flow will stop.

276 Well Cementing

8-3.7 Set-cement strength and set-cement failure ing a limit beyond which instability or failure occurs. The
Set-cement failure occurs when cracks or discontinu- Terzaghi effective stress is used in the failure criterion
ities appear in the matrix. To achieve set-cement failure, σ ′ = σ + p pore . (8-26)
the cracks must be large enough to cause the separation
of fragments. The failure can be brittle or ductile Several criteria have been proposed in the literature
(Fig. 8-11). A brittle failure means that failure occurs in for various applications. The more popular criteria
the elastic section of the stress-strain deformation include the following.
curve. Cracks initiate and propagate very quickly. Brittle ■ The maximum tensile stress criterion states that
failure is often unstable because the elastic energy failure initiates as soon as the maximum effective
released by the cracking process is higher than the sur- principal stress reaches the tensile strength, Stens,
face energy consumed when cracks are created. In a of the material:
crushing test, this is characterized by shattering of the
set cement and fragments flying away. Set cements are σ 1′ = Stens . (8-27)
brittle in tension, under impact loading, and during a
compressive strength test at low confining pressure. ■ The Tresca criterion states that failure occurs when
Ductile failure means that failure occurs in the plastic the maximum shear stress, (σ1 – σ3)/2, reaches a
regime, when permanent deformation is created. Set characteristic cohesion value, Yco:
cements exhibit a ductile failure under confining pres-
sure. Ductile failure is more advantageous for cement- σ 1 − σ 3 = 2Yco . (8-28)
sheath integrity in a tectonic environment, because the
■ The Mohr-Coulomb criterion states that, for com-
set cement will deform significantly before being dam-
aged. In some cases the cement will be able to flow with- pressive failure, the shear stress tending to cause
out cracking. This is especially common with porous or failure is opposed by the cohesion of the material and
flexible cements at high confining pressure. by a factor analogous to the coefficient of friction
multiplied by the effective normal stress acting
across the failure plane.

)
τ = Yco − tan ( Φ σ n ′ , (8-29)

where Φ is the angle of internal friction and Yco is

the cohesion. The Mohr-Coulomb failure criterion
Ductile can be rewritten in terms of principal stresses to give
Axial stress σ3 at failure in terms of σ1 (recall that tensions are
Brittle positive, and under this convention, σ1 is the confin-
ing stress):

⎛ π Φ⎞
( )
σ 3 ′ = − σ c + tan 2 ⎜ + ⎟ σ 1′
⎝4 2⎠
(8-30)

where σc is the compressive strength.

Axial strain
This criterion shows an increase of compressive
Fig. 8-11. Brittle versus ductile failure. strength with an increase of confining pressure. As
shown above, the Tresca and Mohr-Coulomb criteria do
not include the influence of the intermediate stress, σ2.
To estimate the initiation of cement failure, one uses a Experimental evidence shows that, in many cases, it is a
failure criterion. A failure criterion is usually a relation- good approximation. However, other criteria include the
ship between the principal effective stresses, represent- effect of σ2.

Chapter 8 Mechanical Properties of Well Cements 277

8-3.8 Influence of confining pressure on set- 8-3.9 Postfailure behavior of set cement
cement behavior and failure There are cases when the set cement will be subjected
In a downhole situation, especially when the rock is to excessive loads and vibration (e.g., during window
creeping or when large tectonic stresses are present, drilling), leading to failure. In such cases, to ensure that
the cement will be subjected to confining pressure. the damaged set cement does not fall apart, it is impor-
Increased confining pressure leads to increased cement tant to control the postfailure behavior. As mentioned
strength, as predicted by the Mohr-Coulomb failure above, conventional set cements are brittle materials at
criterion and also by the transition from a brittle to low confining pressures. When cracks are initiated they
a ductile material. Fig. 8-12 shows an example of the can propagate very quickly. Unstable failure occurs
influence of the confining pressure on a cement sample. because the elastic energy released by the cracking
The tests were carried out as shown in Fig. 8-4, but each process is higher than the surface energy consumed by
test was performed at a different confining-pressure the creation of cracks. To control the failure, one must
value, pcon. The set-cement strength does indeed increase unload the cement sufficiently to release the elastic
with confining pressure, but in reality this increase energy. In a controlled experiment, the area below the
is not the linear one predicted by the Mohr-Coulomb stress-strain curve is exactly the energy per unit volume
criterion. An empirical law relating the cement strength consumed to propagate the crack. In a bending test
to the confining pressure might be preferred. (Fig. 8-13) or a direct-tension test, this energy per
One key feature of set cements under confining pres- unit volume is called toughness. Models have been
sure is that they quickly become fully plastic with strain developed to describe tensile-crack propagation based
hardening, and thus failure is no longer observed. on this type of relationship (Bosma et al., 1999).
Another characteristic is that, because cements are However, in difficult cases in which conventional
highly porous, they will compact quickly under confining set cements are so brittle that uncontrolled failure is
pressure (if the sample is drained). During compaction, inevitable, high-toughness cements are required. High-
the set-cement porosity decreases. This is the main toughness cements are designed so that the surface
cause of strain hardening, as the cement becomes area below the stress-strain curve is sufficiently great
stronger with decreased porosity (unless the porosity is to ensure that, when cracks propagate, they do so in
so high that the matrix collapses). a stable manner. Ideally the cracks do not extend or
coalesce; instead, small microcracks form.

0
Uniaxial
–10 17.24-MPa confinement

–30

Differential stress –40

(MPa)
–50

–60

–70
–50,000 –40,000 –30,000 –20,000 –10,000 0 10,000 20,000
Microstrain

Fig. 8-12. Radial and axial strain as a function of axial differential stress (axial stress – confining stress)
and confining pressure. Stress and strain values follow the convention that tensile stress is positive
(i.e., compression negative). Conventional neat system.

278 Well Cementing

 total chemical shrinkage is the volume reduction that
occurs as a consequence of cement hydration.
At 100% hydration, the total chemical shrinkage of
F Portland cement is about 6.25 cm3/100 g of cement
(Powers, 1958). The total chemical shrinkage can
be measured by placing cement paste in a container,
surrounding it with water, and monitoring the water
level versus time. This technique tends to underestimate
the total chemical shrinkage because it assumes that
the water has full access to the pores. This is not neces-
Flexural strength sarily the case, as the permeability of the matrix
decreases with time. Justnes et al. (1995) measured the
total chemical shrinkage of various Portland cement
Stress
compositions at ambient conditions (68°F [20°C] and
atmospheric pressure). After 48 hr the total chemical
shrinkage was found to be independent of the water-to-
cement ratio—about 2.17 cm3/100 g of cement for neat
E
Class G slurries (or 2.8 vol% for a water-to-cement ratio
Strain of 44 wt%). This difference between the results of
Fig. 8-13. Bending test or flexural test: The area below the stress-
Justnes and those of Powers indicates that at 48 hr the
strain curve measures the toughness of the sample. degree of hydration is only 2.17/6.25 = 0.35.
Once the cement begins to develop compressive
strength, the total chemical shrinkage can occur as
bulk shrinkage. Bulk shrinkage is the external volume
The most efficient way to promote high cement reduction that can occur during the hydration of Portland
toughness involves the addition of fibers or microribbons cement. It is a mechanical response to the stresses and
to the cement matrix (Chapter 7). Fibers resist breakage pore-pressure changes generated by chemical shrinkage.
when traversed by a microcrack and, up to a certain Bulk shrinkage can create a microannulus between
load, remain bonded to the cement matrix. This also the formation and the cement, preventing the cement
improves the tensile strength of the set cement. But the from fulfilling its role as a sealing material (Justnes
most dramatic effect is that the fibers hold the broken et al., 1994; Justnes et al., 1995). When the set cement
cement fragments together and prevent the generation shrinks, it moves toward the casing. Formation of a
of larger cracks (Xenakis et al., 1997). The toughness of microannulus can lead to gas leaks at the surface (even
set cements containing fibers can be significantly when production has ceased: Dusseault et al., 2000),
higher than that of conventional cement systems, even or incorrect leakoff-pressure estimation (Zhou and
at low fiber concentrations; however, its efficiency Wojtanowicz, 1999).
depends on the chemical nature of the reinforcement Bulk shrinkage can be measured by placing cement
(Park et al., 1999; Nataraja et al., 2000). High-toughness paste in a flexible membrane, sealing the membrane,
cements based on metallic microribbons have been and monitoring the volume variations of the membrane.
used as shock-resistant kickoff plugs (Babasheikh At the beginning of the experiment, the bulk shrinkage
et al.; 2003) (Chapter 14). curve matches that of the total chemical shrinkage.
Then the bulk-shrinkage curve flattens out as a rigid
structure starts to form, preventing the total collapse
8-3.10 Shrinkage and expansion of the material (Fig. 8-14). Total chemical shrinkage
Cement shrinkage and expansion are discussed in continues to occur, because cement hydration does not
Chapters 2 and 9; however, in the context of the present stop after the cement sets.
discussion, it is necessary to revisit the subject. A review Using similar techniques in which the cement paste
of the terminology surrounding this subject is useful to is isolated from the surrounding medium, several
prevent confusion. authors (Parcevaux and Sault, 1984; Chenevert and
The fundamental mechanism responsible for cement Shrestha, 1991; Bensted, 1991; Sabins and Sutton, 1991;
shrinkage is the total chemical shrinkage (following Justnes et al., 1994; De Rozières and Sabins, 1995;
the terminology of Justnes et al., 1994). As discussed Justnes et al., 1995; Justnes et al., 1996; Backe et al.,
in Chapter 2, the volume of the reaction products of 1997) have reported bulk shrinkage values of 0.5 to 5.0%
cement hydration is less than that of the reactants. The by volume. Although these results were obtained under

Chapter 8 Mechanical Properties of Well Cements 279

 2.5 enough to allow the external fluid to invade the pore
structure (Appleby and Wilson, 1996). Restated in solid
2.0 mechanics terms, the pressure and saturation of the
1.5
fluid in the cement pores must remain constant during
Shrinkage cement hydration to avoid bulk shrinkage.
(mL /100 g) 1.0 When cement has no free access to water (e.g.,
because of the presence of an impermeable membrane),
0.5 total chemical shrinkage leads to pore-pressure reduc-
tion and eventual pore collapse. Restated in solid
0
mechanics terms, the cement compacts if the external
0 10 20 30 40 50 60
total stress remains constant and the internal pore
Time (hr)
pressure decreases during cement hydration, leading
Fig. 8-14. Chemical shrinkage (green line) and bulk shrinkage (red to an increase of effective mean stress. Note that the
line) as a function of time for a cement with a water/cement ratio chemical shrinkage can be so high that pore saturation
by weight of 0.4 (after Justnes et al., 1995). Reprinted with permis- falls below 100%, and, because of capillary effects, a
sion from Thomas Telford Limited. negative pore pressure is observed.

a variety of temperature and pressure conditions, such a 8-4 Mechanical behavior of a cement
wide range of results is probably caused by the formation
of free water during the experiments (Justnes et al., cased wellbore
1996) and the failure to control the boundary conditions. 8-4.1 State of stress in the cement sheath
If bulk-volume variations are measured when the To determine whether a cement sheath will fail or debond
cement has access to additional water during the test in the annulus, one must calculate the state of stress.
(e.g., by measuring the dimensional variation of an The calculated stress is then entered into an expression
annular ring mold or cylindrical sleeve filled with the to determine whether failure is attained. To calculate
cement paste and placed in water), a bulk expansion as the state of stress, one must assume a deformation
high as 0.3% by volume is observed after the cement sets behavior (e.g., elasticity) and consider the various
(de Rozières and Sabins, 1995). Uncontrolled bulk applied loads at specific boundaries such as the
expansion can be as harmful as bulk shrinkage, because casing/cement and cement/formation interfaces. In some
it can disrupt the casing/cement interface (Beirute et cases, the influence of temperature and pore pressure
al., 1988; Baumgarte et al., 1999). To avoid this prob- must also be considered.
lem, one must ensure that the cement has a lower In recent years, various models have been devel-
Young’s modulus value than the surrounding rock oped to analyze the state of stress in the cement
(Baumgarte et al., 1999, Le Roy-Delage et al., 2000). (Thiercelin et al., 1997; Bosma et al., 1999; Gino di
In the context of the mechanical properties of well Lullo and Rae, 2000; Fleckenstein et al., 2000;
cements, the initial phase of shrinkage is not relevant Philippacopoulos and Berndt, 2002; Pattillo and
(Setter and Roy, 1978). During this phase, the cement is Christansen, 2002). They are based on analytical solu-
still a liquid slurry. During primary cementing, before tions, numerical solutions, or a combination of both.
the cement slurry begins to set, the top of the cement
column moves downward to compensate for the volume
reduction (Chenevert and Shrestha, 1991). The bulk 8-4.2 Modeling the cement sheath using
shrinkage that occurs after a rigid network of hydration thermoelasticity
products has formed and compressive strength begins to In this section the modeling of stresses in a cased
develop is relevant to the mechanical properties. wellbore containing a finite number of concentric
Previous studies of cement shrinkage have shown two casings is briefly described. A cross section of the well-
key behaviors that can be easily linked to porous elasto- bore is shown in Fig. 8-15. The stresses in the cement
plastic solid behavior controlled by effective stress are calculated assuming that casing, cement, and rock
(Thiercelin et al., 1998). When cement has free access to are thermoelastic materials. The casing/cement and
additional water, the external water flows into the cement/rock interfaces are also assumed to be either
cement pore space to compensate for the total chemical fully bonded or unbonded. Finally, it is assumed that the
shrinkage, and almost no bulk shrinkage is observed. In cement is under no internal “effective” stress after
some formulations, an expansion might even be observed. setting. This final assumption is obviously a strong
For this to occur, the cement permeability must be high

280 Well Cementing

 Depth of interest
Rock Cement
σθ
Section at depth of interest

Cement

σr r

Casing

Fig. 8-16. Radial stress and tangential stress.

Casing

σr ν
)
Fig. 8-15. A cross section of the well.
εr −βT = − σ +σz
E E θ
(
σ ν
simplification. It is known that, after placement, the
cement slurry unloads as its gel strength develops
εθ − β T = θ − σ r + σ z
E E
( )
σ ν
)
(Chapter 9). Field observations (Cooke et al., 1983;
Morgan, 1989) tend to confirm that the total stress in ε z − β T = z − σ r + σθ
E E
(
the cement drops to at least the hydrostatic pressure
1
given by the saturating formation fluid (mainly water), γ rz = τ rz , (8-31)
justifying the zero-effective stress assumption. However, G
in some cases the pore pressure can drop below the
hydrostatic pressure, especially in a casing-to-casing where E, ν, and G are respectively the Young’s modu-
configuration. One can imagine a variety of situations, lus, Poisson’s ratio, and shear modulus, and β is the
depending on the cement properties, cementing coefficient of linear thermal expansion.
procedure, formation permeability, and nature of The temperature distribution as a function of time is
the saturating fluid. Nevertheless, in the absence of obtained from the heat diffusion equation, which is
better information, this simplification is appropriate. expressed under the assumption that the initial and
Consequently, to study the cement behavior, only the boundary conditions do not depend on θ, as
variations of pressure, stress, or temperature that occur
∂2 T 1 ∂ T ∂2 T ρ C ∂ T
after the cement sets are considered. + + = , (8-32)
The geometry of the problem is axisymmetric, with ∂r 2 r ∂r ∂z 2 λ ∂t
the axis of symmetry being the wellbore axis, allowing
the use of cylindrical coordinates r, θ, and z. The sim- where λ is the thermal conductivity, ρ the density, and
plest situation is when the boundary and initial condi- C the specific heat.
tions (wellbore and far-field states of stress and temper- Plane strain is also assumed, meaning that there is no
ature) are independent of θ. The variables of interest are axial movement. This is usually a good assumption,
then the radial displacement; radial stress, σr; tangential although axial movement could develop when casing
stress, σθ; axial stress, σ z; the shear stress, τrz; and the sections are being heated and axial casing deformation
temperature, T (which in practice is the temperature dif- is not prevented at the surface.
ference from a reference state). The tangential stress is The stress model uses analytical solutions that have
a principal stress. The radial and tangential stresses are been presented in Thiercelin et al. (1997). The solution
shown in Fig. 8-16. The sign convention is that tensile is constructed with the conditions that the radial
stresses are positive. Thermoelasticity provides a linear displacements and radial stress are continuous across
. the interface between two materials and the radial
relationship between the strains ε r, ε θ, ε z, and γrz,
stresses, and temperature, T. stress is compressive. If the radial stress is tensile (or

Chapter 8 Mechanical Properties of Well Cements 281

above a small value that can be considered as cement 3.0
adhesion at the interface), debonding occurs, radial
displacement is discontinuous, and the radial stress is 2.5
set to zero. The inner surface of the inner casing has an
imposed radial stress condition, given by the variation of 2.0
wellbore pressure on its inner surface. Stress 1.5
Heat transport takes place by conduction, as described (MPa)
above. To calculate the evolution of temperature with 1.0
time, it is often more convenient to use a numerical
technique such as the finite-difference technique or the 0.5
finite-element technique. The influence of temperature
on the state of stress is a function of thermal expansion 0
of the various materials. Time is introduced by the heat- 60 65 70 75 80 85 90
diffusion process. Distance from well axis (mm)

8-4.3 Influence of wellbore pressure increase Fig. 8-18. Tangential stress in the cement sheath as a function of
distance from the wellbore.
The most damaging wellbore pressure increases often
occur during a pressure test of the casing. It is indeed
unfortunate that, by checking the casing integrity, one
can damage the cement sheath. An increase of mud respectively. The wellbore pressure increase is 2,900 psi
weight, a hydraulic fracturing treatment, or a perfora- [20 MPa]. The openhole diameter is 7 in. [178 mm], the
tion test can also generate large wellbore pressure casing outside diameter (OD) is 5 in. [127 mm], and the
increases. A wellbore pressure increase induces a casing weight is 23.20 lbm/ft [34.53 kg/m]. Because we
radial elastic expansion of the casing, which in turn are using linear elasticity with a single loading condi-
loads the cement. tion, the stress values are a linear function of wellbore
The effect of a pressure increase on the state of stress pressure. Thus, doubling the wellbore pressure results in
in the cement sheath is shown in Figs. 8-17 and 8-18, in doubling the value of the stresses in the cement.
which the radial and tangential stresses in the cement These figures show that, in this case, the radial stress
are shown as a function of the distance from the well- is compressive and the tangential stress is tensile. The
bore axis. In this case, the Young’s moduli of the steel, tangential stress is about half the absolute value of the
cement, and rock are 29.0 × 106 psi [200 GPa], 0.725 radial stress. Because cements are about 10 times
× 10 6 psi [5 GPa], and 1.45 × 10 6 psi [10 GPa], weaker in tension than in compression, the cement
respectively. The Poisson’s ratios are 0.27, 0.15, and 0.2, failure will occur in tension. Tensile failure appears
when the tensile stress is greater than or equal to the
tensile strength. The highest value of the tangential
1.0 stress is at the steel/cement interface; therefore, this is
where failure should first occur. In this case, cement
0.5 failure will correspond to the initiation and propagation
of tensile radial cracks, because tensile cracks propagate
0 perpendicular to the direction of the maximum tensile
Radial stress. If the wellbore pressure increases by 2,900 psi
stress –0.5 [20 MPa], the value of the tangential stress at the
(MPa) steel/cement interface can be used to calculate the
–1.0 tensile strength the cement must have to avoid failure.
The cement Young’s modulus has a strong influence
–1.5
on the cement sheath response. This is demonstrated in
–2.0
Fig. 8-19, which shows the tangential stress as a function
60 65 70 75 80 85 90 of distance from the wellbore axis. In this case the
Young’s modulus of the cement is 0.725 × 10 5 psi
Distance from well axis (mm)
[500 MPa]. This time the tangential stress in the cement
Fig. 8-17. Radial stress in the cement sheath as a function of is less tensile, even compressive, near the cement/rock
distance from the wellbore. interface. This is because of the mechanical support pro-

282 Well Cementing

 0.20 Table 8-2. Cement Failure as a Function of Cement
Properties During a Wellbore Pressure Increase
0.15
Formulation A (Conventional) B (Elastic)
0.10 Tensile strength (psi [MPa]) 566 [3.9] 305 [2.1]
Stress 0.05 Young’s modulus (psi [MPa]) 1,311,000 [9,041] 376,000 [2,594]
(MPa)
Poisson’s ratio 0.15 0.22
0
Tangential stress at 998 [6.88] 138 [0.95]
–0.05 wellbore (psi [MPa])

–0.10 Failure Yes No

60 65 70 75 80 85 90
Distance from well axis (mm)
Fig. 8-19. Tangential stress in the cement sheath as a function strength, and that cement failure in tension is expected.
of distance from the wellbore for a cement Young’s modulus of A less stiff cement containing flexible particles, which
500 MPa. gains support from the rock, is more appropriate
(Formulation B). In this case, the tangential stress does
not exceed the tensile strength of the cement, even
vided by the rock. In fact, the resistance of the cement though the actual tensile strength is less than that of the
sheath is increased, as this effect usually compensates conventional cement.
for the decrease of cement strength often associated
with a decrease of Young’s modulus. The mechanical 8-4.4 Influence of wellbore temperature increase
support is actually related to the ratio of the set
cement’s Young’s modulus to the rock’s Young’s modulus, The calculation of stresses on set cement caused by a
so a rock with a higher Young’s modulus will show similar temperature increase involves additional parameters
behavior in reducing the tensile stress in the cement. such as the density, specific heat, thermal conductivity,
This also happens when the set cement’s Poisson’s ratio and the coefficients of thermal expansion of the various
is higher. Obviously other issues, such as casing protec- materials. For simplification one can assume that the
tion and casing support, must be considered. coefficients of thermal expansion of the steel, cement,
These results demonstrate that, to determine and rock are the same: 1.3 × 10–5 K–1. The complete set
whether the cement will fail owing to a wellbore pres- of thermoelastic properties is given in Table 8-3.
sure increase, one must know the tensile strength of Other material parameters are the same as in the exam-
the cement and the elastic properties of the cement and ple in Section 8-4.3, except for the stiffness of the
the rock. The geometry of the cased wellbore is also cement, which, in this example, is higher than that of
an important parameter. For example, increasing the the rock.
casing thickness will decrease the tensile strength The radial and tangential stresses in the cement show
requirement. the same pattern as that during a wellbore pressure
A field example was described by Le Roy-Delage et al. increase, with the generation of high tensile tangential
(2000). A cemented cased section of a well, in which stress in the set cement. The stresses are first generated
the slurry was displaced with a low-density mud, was to by the rapid thermal dilation of the casing. Then, as heat
be subjected at a later stage to a pressure increase of diffuses through the set cement and eventually into
6,500 psi [44.8 MPa]. This pressure increase was the
consequence of a mud-density increase, required to drill
through a salt zone located at a deeper location. Table 8-3. Assumed Thermal Properties of Steel, Cement,
The openhole diameter was 12 in. with a 9 5⁄8-in., and Rock
53.5-lbm/ft casing. The Young’s modulus and Poisson’s
Steel Cement Rock
ratio of the rock were 3.6 × 106 psi [24,800 MPa] and
0.25, respectively. The mechanical properties of the Solid density (kg/m3) 8,000 1,900 2,100
cement formulations used for this application are shown Specific heat (J/kg-K) 500 2,100 1,900
in Table 8-2. The slurry-density requirement was 16
lbm/gal [1,920 kg/m3]. Table 8-2 shows that the tangen- Thermal expansion coefficient (K–1) 1.3 × 10–5 1.3 × 10–5 1.3 × 10–5
tial stress imposed on the conventional cement
Thermal conductivity (W/m-K) 15 1.0 1.0
(Formulation A) largely exceeds the cement’s tensile

Chapter 8 Mechanical Properties of Well Cements 283

the rock, stresses are generated by the thermal dilation cement density varied from 12 lbm/gal [1,440 kg/m3] to
of the cement. Fig. 8-20 shows the tangential stress 14 lbm/gal [1,680 kg/m3]. The operator required a low-
in the cement, 1 min after a 360°F [200°C] temperature permeability cement with a final compressive strength
increase in the well. The tangential stress is compressive of at least 1,500 psi [10 MPa]. As shown in Table 8-4,
near the steel/cement interface and tensile near the a conventional 14-lbm/gal [1,680 kg/m3] cement
cement/rock interface. This behavior is caused by the (Formulation C) would not support the stress generated
nonuniform thermal expansions of the materials. The by such a temperature increase. The thermal expansion
temperature in the region near the well is higher than of the casing imposes a tensile tangential stress well
that further away. Compressive tangential stress is gen- above the tensile strength of the cement sheath.
erated in this region because the thermal expansion is Moreover, because the well was shallow, the formation
confined by the cooler materials in the surrounding (sandstone) had a low Young’s modulus and would not
region. Similarly, tensile stress is generated in the sur- prevent large deformations of the cement sheath. The
rounding region that is being pressurized by the near- model shows when and where the tensile stress reaches
wellbore region. As temperature diffuses through the the most critical value in the cement sheath. In this case,
cement, the stress pattern will evolve. The cement the most critical location is 5.2 in. [132 mm] from the
achieves a more uniform temperature and is therefore wellbore axis. Various formulations were tested.
under less stress. Formulation D, containing flexible particles, was able to
provide very low stiffness while retaining sufficient com-
pressive strength and low permeability. Table 8-4 demon-
8 strates that this cement will not crack under the thermal
loading, because the tangential stress at the critical loca-
6 tion is always below the value of the cement’s tensile
strength.
4

Stress 2
(MPa) Table 8-4. Cement Failure as a Function of Cement
Properties During a Wellbore Temperature Increase
0
Formulation C D
–2 Tensile strength (psi [MPa]) 334 [2.3] 218 [1.5]

–4 Young’s modulus [psi [MPa]) 545,000 [3,759] 177,000 [1,221]

60 65 70 75 80 85 90
Poisson’s ratio 0.16 0.24
Distance from well axis (mm)
Maximum value of 799 [5.51] 149 [1.03]
Fig. 8-20. Tangential stress as a function of wellbore radius tangential stress (psi [MPa])
after 1 min.
Failure Yes No

This process is similar to a pressure increase, but is

more complex owing to the influence of the temperature 8-4.5 Influence of wellbore pressure and
diffusion in the set cement and in the rock and the temperature decrease
existence of a temperature gradient resulting in the non- A decrease of wellbore temperature and pressure
uniform thermal dilation of a given material. A detailed creates tensile stress at the cement/rock and
analysis as a function of time is therefore required. cement/casing interfaces, if the cement is assumed to
A second field case addresses one of the most severe be fully bonded to the rock and casing. The radial and
conditions a cement sheath would ever experience in a tangential stresses are then exactly the same as in
well: steam injection (Le Roy-Delage et al., 2000). For Fig. 8-17 and Fig. 8-18, but with a sign change (again a
this well, the temperature was increased to 400°F consequence of elastic behavior). Radial stress is now
[204°C] from 120°F [49°C]. The openhole diameter as tensile and, because the adhesion forces are often very
measured by the caliper was 13 in. [330 mm]. The casing small, debonding will occur. Whether debonding will
was 9 5⁄8 in. [244 mm] OD and 40 lbm/ft [59.53 kg/m], and occur at the rock /cement or cement/casing interface
the Young’s modulus and Poisson’s ratio of the rock were depends on numerous parameters, such as the presence
580,000 psi [4,000 MPa] and 0.15, respectively. The

284 Well Cementing

of a thick mud cake at the cement/rock interface. As bonded to the casing has sufficient strength to support
demonstrated by bond logs, debonding often occurs at the the weight of the casing plus the weight of the drillstring
casing/cement interface, and a microannulus is created. during shoe drilling or other events. The strength of
The following field case is an example of the problem, interest is the shear bond strength, often empirically
although more severe cases have been encountered. An associated with the tensile strength of the cement.
intermediate section of a gas well is analyzed. The open- Casing support is usually calculated using the follow-
hole diameter is 12 1⁄4 in. [311.1 mm], with a 9 5⁄8-in. ing equation.
[244-mm], 66.8-lbm/ft [99.41-kg/m] casing. During
completion the heavy mud used to drill the deeper sec- F = 0.969 S × d × h, (8-33)
tion was replaced by a completion fluid with a density where F is the load to break the cement bond in lbf, S is
close to water. This resulted in a 3,290-psi [23-MPa] the compressive strength in pounds per square inch
pressure drop in front of the casing. However, during (assuming it is 10 times the shear-bond strength), d is
production, the temperature of the completion fluid the outer diameter of the casing string in inches, and h is
increased by 36°F [20°C]. The combined effect of the the height of the cement column in feet.
pressure decrease and temperature increase is shown in Assuming that only 5% of the cement is supporting the
Fig. 8-21. First, a 58-μm microannulus was created. load (which can be viewed as a very strong safety factor),
However, as the casing was being heated, it dilated and it can be shown (di Lullo and Rae, 2000) that, even with
reduced the size of the microannulus to 28 μm, which the heaviest casings, the required compressive strength
is still sufficient to create a gas leak. The only solution is around 143 psi [1 MPa]. This is well below the
would be to use an expanding cement (Seidel and accepted 500-psi [3.5045-MPa] standard. In practice,
Greene, 1985; Moran et al., 1991; Baumgarte et al., 1999, with normal casings, 50-psi [0.35-MPa] compressive
Le Roy-Delage et al., 2000; Chapter 7). However, as strength would be sufficient, especially if one considers
mentioned earlier, the cement must be more flexible than that the confining pressure might even increase the
the surrounding rock (Baumgarte et al., 1999, Le Roy- strength.
Delage et al., 2000) to ensure good bonding. For perforating and fracturing, a required compres-
sive strength of around 1,000–1,400 psi [7–10 MPa] is
often mentioned. The earlier discussion shows that this is
60 not necessary, because the pressures generated during
50 perforating and hydraulic fracturing result in a wellbore-
pressure increase. In many cases, a compliant or highly
40 flexible cement with low compressive strength will pro-
Micro- vide better support than a strong stiff cement. The
annulus 30 cement does not need to have high compressive strength
size (µm) to fulfill its role in casing support or zonal isolation.
20
Field observations support this statement (Deeg et
10 al., 1999). However, the flexible cement must be of good
quality and proven resistant to shock. For example,
0 highly permeable cements saturated with water might
0 10 20 30 40 50
suffer excessive shock because of the generation of
Time (min) excessive pore pressure in the cement matrix.
Fig. 8-21. Microannulus as a function of time.
8-4.7 Tectonics, subsidence, and formation creep
Tectonic stresses, subsidence, and formation creep can
8-4.6 Casing support, perforation and fracturing lead to loading of the cement by the formation and even-
tually the collapse of the casing. The loads involved are
The compressive strength required to support casing and often so large (except in the case of formation creep)
survive perforation and hydraulic fracturing treatments that it is practically impossible to prevent failure alto-
has been a subject of debate for many years. Traditionally gether. However, by using a careful completion design,
it was believed that the higher the compressive strength one can at least retard the collapse.
of the cement, the better the result would be. Recent When rock deformation or rock loading occurs, the
developments allow a clarification of the subject. cement can be strongly confined and compressed.
The situation concerning casing support is straight- Plastic flow of the cement leading to shear ductile failure
forward. Casing support means that the set cement would be the main deformation and failure mechanisms.

Chapter 8 Mechanical Properties of Well Cements 285

For these situations, it is necessary to know the cement
behavior as a function of the confining pressure.
The influence of rock stresses or rock deformation on
cement and casing failures can be grouped in three main
loadings.
1. The first includes a more or less continuous loading of
the cement bond and casing. It can be very isotropic if
rock creep is involved, or strongly directional if tec-
tonic stresses are involved (Fig. 8-22). The most
remarkable example is the influence of tectonic
stresses on cased wellbores in the foothills of
Colombia, South America (Last et al., 2002), which
leads to the ovalization of the casing string. Field data
show that the presence of the cement sheath
improves the stability of the casing owing to better
load distribution on the casing, and that a ductile
cement that deforms before failure also helps. Their
analysis is supported by a numerical study based on
the finite element method, in which casing and
cement elastoplastic behavior are taken into account.
Overall, the main recommendation is to use a dual
cemented string. The cement between the two strings
has the goal to strengthen and stiffen the dual string,
so it must be as strong and stiff as possible. Fig. 8-23. Casing shearing owing to shear band activation.

3. The last loading is axial compression of the casing

linked with the compaction of the formation adjacent
to it. This leads to casing buckling (Fig. 8-24), essen-
tially in poorly cemented zones. Heating of the casing
will have the same consequences because of thermal
vertical expansion. Experimental studies (Veeken
et al., 1994) again show that the presence of the
cement is essential to provide casing support and
avoid buckling.
Fig. 8-22. Casing ovalization caused by tectonic loads.

2. The second loading is linked to the activation of

natural faults (Fig. 8-23) or the creation of new
ones. This is typical of the overburden response to Zone with
strong compaction and leads to severe shearing of the no cement
casing. Recommendations (Dusseault et al., 1998) are
essentially based on either a strengthening of the
casing (dual cemented string or thicker casing) or an
increase of system compliance such as the under-
reaming of the zone, the use of ductile cements, and Fig. 8-24. Casing buckling owing to compaction.
even the avoidance of cementing. Pattillo and
Kristiansen (2002) used the finite element method to
analyze the stress field in the casing and cement and
showed that ductile cement aids stability by avoid-
ing extreme casing deformation.

286 Well Cementing

 The overall recommendation to maintain wellbore sta-
bility as long as possible is to strengthen the casing as
(2.109 − 1.174 ) × 914.4 × 9.81 = 8.39 MPa
much as possible and to ensure that the annulus is fully 1, 000
cemented. The cement must have good plastic/ductile or (0.914 − 0.509) × 3, 000 = 1, 215 psi
behavior and flow easily under stress to avoid casing
■ The required cement mechanical properties to sur-
pitching and to provide casing support.
Poor cement or microannulus formation can also lead vive such a change of downhole conditions are then
to a gas or oil leak between the reservoir and the over- determined using a stress model of the cemented
burden. In overpressured reservoirs, such fluid leaks can cased wellbore. The input data to the model include
lead to the pressurization of natural faults, which in turn the well geometry; steel, cement, and rock properties;
become unstable. Shear reactivation and shear collapse and the expected variation of downhole conditions.
of the casing can therefore be either promoted or The cement properties are obtained from the data-
induced by these leaks. Good bonding must be ensured. base mentioned above.
■ Cement selection using such a model is performed
with an iterative process, though one can imagine
8-5 Guidelines for cement design that, in the future, the model will be able to deter-
mine the best configuration automatically. The best
To design a cement system that resists the downhole solution is the most cost-effective cement that does
conditions, the following methodology is proposed. not fail under the input loading condition. The model
■ It is first essential to have a database of cements can be used to provide recommendations on cement
or alternate sealants with the desired mechanical design by tuning the properties of the cement to
properties (elastic parameters, compressive strength, minimize the expected problem. Unfortunately, in
tensile strength, shrinkage/expansion) at various complex situations, such as in tectonically active
temperatures and confining pressures. For example, areas, the modeling can be very involved.
having a variety of cement systems with low, medium, ■ The model must also be able to predict whether the
and high Young’s moduli is recommended. The data- loading can induce the formation of a microannulus,
base allows us to know what can be achieved under and in that case it is necessary to select the cement
various constraints such as slurry density, slurry accordingly (for example expanding cement with a
rheology, and cost. Rock properties are also needed Young’s modulus lower than the rock properties).
and can be obtained from sonic logs.
■ Finally, the integrity of the cement sheath should be
■ The second step is to properly understand and
monitored over time to validate the selection.
describe the downhole situation to identify the cause
of potential problems. This requires a review of all the To help the field engineers design cements based
events that may occur after the cement sets, such on mechanical considerations, software tools that follow
as a change of wellbore fluid, change of temperature, the above guidelines have been developed. An applica-
formation creep, subsidence, or other tectonic event. tion of such software tools for the cementation of steam-
In situations in which the casing can be damaged assisted gravity drainage wells is given in Stiles and
by the external loading, determination of the casing Hollies (2002). This stress analysis model software con-
failure mode (Section 8-4.7) from downhole informa- tains databases on cement, rock, and steel mechanical
tion is necessary. properties; a calculator to determine downhole stress
Example: A casing has been cemented with a conditions; a stress model to predict whether the
0.821-psi/ft or 15.8-lbm/gal [1,900-kg/m3] cement cement will fail (as described in Section 8-4); and a tool
displaced with a 0.509-psi/ft or 9.8-lbm/gal to help in designing the cement.
[1,174-kg/m3] mud. The reference pressure is the This software tool greatly facilitates the design of spe-
pressure applied inside the casing by the mud during cial cement for long-term zonal isolation, and this type of
cement setting, not the pressure applied by the software tool is recommended to design critical cases.
cement in the annulus. The drilling of the next sec-
tion, in which overpressured shales are encoun-
tered, requires a mud-weight increase to 0.914 psi/ft
or 17.6 lbm/gal [2,109 kg/m3]. The maximum pres-
sure increase is at the shoe, located at 3,000 ft
[914.4 m]. The pressure increase is

Chapter 8 Mechanical Properties of Well Cements 287

8-6 Conclusion in preventing microannulus formation. Other configura-
The determination of optimal long-term mechanical tions require high-strength, high-stiffness cements, such
properties that cements must have to ensure zonal as the cement used in a double-string casing and in
isolation is a relatively new approach in the oil and kickoff plugs. In a high-stress environment, it seems
gas industry. It is therefore an area in which new that the main requirement is that the cement must
development occurs at a high pace. The determination ensure isolation and keep the rock away from the casing
of cement mechanical properties under downhole as much as possible. To achieve its goal, it must be
conditions and the modeling of cased wellbore deforma- able to flow without cracking. Soft cements might help
tions are current topics of research. Numerous aspects reduce the load to the casing when shear fault reacti-
are still poorly known, such as the influence of the vation is a problem, in cases in which the slip of the fault
slurry-to-solid transition on the initial state of stress, is limited.
the role of cement pore pressure, and the modeling Cements and alternate sealants are being developed
of cement expansion or shrinkage in a downhole with properties that are more appropriate for downhole
environment. Nevertheless it is clear that cement conditions than conventional oilwell cements. There is
strength is not the only key mechanical property, and no doubt that new products with enhanced properties
that the cement’s elastic properties play a significant will continue to appear in the coming years.
role in maintaining the integrity of the cased wellbore.
In most configurations, especially when the casing
expands because of pressure increase or temperature 8-7 Acronym list
increase, soft elastic cements are the best solution. OD Outside diameter
Expanding cements must also have low Young’s moduli UCA Ultrasonic cement analyzer
to be properly confined by the rock and to play their role

288 Well Cementing

9-1 Introduction
Annular Formation Fluid Migration

Annular fluid migration during drilling or well-comple-

tion procedures has long been recognized as one of the
David Stiles—Schlumberger

Well 1
Interzonal communication
9
Well 2
Gas to surface

most troublesome problems of the petroleum industry. It

is defined as the invasion of formation fluids into the
annulus owing to a pressure imbalance at the formation
face. The fluids may migrate to a lower pressure zone or Low-pressure
possibly to the surface (Fig. 9-1). Gas migration is the zone
most common form of annular fluid migration and no
doubt the most dangerous (Carter and Slagle, 1970;
Sutton and Faul, 1984). Gas migration is the primary
focus of this chapter; however, most of the concepts pre-
sented here also apply to other formation fluids, such as High-pressure
shallow water (Stiles, 1997; Watters and Beirute, 1998). gas zone
Gas migration—also called gas communication or gas
leakage (Carter and Slagle, 1970), annular gas flow
(Garcia and Clark, 1976), gas channeling (Parcevaux et
al., 1983), flow after cementing (Webster and Eikerts,
1979), or gas invasion (Bannister et al., 1983)—is a
problem for both gas-producing and gas-storage wells.
The severity of the problem ranges from the most haz-
ardous (e.g., a blowout situation arising from a severe
pressure imbalance during drilling or cementing) to the Fig. 9-1. Two scenarios of annular gas migration.
most marginal (e.g., a residual gas pressure of a few psi
at the wellhead). In addition to surface-related difficul-
ties, communication between two or more subsurface
zones can occur. Such problems are more difficult to sure or gas flow at the wellhead, may dictate well aban-
detect. donment. More frequently, remedial cementing is per-
This chapter concentrates on gas migration after pri- formed until gas flow is stopped and gas pressure is
mary cementing. However, many of the elements of gas reduced to a level compatible with the operator’s safety
migration are similar to well control during drilling policy and local regulations. However, the efficiency of
(Moore, 1974). Although drilling is beyond the scope of squeeze cementing in such circumstances is very poor
this chapter, some similarities between gas migration for three essential reasons: (1) gas channels are diffi-
during cementing and gas migration during drilling are cult to locate, especially if they are less than 1 mm in
highlighted. size; (2) gas channels may be too small to be effectively
filled by cement; and (3) the pressure exerted during the
squeeze job is sometimes sufficient to break cement
bonds or even to initiate formation fracturing, further
9-2 Practical consequences of gas worsening downhole communication problems.
migration Furthermore, cement repair operations are expensive,
The potential consequences of gas migration following especially offshore or at remote locations (Cooke et al.,
primary cementing are numerous, but not always imme- 1982). Therefore, preventing a gas migration problem is
diately detectable. At the extreme, those that manifest definitely preferable to attempts to repair it. A thorough
themselves at the surface, e.g., sustained casing pres- discussion of remedial cementing appears in Chapter 14.

Well Cementing ■ Chapter 9 Annular Formation Fluid Migration 289

 Gas migration between two or more subsurface zones, ■ Applications of new products and techniques in the
with no surface manifestations, is very difficult to detect field (Kucyn et al., 1977; Watters and Sabins, 1980;
(Fig. 9-1). In such cases, gas production may be Cheung and Myrick, 1983; Seidel and Greene, 1985;
impaired, gas may be diverted to an upper depleted zone Sepos and Cart, 1985; Matthews and Copeland, 1986)
(possibly followed by gas migration to the surface ■ Empirical qualitative prediction techniques (Sutton
through another well), or the efficiency of stimulation et al., 1984; Rae et al., 1989).
treatments may be reduced (Cooke et al., 1982). Such
downhole channeling can sometimes be evaluated by Successful numerical simulations of the process or
special methods such as noise logs (Garcia and Clark, scaled laboratory experiments that could allow a gener-
1976) or acoustic logs (Catala et al., 1984; Rang, 1987) alized and quantitative prediction of gas migration have
(Chapter 15). Hydraulic communication testing is not not been reported.
recommended. If not properly designed and controlled,
these tests may induce communication across properly 9-3.1 Root causes for gas migration
cemented zones or aggravate minor cement-job defects.
A complete discussion of cement-job evaluation is pre- Annular gas migration has three distinct root causes
sented in Chapter 15. (Fig. 9-2).
1. The hydrostatic pressure in the annulus falls to a level
that is less than or equal to the pore pressure of a gas-
9-3 Physical process of gas migration bearing zone.
2. Space in the annulus allows gas entry.
Gas migration is a complex process that is influenced by 3. A path is present in the annulus through which the
many factors: fluid density control, mud removal, gas can migrate.
cement-slurry properties, cement hydration, and inter-
actions between the cement, casing, and formation. The All three root causes must be satisfied for annular gas
cementing industry first recognized gas migration in the migration to take place. The root causes involve various
early 1960s, when major gas-communication problems factors inherent to the cementing process. These factors
occurred in gas storage wells in the United States (Stone are described in this chapter.
and Christian, 1974). Since then, the industry has
worked diligently to understand the problem and find
solutions.
Extensive research has been performed to under-
stand the fundamentals of the physical process, and a Path to
vast quantity of literature describes various aspects of migrate
gas migration.
■ Field case study analyses and experiments for making
practical recommendations (Vidovskii et al., 1971; Gas
Stone and Christian, 1974; Garcia and Clark, 1976; migration
Cooke et al., 1982; Lukkien, 1982; Al-Buraik et al.,
1998)
■ Laboratory investigations to improve the understand-
Annular ≤ Formation
Space for entry
pressure pressure
ing of gas-migration fundamentals (Guyvoronsky and
Farukshin, 1963; Carter and Slagle, 1970; Carter et Fig. 9-2. Root causes of annular gas migration.
al., 1973; Webster and Eikerts, 1979; Sabins et al.,
1982; Bannister et al., 1983; Parcevaux, 1984b;
Beirute and Cheung, 1990; Moroni et al., 1997; Calloni
et al., 1999; Barlet-Gouédard et al., 2001) 9-3.2 Gas migration categories
■ Development of technical solutions (Levine et al., Understanding and modeling the gas migration phenom-
1979; Tinsley et al., 1979; Cheung and Beirute, 1982; enon is difficult because the conduit through which the
Parcevaux et al., 1983; Stewart and Schouten, 1986; gas can channel (i.e., an annular column full of cement,
Sykes and Logan, 1987; Dean and Brennen, 1992) with possibly some spacer and drilling fluid left in the

290 Well Cementing

hole) evolves with time. The physical state of the cement the annulus falls below the formation gas pressure at any
progresses from a liquid slurry during placement to a time, a gas release may occur that, by further relieving
permeable gel during a limited static period, to a per- the hydrostatic pressure, may lead to an irreversible gas-
meable, weak solid when setting, and finally to an imper- entry process. Consequently, the cement-job design
meable solid after hardening. Thus, the physical process should be performed with a free-fall computer simulator
of gas migration is convenient to categorize according to to ensure that the hydrostatic pressure against critical
when it occurs during the cementing operation zones is maintained between the pore and the fracturing
(Fig. 9-3). Three major categories can be defined: pressure at all times during the cement job. An example
1. immediate gas migration (during placement) is shown in Fig. 9-4 (Drecq and Parcevaux, 1988).
It should also be noted that casing reciprocation may
2. short-term gas migration (postplacement)
cause the hydrostatic pressure in the annulus to fall
3. long-term gas migration (postsetting). because of a swabbing effect. This is especially probable

9-3.2.1 Immediate gas migration Placement Pressure Limits

Immediate gas migration, also referred to as gas migra- Fracture pressure
4,000
tion during placement, occurs between the start of the Maximum dynamic
cementing operation and the end of cement placement, 4,500 pressure
which is normally marked when the top wiper plug Minimum hydrostatic
5,000 pressure
lands. Pore pressure
During this time period, gas migration results from 5,500
the loss of hydrostatic-pressure overbalance against gas- 6,000
bearing formations. Preventing gas migration during this 6,500
stage is a relatively straightforward well-control matter, Depth
similar to that practiced during drilling. One of the first (ft) 7,000
approaches to solving immediate gas migration was to 7,500
simply increase the density of the fluid in the annulus. 8,000
Such an approach can be dangerous, because the result-
ing increase in hydrostatic pressure can lead to lost cir- 8,500
culation or formation fracturing. In 1970, Carter and 9,000
Slagle recommended circulating the well before cement- 9,500
Formation
fractured
ing to help remove any trapped gas bubbles that, if not
removed before cement placement, would reduce the 10,000
hydrostatic head of the fluid column. 2,000 4,000 6,000 8,000 10,000
The principal difference between well control during Annular pressure (psi)
drilling and that during cementing is the free-fall or U-
Fig. 9-4. Computer-aided program output (from Drecq and
tubing phenomenon that occurs during the cement job Parcevaux, 1988).
(Chapter 12). Because of the density differences
between the mud, preflushes, spacer, and cement slurry
(or slurries), the hydrostatic pressure exerted at the for-
mation face is not constant during the job (Beirute,
1984; Smith et al., 1987). If the hydrostatic pressure in
et
ug

ts
pl

en
p
m

m
Bu

Ce

Placement Postplacement Postsetting

Minutes Hours Days Months/years

Immediate Short-term Long-term

gas migration gas migration gas migration

Fig. 9-3. Categories of gas migration.

Chapter 9 Annular Formation Fluid Migration 291

during periods when the fluids in the wellbore are in a has increased in recent years because of environmental
static state owing to free fall. As long as the hydrostatic concerns about abandoned gas wells leaking gas into the
pressure in the annulus remains greater than the forma- atmosphere.
tion gas pressure, no gas migration (other than that The primary driver for gas to migrate in the long term
which occurs though dissolution and diffusion processes is the formation of a pathway for the gas to travel after
at the molecular level) should occur during the place- the cement has set. After setting, a normal-density
ment operation. cement system becomes a solid with microdarcy perme-
One final point should be made concerning density ability. As a result, gas cannot migrate at any detectable
control during the cementing operation. Many large rate within the partially water-saturated pores of the
cement jobs are performed on a continuous-mix basis cement matrix. It should be noted that low-density
(i.e., “on the fly”). Density fluctuations may occur during cement systems with high water-to-cement ratios might
the course of the job, resulting in the placement of a exhibit fairly high permeabilities (0.5 to 5.0 mD).
nonuniform column of cement in the annulus Therefore, it is possible for gas to flow, albeit at low
(Granberry et al., 1989). Such a condition may cause rates, within the matrix of such cements and to eventu-
solids settling, free-water development, or premature ally reach the surface. Such events may take weeks or
bridging in some parts of the annulus. When the poten- months to appear as measurable phenomena at the sur-
tial for annular gas migration exists, either batch mixing face, where they usually appear as slow pressure
or process-controlled continuous mixing is recom- buildups in the shut-in annulus.
mended to ensure a homogeneous cement column. The path for long-term gas migration is more likely to
be through a microannulus, a mud channel, a channel of
bypassed lead cement slurry, a free-water channel, a dehy-
9-3.2.2 Short-term gas migration drated filtercake, or any mechanical failure of the cement
Short-term gas migration, also called postplacement gas sheath caused by imposed stresses. Space for entry can
migration, occurs between the end of the primary come from chemical shrinkage of the cement, bulk
cementing operation (normally marked by the landing of shrinkage of the cement, and dehydration of mud chan-
the top wiper plug) and the setting of the cement. Short- nels, free-fluid channels, and filtercakes. Because the
term gas migration may occur anytime between a few cement has already set during this phase, it is no longer
minutes to a few days after the end of the cementing transmitting hydrostatic pressure across the gas zone.
operation. Gas migration that occurs during this time
period is perhaps the most complex to understand, diffi-
cult to predict, and problematic to prevent. For this 9-3.3 Factors affecting gas migration
reason, most industry work concerning gas migration Numerous factors, many of which have been mentioned
has focused on this category. previously, may contribute to gas migration. A synopsis
The primary driver for the occurrence of gas migra- of these factors and how they relate to both the root
tion during this stage is believed to be the decay of the causes and the categories of gas migration is presented
annular pressure. This pressure decay can be attributed in Table 9-1. It is important to note that one single factor
to a combination of several factors, including fluid loss, does not cause gas migration, but rather a combination
gel strength development, and chemical shrinkage of the of several factors, depending upon the unique conditions
cement during hydration. It can also result from the for- in each well.
mation of annular bridges or the setting of mechanical
devices such as packers that isolate hydrostatic-pressure
transmission. Space for gas to enter the annulus during 9-3.3.1 Fluid loss
this stage can be created by fluid loss, free fluid, chemi- Fluid loss from the cement slurry into the formation
cal shrinkage, and the inherent porosity of the slurry. directly affects all three of the root causes of gas migra-
The gas-migration path during the short-term stage is tion and therefore must be considered one of the main
first through the cement-filtercake permeability and contributing factors. First, it may be responsible for a
then through the permeability of the cement matrix. decrease in annular pressure owing to
■ annular bridging

■ increased slurry gelation effects caused by a reduced

9-3.2.3 Long-term gas migration
water content in the slurry
Long-term gas migration occurs after the cement has
■ a decrease in the height of the hydrostatic column
set, which may occur within a few hours after the end of
the cement job. However, days, months, or even years are owing to a slurry-volume decrease
the usual time frame. Industry interest in understand- ■ friction-pressure losses during the compaction
ing, predicting, and preventing long-term gas migration resulting from a slurry volume decrease.

292 Well Cementing

Table 9-1. Factors Responsible for Gas Migration
Annular Pressure –< Pore Pressure Space for Entry Path for Migration
Immediate Hydrostatic underbalance Fluid displaced from wellbore Fluid displaced from wellbore

Short term Fluid loss Fluid loss Slurry permeability

Gel strength development Free fluid Slurry permeability

Chemical shrinkage of cement Chemical shrinkage of cement Filtercake permeability

Annular bridging Slurry porosity Filtercake permeability

Annular packers Slurry porosity Filtercake permeability

Long term Chemical shrinkage of cement Chemical shrinkage of cement Microannulus

Mud channel Mud channel

Free fluid Free-fluid channel

Strength development of cement Dehydrated filtercake Dehydrated filtercake

Bulk shrinkage of cement Bulk shrinkage of cement

Low cement tops

Cement sheath mechanical failure

Second, owing to a volume decrease, fluid loss may into the formation had been neglected. Baret (1988)
create space within the cement matrix that gas can confirmed the critical importance of fluid loss by more
occupy. Finally, fluid loss may be responsible for control- precise direct computations based on Darcy flow. He
ling the filtercake permeability, which ultimately influ- determined that, even in the presence of drilling mud
ences the migration path. Fluid-loss additives may also filtercake, API fluid-loss rates as low as 10 mL/30 min
act indirectly to reduce the permeability of the cement would sometimes be required to prevent annular bridging.
slurry (Chapter 6). It is important to mention that poor fluid-loss control
The importance of fluid loss as a contributing factor across permeable formations further up the hole can
to gas migration was first recognized by Carter and also impair full transmission of the hydrostatic pressure
Slagle (1970). At that time, the respective influences of to a gas zone. In 1976, Garcia and Clark observed gas
fluid-loss control and cement-slurry gelation were not migration when fluid loss occurred high in the hole,
fully understood. It was, however, pointed out that bridg- and hydrostatic pressure was no longer transmitted
ing or gelation owing to fluid loss could restrict the from the column above the bridging point to the bottom
transmission of hydrostatic pressure. In 1975, Christian of the hole.
et al. derived a method for calculating the fluid-loss con- Parcevaux (1987) discussed how fluid loss causes a
trol required to prevent bridging of the cement across pore-pressure decline and the formation of a void space
permeable formations during and after cement place- in the cement. The interstitial water in cement slurry is
ment. Christian et al. concluded that reducing the mobile; therefore, some degree of fluid loss always
American Petroleum Institute (API) fluid-loss rate to occurs when the annular hydrostatic pressure exceeds
less than 50 mL/30 min would reduce gas invasion and that of the formation. The process slows when a low-per-
lessen cement permeability. In 1977, Cook and meability filtercake forms against the formation wall
Cunningham described a procedure to analyze the gas and can stop altogether when the annular and formation
leakage potential based on a similar fluid-loss-rate com- pressures equilibrate. Once pressure equilibrium is
putation. However, Webster and Eikerts (1979) pointed attained, any volume change within the cement will pro-
out that, because earlier work was not based upon flow voke a sharp pore-pressure decline; consequently,
equations, the relative importance of fluid loss may because of the low cement compressibility, a void space
have been overemphasized. The positive influences of forms within the cement matrix, potentially inducing gas
the drilling mud filtercake and mud-particle invasion influx into the cement.

Chapter 9 Annular Formation Fluid Migration 293

 In 2002, Nishikawa and Wojtanowicz discounted the 8,000
validity of previously published pressure-decline theo- 7,000
ries because the correlation with field data was poor. 6,000
They proposed a different model with the following
assumptions. Once the cement is in place in the annulus, 5,000
fluid is initially lost by filtration driven by a hydrostatic Pressure 4,000
(psi)
overbalance. The resulting volume change causes pres- 3,000
sure reduction owing to the compressibility of the 2,000
system. As the volume reduction continues, the cement
1,000
slurry moves downward in plug flow, causing friction
drag at the wellbore walls and an annular pressure 0
decrease. This pressure transient effect is transmitted 0 100 200 300 400
upwards from the zone of fluid loss to the surface as a Time (min)
function of depth and time. A mathematical model was Recorded Predicted
derived for the pressure unloading. at 8,754 ft at 8,754 ft
at 6,909 ft at 6,909 ft
∞ at 5,488 ft at 5,488 ft
0.052
p( h, t ) = 0.052ρ eq h + 2∑ (ρ slurry − ρ eq ) at 4,787 ft
at 4,632 ft
at 4,787 ft
at 4,632 ft
n=1 Dboc at 3,636 ft at 3,636 ft
⎡ (−1) n+1 ⎤
⎢ 2
⎥ sin ( αh e)c2α 2 t
, (9-1) Fig. 9-5. Predicted and recorded loss of hydrostatic pressure at
⎢⎣ α ⎥⎦ depth (from Nishikawa and Wojtanowicz, 2002). Reprinted with
permission of SPE.
where
c = compressibility (psi–1)
thin, low-permeability filtercakes are more effective at
Dboc = Depth at bottom of cement controlling gas migration.
h = depth (cm)
t = time after pumping slurry (min) 9-3.3.2 Gel strength development
α = constant defined as As early as 1970, Carter and Slagle noticed that
thixotropy or gelation of wellbore fluids was relevant to
⎛ π⎞ 1
α = ⎜ nπ − ⎟ the reduction of hydrostatic pressure but provided no
⎝ 2 ⎠ Dboc explanation. Experiments to quantify the effect of gela-
tion on hydrostatic pressure transmission gave inconclu-
ρeq = equivalent density of normal formation sive results (Carter et al., 1973). Some pressure restric-
pressure gradient (lbm/gal) tion was observed at low curing pressures, but
ρslurry = density of the cement slurry (lbm/gal). experiments at greater pressures (500 to 1,000 psi or 3.5
This model was validated in the field when six pres- to 7 MPa) indicated no pressure change. This was prob-
sure gauges were distributed throughout the interval of ably related to deficiencies in the experimental design.
interest in the annulus of a well. The bottomhole pres- It is interesting to note that, much earlier, hydrosta-
sures calculated from the model closely matched the tic pressure reduction during cement hydration had
pressure reduction in the well (Fig. 9-5). been demonstrated in the laboratory and confirmed by
As shown in Chapter 6, cement filtercakes generally field measurements (Guyvoronsky and Farukshin, 1963;
have very low permeability. Bannister et al. (1983) con- Vidovskii et al., 1971) in the former Soviet Union. Similar
cluded that low-permeability cement filtercake deposi- field measurements were performed later by Cooke et al.
tion at the point of gas invasion could hinder gas flow. It (1982), in which the use of external casing sensors per-
is important to note that API fluid-loss tests are per- mitted the observation of downhole temperature and
formed at a fixed differential pressure across a perm- pressure fluctuations as well as the transmission of
eable medium. Thus, according to Darcy’s law, the same applied surface pressure (Fig. 9-6). From this informa-
fluid-loss rate could be measured for two very different tion, it was possible to derive the extent of vertical fluid
slurries: one with a relatively thin, low-permeability movement into the wellbore, locate the top of the
filtercake, and another with a relatively thick, high- cement column, and measure the cement setting time at
permeability filtercake. From the preceding discussion, different depths.
it is clear that low-fluid-loss cement systems that form

294 Well Cementing

 5,000 process involves the breakage of the slurry’s gel strength.
4,000
Sensor depths (ft)
Perforated
However, gas may also flow at the microscopic level within
6,885 ft
6,585 ft the pores of the gelled cement structure (Section 9-3.3.4)
Pressure
3,000 6,659 and 6,585 ft or directly along the cement/pipe and cement/formation
(psi) 5,969 ft

2,000 4,779 ft 6,659 ft interfaces (Section 9-3.3.7). Any or all of these processes
3,459 ft may contribute to the overall phenomenon of gas migra-
1,000 tion, and this limits the applicability of Eq. 9-2.
225 Geothermal Grachyov and Leonov (1969), Parcevaux (1987), and
Temperature 205 6,659 and 6,585 ft • 6,885 ft

(°F) 175 6,885 ft • 6,659 and 6,585 ft

• 4,779 ft
Drecq and Parcevaux (1988) further formalized the pres-
145 3,459 ft 5,969 ft • 3,459 ft sure-reduction process by taking advantage of the simi-
1 2 3 0 0.4 larities between a gelling cement column and a layer of
Time (thousands of minutes) soil undergoing consolidation. Using the theory of soil
mechanics and assuming that the cement slurry behaves
Fig. 9-6. Annular pressure and temperature measurements from as a virgin sedimentary soil before significant hydration
external casing sensors (from Cooke et al., 1982). Reprinted with occurs, the state of stress in the slurry can be described by
permission of SPE.
Terzaghi’s law (Vyalov, 1986). Either U.S. or SI units may
be used in the following two equations, as long as the unit
system is consistent.
In 1982, Sabins et al. related the kinetics of hydrosta-
tic pressure reduction to the cement-slurry gel strength σ = σ ′ + p, (9-3)
development, fluid-loss volume, volume reduction owing
where
to hydration, and slurry compressibility factor. This work
resulted in an empirical method to predict gas migra- p = interstitial (pore) or hydrostatic pressure
tion, and the following equation was derived. σ = total stress exerted at a given linear depth, z
σ′= intergranular or effective stress related to
⎛ 4 × S gel × L ⎞ ⎛ ΔV fl + ΔVhyd ⎞ gel-strength development.
Δp = ⎜ ⎟ or ⎜ ⎟, (9-2)
⎝ dhole − d pipe ⎠ ⎝ fc ⎠ σ is constant and equal to the full overburden pres-
sure exerted by the fluid column.
where
D
dhole = hole diameter (m)
σ = g ∫ ρ slurry ( z )cos θ( z )d x, (9-4)
dpipe = pipe diameter (m) 0
fc = slurry compressibility factor (dimensionless)
where
L = cement column length (m)
D = total linear depth
Δp = hydrostatic pressure change (kPa)
g = gravitational coefficient
Sgel = static gel strength (Pa)
θ = angular deviation
ΔVfl = fluid-loss volume reduction (L)
ρslurry = specific gravity of the slurry at depth z.
ΔVhyd = hydration volume reduction (L).
The effective shear stress, σ′, is related to the static
In 1979, Tinsley et al. introduced the concept of transi-
gel strength determined in the laboratory, using the
tion state, an intermediate period during which the
method described by Sabins et al. (1982) or by Hannant
cement behaves neither as a fluid nor as a solid and the
and Keating (1985), which incorporates the classic
slurry loses its ability to transmit hydrostatic pressure.
shear-stress equation.
The concept of transition state was quantified by a tran-
sition time, beginning with the first measurable gel 4 × L × S gel
strength (about 21 lbf/100 ft2 [10 Pa]), and ending when σ′ = , (9-5)
dhole − d pipe
gas could no longer percolate within the gelled cement.
They showed that a gel strength range from 250 to
500 lbf/100 ft2 [120 to 240 Pa] was sufficient to prohibit where
gas percolation. Gas percolation can be considered a par- dhole – dpipe = width of the annular gap (m)
ticular type of gas migration in which gas in the form of L = length (m)
macroscopic bubbles invades the slurry and rises owing to
Sgel = static gel strength (Pa)
buoyancy effects in accordance with Stokes law. Cement
slurries behave as non-Newtonian fluids; therefore, this σ′ = shear stress (Pa).

Chapter 9 Annular Formation Fluid Migration 295

 Equations 9-4 and 9-5 can thus be combined to obtain Powers studied the shrinkage of pure cement phases
as early as 1935 and found it to increase along the series
4 × L × S gel C2S-C3S-C4AF-C3A from 1% for C2S up to 16% for C3A. He
p = σ − σ ′ = ρ slurry gD cos θ − . (9-6)
dhole − d pipe found that the absolute shrinkage of Portland cement
pastes varies between 2.3% and 5.1%, according to the
following equation.
The hydrostatic pressure, p, exerted by the slurry in
front of the formation varies as a function of the static ΔVab = K hyd1[C3 S] + K hyd 2[C2S] +
gel strength, σ′. However, the exact value of p at time t
may be different from that given by Eq. 9-6 because of K hyd 3[C3 A] + K hyd 4[C 4 AF] (9-9)
kinetic effects.
When gelation occurs during the induction or dor- Powers assumed that, for each type of cement, the
mant period, there is no significant hydration of the shrinkage is a linear function of the percentages of the
cement grains but essentially a buildup of intergranular four major clinker phases. The values Khyd1, Khyd2, Khyd3,
forces owing to interparticle electrostatic forces and the and Khyd4 are coefficients with values varying with the
precipitation of hydrates (Chapter 2). In a first approxi- age (degree of hydration) of the specimen.
mation, the total stress, σ, remains the same, but a In 1982, Geiker and Knudsen found the rate and mag-
transfer from p to σ′ occurs. Eventually, σ′ increases to nitude of the chemical shrinkage to increase slightly
a point at which the cement becomes self-supporting. At with the water-to-cement ratio but the ultimate degree
this time, the interstitial pressure drops to that of a of shrinkage to decrease with increasing curing temper-
column of water, as shown by Eqs. 9-7 and 9-8. ature. The total chemical contraction is split between a
bulk or external volumetric shrinkage, less than 1%, and
p = ρ slurry gD cos θ (9-7) a matrix internal contraction representing from 4% to 6%
by volume of cement slurry, depending upon the cement
σ ′ = (ρ slurry − ρ w ) gD cos θ (9-8) composition (Parcevaux and Sault, 1984). Thus, when
considering cement shrinkage, a distinction should
where always be made between the two types. In most cases,
data reported in the literature refer to total chemical
ρw = the specific gravity of the interstitial water. contraction.
Shrinkage values of less than 4% were reported by
9-3.3.3 Cement shrinkage Chenevert and Shreshtha (1987); however, their experi-
mental design suggests that the phenomenon measured
Cement shrinkage contributes to gas migration by caus- was not the total chemical contraction, but a combina-
ing an annular-pressure reduction and by providing tion of bulk shrinkage and reabsorption of cement free
space for gas to enter the wellbore. When cement enters water. Chemical contraction is a time-dependent phe-
the setting period and hydration accelerates, intergran- nomenon (Fig. 9-7) that begins during the initial setting
ular stresses increase above the value given in Eq. 9-8 and levels off after the final set (Stewart and Schouten,
because of the intergrowth of calcium silicate hydrates. 1986).
Were no volume change to occur at this stage, the pore In 1979, Levine et al. made a significant contribution
pressure would remain at the level given by Eq. 9-7, and by relating shrinkage to pressure reduction. They mea-
the cement would behave as a porous formation. sured the hydrostatic pressure transmission of cement
However, this is not the case. Cement hydration is slurries in a 47-ft long cell with no external pressure
responsible for an absolute volume reduction of the source (Fig. 9-8). They demonstrated that the hydrosta-
cement matrix, also called cement chemical contrac- tic pressure gradient gradually decreases to that of the
tion, which was first identified by LeChâtelier in 1887 mix water. Later, when the cement slurry begins to set,
(Chapter 2; Appendix B). He showed, for normal the hydrostatic pressure quickly approaches zero
Portland cement, a volumetric shrinkage of 4.6%. This (Fig. 9-9). The hydrostatic pressure reduction is the
shrinkage is well documented in the civil engineering lit- result of shrinkage within the cement matrix caused by
erature (Setter and Roy, 1978) and occurs because the hydration and fluid loss; at this point, the fluid column
volume of the hydrated phases is less than that of the ini- above cannot reestablish the pore pressure.
tial reactants.

296 Well Cementing

 7 30
Pressure data
25 150°F laboratory test
6
20 16.4 lbm/gal Class H (neat)
5 Temperature Hydrostatic 15
pressure Static
4 10
(psi)
Shrinkage Total contraction 5 Circulating
(%) 3 0
2 100
1 80
Bulk shrinkage Slurry 60
0 consistency
40
1 5 10 50 100 (Bc) API thickening time
20 150°F atmospheric pressure
Time (hr) 16.4 lbm/gal Class H (neat)
0
Fig. 9-7. Typical contraction and shrinkage (from Parcevaux, 1987). 0 1 2 3 4 5 6
Temperature is 10°C. Time (hr)

Fig. 9-9. Annular gas flow test results (from Levine et al., 1979).
Reprinted with permission of SPE.

Fresh water
10

150°F bath

20
.. p1
. . Pressure
8.34-lbm/gal water
.. . .
. . . transducers and Depth
. . . .
thermocouples (ft)
Tubing

. .. . 30
. .. . .
. . .
. . ..
p2 16.4-lbm/gal cement
Porous plate 40

Meter
4.0 hr

3.0 hr

2.5 hr
2.0 hr
1.5 hr
1.0 hr

0.5 hr
0.2 hr
0 hr

50
Regulator

0 4 8 12 16 20 24 28
Nitrogen Pressure (psi)

Fig. 9-8. Schematic diagram of apparatus to measure hydrostatic pressure transmission of cement
slurries (from Levine et al., 1979). Reprinted with permission of SPE.

Chapter 9 Annular Formation Fluid Migration 297

 Chemical contraction is also responsible for a sec- the existence of free porosity composed of well-con-
ondary porosity, mainly composed of free and conductive nected pores that begin to appear upon the initiation of
pores (Parcevaux, 1984b). At the same time, interstitial the setting period. The same author went on to confirm
water is trapped within the pores through physicochem- that gas migration is driven by an unsteady permeability
ical and capillary forces and can no longer move when effect through the cement pores. After an initial enlarge-
submitted only to its own hydrostatic pressure gradient. ment of the cement pores, a pseudo-steady state is
The combination of chemical shrinkage and secondary achieved when communication has been established
porosity is responsible for the sharp cement pore-pres- throughout the cement column and gas channels have
sure decrease from the water gradient to the formation reached a stable size.
pressure, or even to less than the atmospheric pressure In 1986, Stewart and Schouten confirmed and
if the system is isolated. This was observed by Levine et expanded upon the earlier results of Levine et al. (1979).
al. (1979) and described by Stewart and Schouten They concluded that, when a stable cement slurry (i.e.,
(1986). featuring negligible particle settling) enters the transi-
Contrary to earlier work, Nishikawa and Wojtanowicz tion state, it begins to gel, and the hydrostatic pressure
(2002) reasoned that chemical shrinkage was not decreases ultimately to that of its water phase. When ini-
responsible for annular pressure reduction. Their con- tial setting commences, this pressure (now a pore pres-
clusion was based on experiments involving a solid steel sure) decreases further. In the same paper, Stewart and
rod suspended in a cylindrical container filled with Schouten questioned the validity of static gel strength
cement. The weight of the rod was continuously mea- for describing the potential pressure restriction in
sured as the cement was setting. The theory behind the Eq. 9-1, arguing that this equation assumes the slurry
experiment was that cement shrinkage would instigate a acts as a coherent “one-phase body.” Such an assump-
downward movement of the cement column while gela- tion is valid for pumping applications but not for cases in
tion would increase the friction at the surface of the rod. which the slurry is depressurized internally by fluid loss
Therefore, if gelation and shrinkage occurred concur- or hydration.
rently, the rod would be pulled down by friction, causing
the measured weight of the rod to increase over time. In
the experiments, they did observe a drop in the level of 9-3.3.5 Free fluid (free water)
the cement column; however, they recorded no weight The effect of cement free-water separation was studied
change, leading them to the conclusion that chemical and discussed by Tinsley et al. (1979) and Webster and
shrinkage does not contribute to annular pressure Eikerts (1979). The former concluded through pilot-
decline. This experiment did not take into account the scale experiments that, although undesirable, free water
reduction in pore pressure that would have occurred in is not an influential factor with respect to annular gas
the cement slurry, which would have had an offsetting flow. The latter group studied the problem by construct-
effect on the frictional forces. The experimental setup ing a 9-ft long acrylic model, inclined up to 70° and con-
may also not have had sufficient hydrostatic overburden nected to a gas entry source and several pressure sensors
to cause a significant degree of compaction. (Fig. 9-10). They observed that, in deviated holes, the
free water could coalesce to form a continuous channel
on the upper side of the hole, forming a privileged path
9-3.3.4 Permeability by which the gas may migrate. Thus, they recommended
The concept of gas migration through the pore structure cement slurries that develop essentially no free water.
of a very permeable gelled or set cement, as well as the Despite their observations in the laboratory-scale
potential gas percolation within the gelling slurry that model, Webster and Eikerts were unable to establish a
can occur beforehand, was first introduced by clear relationship between the angle of deviation and
Guyvoronsky and Farukshin (1963). During the period of the importance of the water channel. The nomenclature
hydrostatic pressure reduction, the cement matrix per- for this laboratory test was changed from free water to
meability was measured to be as high as 300 mD. In free fluid in the 1997 edition of API Recommended
1982, Cheung and Beirute proposed a gas migration Practice (RP) 10B, Recommended Practice for Testing
mechanism, based on laboratory experiments, by which Well Cements (Appendix B). The new procedure pre-
the gas first invades cement pore spaces and eventually scribes cement-slurry conditioning in a pressurized con-
permeates the entire cement matrix; consequently, the sistometer at bottomhole circulating temperature
hydration process is prevented from closing the pore before placement in a tube tilted at the same angle as
spaces. Parcevaux (1984b) further refined this mecha- the hole. Webster and Eikerts (1979) and Bergeron and
nism by studying the pore-size distribution of cement Grant (1989) recommended that testing be performed at
slurries during thickening and setting. He demonstrated a 45° angle, the most severe test condition.

298 Well Cementing

 Free water expansion of the casing. Later, when the casing is
opened after the cement has set, the pressure inside the
Channeling effect casing will drop. The heat generated by cement hydra-
tion will also eventually dissipate. Consequently, the
Pressure
measurement
casing diameter will return to its original size.
valve Finally, a bulk volume reduction of the cement sheath
Angle of owing to chemical shrinkage may result in a microannu-
inclination Nitrogen source lus. However, laboratory measurements have shown this
to be negligible under conditions normally encountered
in a cemented annulus (Baumgarte et al., 1999).

9-3.3.8 Cement sheath mechanical failure

Fig. 9-10. Schematic diagram of a model showing fully developed Cracking of the cement sheath caused by imposed ten-
water channeling (from Webster and Eikerts, 1979). Reprinted with
permission of SPE.
sile stresses, compressional stresses, or both (Fig. 9-11)
can form a gas migration path in the annulus. Wellbore
stresses that may be responsible for cement sheath fail-
ure can result from changes in wellbore temperature
9-3.3.6 Mud removal and pressure, tectonic stresses, subsidence, and forma-
tion creep. The relative strength of the formation behind
Early work attributed gas migration problems to poor
the cement sheath is known to have an impact on deter-
mud removal, poor cement/casing or cement/formation
mining whether a given cement sheath will fail when
bonding, or both (Carter and Evans, 1964; Carter and
exposed to wellbore stresses. Formations that are hard,
Slagle, 1970). While several other important contribut-
as characterized by a high Young’s modulus, will confine
ing factors have since been identified, effective mud
the cement sheath and make it less susceptible to crack-
removal is still recognized as imperative for controlling
ing. Formations that are relatively soft, as characterized
annular gas migration. The importance of effective mud
by a low Young’s modulus, will not provide sufficient con-
removal in controlling gas migration cannot be underes-
finement to prevent cracking. A more detailed discus-
timated because, regardless of the properties of the
sion on cement sheath mechanical failure is presented
cement placed in the annulus, a continuous mud chan-
in Chapter 8.
nel between two permeable zones will favor fluid migra-
tion. A detailed discussion on the fundamentals of mud
displacement mechanics and guidelines can be found in
Chapter 5.
Tangential stress Microannulus

9-3.3.7 Microannulus
Tensile
Another path for gas to migrate is through a microan- ΔT failure
nulus—a gap that may form between the cement sheath
and the casing or the formation after the cement has set.
Compressional
A microannulus is a common occurrence that can result Radial stress failure
from any number of events during the life of a well. It has
long been known that a pressure decrease inside the
wellbore after the cement has set will result in a casing- Fig. 9-11. Tensile and compressional failure of the cement sheath
from wellbore stresses.
diameter reduction, leading to the formation of a
microannulus. This commonly occurs when the density
of fluid inside of the casing is reduced after the cement
job. A wellbore-temperature decrease will also reduce
the casing diameter. 9-4 Predicting short-term gas migration
An example of a situation in which both pressure and Gas migration is a complex physical phenomenon that
temperature reduction can occur is when the casing is comprises many facets; as a result, physical modeling of
closed in at the end of the cement job. The exotherm this phenomenon is a formidable problem. It is a non-
generated by cement hydration will cause thermal steady-state process involving changing pressure fields
expansion of the steel casing. In addition, fluids trapped and fluid saturations, and an evolving matrix structure.
inside the casing will heat up, causing further thermal

Chapter 9 Annular Formation Fluid Migration 299

Heterogeneities within the cement slurry, or boundary 9-4.2 Expanding the factors for gas flow potential
effects at the casing or formation, can induce singular Later, a more detailed method developed by Rae et al.
events (such as nonuniform gas breakthrough) that are, (1989) focused on four factors whose components are
by definition, unpredictable. Therefore, it is impossible considered fundamental to the occurrence of gas migra-
to predict the occurrence of gas migration, nor its defin- tion.
itive solution, on an absolute basis. Various prediction
approaches have been developed, but none is universally 1. Formation
applicable. The most common approach to prediction is 2. Hydrostatic pressures in the wellbore
a systematic well analysis, resulting in a purely qualita- 3. Mud removal efficiency
tive risk assignment. 4. Slurry performance
The calculations are based on reservoir deliverability,
9-4.1 Gas flow potential annular geometry, gas-zone pore pressure, hydrostatic
Sutton et al. (1984) described one of the first systematic pressure, mud removal efficiency, cement hydration
approaches to gas migration prediction by calculating a kinetics, and fluid-loss control.
relative gas flow potential (GFP). The GFP is the ratio The Rae et al. (1989) method was not based on a
of the maximum pressure restriction (MPR) to the single experimental investigation or on numerical simu-
hydrostatic overbalance pressure of the well. lations; instead, it was a pragmatic compilation of the
state of the art at the time it was developed. A statistical
K mpr analysis of data from a wide variety of gas wells in the
K gfp = (9-10) United States, Canada, Latin America, Europe, Africa,
pob
the Middle East, and the Far East allowed the calcula-
MPR is further defined as tion of semiempirical relationships between the four fac-
tors. Rae et al. claimed that the wide range of field con-
⎛ ⎞ ditions covered by their method justified its use in most
L
K mpr = 1.67 ⎜ ⎟, (9-11) real cases.
⎝ dhole − d pipe ⎠

where
9-4.2.1 Formation factor
The first factor, the formation factor, is a dimensionless
dhole = diameter of the open hole (in.)
ratio of the reservoir’s productive capacity, kh, to the
dpipe = diameter of the pipe (in.) critical volume, Vcrit (Eq. 9-12). The critical volume is the
Kgfp = gas flow potential additional cement porosity that is created during setting
Kmpr = maximum pressure restriction by chemical shrinkage between the top of the gas zone
and the pressure balance point. A slurry porosity of 2% is
L = cement column length (ft)
assumed at this stage of hydration, and gas is assumed to
pob = overbalance pressure (psi). permeate the annulus in a uniform fashion over the
The GFP is a dimensionless number that can vary defined length. Assuming all other factors remain con-
between 0 and infinity, and the severity of the potential stant, as the value of the formation factor increases, the
gas migration problem is rated according to Table 9-2. risk of gas migration increases.

kh 467.7 khρslurry
fform = = , (9-12)
⎡
) − ( d pipe ) ⎥⎦
2⎤
Table 9-2. GFP Severity Ratings Vcrit
(
pob ⎢ dhole
⎣
2

GFP Severity Rating

<4 Minor where
4–8 Moderate dhole = hole diameter (in.)
>8 Severe dpipe = pipe diameter (in.)
h = zone height (ft)
k = zone permeability (mD)
pob = overbalance pressure (psi)
ρslurry = cement-slurry density (lbm/gal).

300 Well Cementing

9-4.2.2 Hydrostatic factor Table 9-3. Mud Removal Guidelines
The hydrostatic factor (HF) stems from the work of Excellent Moderate
Levine et al. (1979), who observed that the hydrostatic
Hole in excellent condition Hole in excellent condition
pressure exerted by cement slurries decreases, before cementing before cementing
approaching that of the interstitial water as gel strength
increases. They assumed that further pressure decay Circulate one hole volume Circulate one hole volume
occurs only after the cement has set. In theory, when No gas Condition mud
cementing to the surface, gas zones with pore pressures
greater than the hydrostatic pressure of water can flow Condition mud
as soon as the cement gels. When a mud column remains
Greater than 67% standoff Greater than 50% standoff
above the cement, this must be taken into account as an
additional hydrostatic pressure source that is summed Minimal U-tubing Minimal U-tubing
with that of the cement interstitial water. Thus, the
Compatible fluids Compatible fluids
hydrostatic factor is represented by the ratio of the gas-
zone pore pressure to that of the annular pressure at the Use of washes/spacers Use of washes/spacers
commencement of the initial set (Eq. 9-13). Again,
Engineered displacement Engineered displacement
greater values of the hydrostatic factor indicate a
increased risk of gas migration in a given well situation. 10-min spacer contact time 10-min spacer contact time
pgas Use of two bottom plugs Rotation and reciprocation
fhyd = 19.281 ,
) ( )(
of pipe
(
⎡ρ ×h
⎣ mud mud
+ ρ sp × hsp + 8.32 × hcem ⎤
⎦ ) Rotation and reciprocation
of pipe
(9-13)

where calculated from the results of standard API laboratory

fhyd = hydrostatic factor fluid-loss and thickening time tests.
hcem = height of cement in the annulus (ft)
hmud = height of mud in the annulus (ft)
N sp =
q API ( t100 Bc − t30 Bc ), (9-14)
hsp = height of spacer in the annulus (ft) 30
pgas = gas zone pressure (psi)
ρmud = mud density (lbm/gal) where
ρsp = spacer density (lbm/gal). Nsp = slurry performance number
qAPI = API fluid-loss value of slurry (mL/30 min)
t30Bc = time to 30 Bearden units of consistency (min)
9-4.2.3 Mud removal factor
t100Bc = time to 100 Bearden units of consistency (min).
The mud removal factor (MRF) described by Rae et al.
(1989) was a subjective scale from 1 to 10, based upon how The term
closely a defined set of industry standards for mud removal
(Table 9-3) were followed. More quantifiable means of pre- t100 Bc − t30 Bc
dicting mud removal have since been developed
(Chapter 5) that supersede these qualitative standards. is known as the transition time. It must be emphasized
Consequently, the MRF has been largely superseded by that this equation does not represent the actual perfor-
more sophisticated methods that are described later. mance of the slurries under static downhole conditions
in which the mud filtercake influences the leakoff.
Instead, the SPN provides a method of comparing slurry
9-4.2.4 Slurry performance number performance on a relative basis and provides a useful
The final factor considered by Rae et al. (1989) was the tool in both the design and evaluation of cement pro-
slurry performance number (SPN), which was devel- grams for gas wells. Slurries with high SPNs are very
oped to rank cement systems according to their hydra- poor candidates for gas migration control. Those with
tion kinetics and fluid-loss control. It is based on an low API fluid-loss rates and short critical hydration peri-
approximation of the interstitial water lost as the ods offer a much greater probability of success.
cement undergoes the initial hydration process, which is

Chapter 9 Annular Formation Fluid Migration 301

9-4.3 Recent developments in gas migration where
prediction p = pressure (psi)
More recently, improvements have been made to the Rae T = temperature (°F).
et al. (1989) gas migration prediction approach. These
changes involve a more detailed definition of the forma- One can assume that there is a critical volume of
tion parameter (FP), a more quantitative determination space available for the gas to occupy in the annulus
of the mud removal parameter (MRP), and the calcula- before the gas migration problem becomes serious. This
tion of a pressure decay limit parameter (PDLP), which is called the critical annular volume. As previously dis-
replaces the previous HF. Additionally, the new approach cussed, this space is created by fluid-loss, shrinkage, and
does not use an SPN as part of the risk assessment slurry permeability. A conservative assumption is that
process. Instead, it recommends slurry types that have the available space for gas to occupy is 5% of the total
been qualified through laboratory testing based upon slurry volume above the gas zone. Therefore, the critical
the predicted level of risk from the other three parame- annular volume is calculated as 5% of the total slurry
ters (N. Quisel, unpublished results). volume from the top of the gas-bearing zone to the top of
the caprock for that zone.

⎡
)2 − ( d pipe ) ⎤⎥⎦ L
9-4.3.1 Formation parameter
(
2
0.05 π ⎢ dhole
The revised FP is based upon the reservoir’s ability to ⎣
deliver gas. To estimate the produced gas volume, Vgas, Vcrit = , (9-17)
4
across a unit section, the basic steady-state relation for
gases is used. where
dhole = hole diameter (m)
⎡
( ) − ( p ) ⎤⎥⎦
2 2
πkht ⎢ p pore ann
dpipe = pipe diameter (m)
Vgas = ⎣ , (9-15) L = length from top of gas zone to top of caprock (m)
⎛ r ⎞ Vcrit = critical annular volume
μ gas × pann ⎜ ln res + s⎟
⎝ rhole ⎠ 0.05 = the coefficient for space available gas in
the slurry.
where The FP can then be calculated from the ratio of the
h = gas zone length (m) produced gas volume to this critical annular volume by
k = formation permeability (m2) combining Eq. 9-15 and Eq. 9-17.
pann = annular hydrostatic pressure at top of gas Vgas
zone (Pa) K fp =
Vcrit
ppore = pore pressure at top of gas zone (Pa)
(
⎡
) ( ) 2⎤
2
rhole = hole radius (m) 80 kht ⎢ p pore − pann ⎥
rres = reservoir radius (m) = ⎣ ⎦
⎛ ⎞
s = dimensionless skin factor (s = 0 by default). ⎡
( ) ( ) ⎤ r
2 2
⎢ dhole − d pipe ⎥ Lμ gas × pann ⎜ ln res + s⎟
t = production time estimated from setting time (sec) ⎣ ⎦ ⎝ rhole ⎠
Vgas = gas volume, downhole conditions (m3)
μ gas = gas viscosity (Pa-s). (9-18)
The skin factor is highly dependent upon the base where
fluid of the drilling mud (oil or water). A typical skin Kfp = formation parameter.
value is equal to 20.
The gas viscosity, μgas, depends on pressure and tem- Various levels of severity may then be empirically
perature conditions. It is given by Eq. 9-16. This particu- assigned to the value of the FP (Table 9-4).
lar equation is for methane.

( )(
μ gas = 9.76 × 10 −6 + 0.0126 × 10 −6 T + )
{ ( )( )}
p ⎡⎢ 3.22 × 10 −9 − 4.84 × 10 −12 T ⎤⎥ ,
⎣ ⎦
(9-16)

302 Well Cementing

Table 9–4. Severity Ratings Versus FP Cement concentration
FP Severity Rating i = nZ 0% 100%

>1 Very critical

>0.75 to 1 Critical

>0.5 to 0.75 High

>0.25 to 0.5 Moderate

0 to 0.25 Low
i=0
j=0 j = nφ MD
2D Wellbore Concentration
Mesh Grid Versus Depth
9-4.3.2 Mud removal parameter
The second parameter is the MRP. With the advent of Fig. 9-12. Mathematical simulator for the MRP determination.
improved two-dimensional mathematical simulators for
fluid displacement (Chapter 5), the ability to quantify
the effectiveness of mud removal has advanced signifi- causing the annular pressure to drop below the hydro-
cantly in recent years. As shown in Fig. 9-12, such a static pressure. However, the fluid column has lower
mathematical simulator can be used to calculate cement pressure limits at all depths along the wellbore:
concentration on a mesh grid representing a wellbore. ■ pressure applied at the top of the annulus (normally
Perfect displacement would be predicted if a cement atmospheric pressure)
concentration value of 100% were indicated across the ■ pore pressure in front of permeable zones
entire length.
■ vapor pressure of water in front of impermeable zones
Using industry-accepted practices that have been
developed from field experience, positive zonal isolation (including casing-in-casing annuli).
will usually be achieved with 500 ft of cement coverage The pressure cannot drop below these limits at any
above the top of the gas-bearing zone. An MRP can thus depth. Gas migration can occur only when the annular
be calculated over this zone. pressure at a given depth drops to a value equal to or less
Dtog +500 ft
than the pore pressure of a gas-bearing zone at that
depth. The shear stress at the wellbore wall that causes
∫ ( Ycc ) dz ,
1
K mrp = (9-19) the pressure to reach this critical value for gas entry is
h Dbog the PDLP. The following equation uses oilfield units.

where
Dbog = Depth to bottom of gas zone K pdlp =
(
pob dhole − d pipe ), (9-20)
4L
Dtog = Depth to top of gas zone
h = length from bottom of gas zone to 500 ft above where
top of gas zone Kpdlp = pressure decay limit parameter
Kmrp = mud removal parameter L = the length of the cement column above the gas-bear-
Ycc = cement concentration value (calculated by ing formation
mathematical simulator). pob = overbalance pressure at the end of cement
placement, further defined as
Although the MRP is a useful design tool, one should
not forget the importance of the good cementing prac- pob = pmud + psp + pcem + pback − p pore , (9-21)
tices outlined in Chapter 5.
where
9-4.3.3 Pressure decay limit parameter pback = backpressure (i.e., atmospheric pressure + any
applied backpressure)
The third parameter is the PDLP, which is based upon
the concept of critical wall shear stress described by pcem = hydrostatic pressure from the cement column
Stiles (1997). When a fluid is placed inside a pipe or pmud = hydrostatic pressure from the mud column
annulus, a shear stress may be created along the wall,

Chapter 9 Annular Formation Fluid Migration 303

ppore = pore pressure of the gas-bearing formation ■ Reduce the matrix permeability of the cement.
psp = hydrostatic pressure from the spacer/wash column. ■ Increase the rate of gel strength development.
■ Increase the PDLP.
Various levels of severity may be empirically assigned to
the value of the PDLP (Table 9-5). Decreasing the upper boundary limit can be achieved
It is very important to note that, by definition, the by reducing the matrix permeability of the cement
PDLP is a function of density only. Other slurry proper- (Fig. 9-14). This is accomplished by altering the cement-
ties do not apply. The PDLP is entirely defined by the size slurry design, for example by adding small-particle con-
of the annular gap, the hydrostatic contributions of the stituents. However, because the y-axis is logarithmic, it is
fluids in the well, the length of the fluid column, and the apparent that a relatively large decrease in matrix per-
relevant pressure boundaries. Therefore, the only change meability will result in a relatively small decrease in the
in slurry design that will affect the PDLP is density. CHP.
Increasing the slope of the gel strength–develop-
ment curve has a major effect on the CHP (Fig. 9-15).
Table 9-5. Severity Rating Versus PDLP This can also be accomplished by modifying the cement-
slurry design.
PDLP Severity Rating
0–25 Pa (0–50 lbf/100 ft2) Very critical
10,000
25–75 Pa (50–150 lbf/100 ft2) Critical
Impermeable matrix
75–150 Pa (150–300 lbf/100 ft2) High
1,000
150–250 Pa (300–500 lbf/100 ft2) Moderate
Gel strength PDLP Gel strength
>250 Pa (>500 lbf/100 ft2) Low (lbf/100 ft2) 100

10
CHP
9-5 Theoretical strategies for combating
short-term gas migration 1
Critical Upper time
Theoretically, the root causes for gas migration time boundary
(described in Section 9-3.1) can be combated by manag- Time
ing the annular pressure decline, reducing the space for
entry, and minimizing the path for migration. For short- Fig. 9-13. Plot of gel strength development versus time to define the
CHP.
term gas migration, these strategies must be addressed
during the postplacement period.
One concept for evaluating the postplacement period
is to plot the evolution of gel strength over time (Stiles, 10,000
1997). Because gel strength development tends to be
logarithmic, the log of gel strength can be plotted as a 1,000
straight line versus time (Fig. 9-13). The PDLP can be Impermeable matrix
calculated from Eq. 9-20 and plotted on the gel strength
graph so that the intersection of this value with the gel Gel strength 100
(lbf/100 ft2) PDLP
strength curve will define a critical time, tc, when gas
can first enter the annulus. An upper time boundary, tf,
10
represents the time beyond which the gel strength is too CHP
high to allow gas migration. This occurs after the cement
begins to set and become an impermeable matrix. tf is 1
plotted in the same manner as tc, by drawing a vertical Critical Upper time
line from the x-axis to the gel strength–development/ time boundary
impermeable-matrix gel strength intersection. The time Time
between tc and tf is the critical hydration period (CHP).
Fig. 9-14. Strategy for shortening the CHP by reducing the matrix
Stiles recognized three distinct strategies for permeability of the cement.
shortening the CHP and thus reducing the risk of gas
migration.

304 Well Cementing

 10,000 10,000
Impermeable Impermeable matrix
matrix
1,000 1,000
PDLP

Gel strength 100 Gel strength 100

(lbf/100 ft2) PDLP (lbf/100 ft2)

10 10
CHP CHP

1 1
Critical Upper time Critical Upper time
time boundary time boundary
Time Time

Fig. 9-15. Strategy for shortening the CHP by increasing the rate of Fig. 9-16. Strategy for shortening the CHP by increasing the PDLP.
gel strength development.

Finally, the CHP can be shortened significantly by 9-6 Practical solutions for combating
increasing the PDLP (Fig. 9-16). By reviewing Eq. 9-20, gas migration
which defines the PDLP, it can be seen that, unlike the
previous strategies, the PDLP cannot be affected by Practical strategies for combating gas migration can be
modifying any of the cement-slurry properties (with the classified according to the factors previously outlined in
exception of the density). Instead, the size of the annu- Table 9-1. Table 9-6 presents the strategies as a function
lar gap, the hydrostatic contributions of the fluids in the of the three critical root causes and the three time-
well, the length of the fluid column, and the relevant based categories for gas migration.
pressure boundaries control this parameter. Increasing
the size of the annular gap by decreasing the pipe size 9-6.1 Physical techniques
or increasing the openhole diameter is one way to
increase the PDLP. Incidentally, this would also result It has long been known that a number of physical tech-
in a reduction of the FP (Eq. 9-18), further reducing the niques can, under certain circumstances, help control
potential for gas migration. In most cases, however, nei- gas migration. These include applying annular backpres-
ther decreasing the casing diameter nor increasing the sure or small pressure pulses to the annulus, using exter-
openhole diameter would be considered economically nal casing packers (ECPs) and liner-top packers, and
viable options for managing gas migration. reducing the cement column height (including multi-
Increasing the overbalance pressure (Eq. 9-21) stage cementing). Such techniques are certainly valid
would also increase the PDLP. This could be achieved by under a variety of conditions, but well conditions often
increasing the density of any of the wellbore fluids or by limit their application.
increasing the length of a relatively higher-density fluid Application of annular backpressure after the cement
in the annulus. The overbalance pressure is also is in place increases the overbalance pressure exerted on
increased if backpressure is applied. It is important to gas zones, thus delaying the time when gas can enter the
remember that setting an annular packer will have the annulus. However, the presence of weak zones may
opposite effect. The packer will reduce the overbalance restrict this technique because of the risk of inducing lost
pressure and drastically reduce the PDLP. The final way circulation (Levine et al., 1979).
to improve the PDLP is to decrease the length of the Another technique for delaying gas entry, first
cement column; however, this would not be desirable described by Haberman and Wolhart in 1997, involves
because the effective length of the annular seal above applying pressure pulses to the annulus after the cement
the gas zone would be shorter. is in place. The pressure pulses are applied with com-
pressed air or water at approximately 100 psi at a fre-
quency of 30 to 60 sec/pulse. The concept behind this
technique is that the pressure pulses will disrupt gel
strength development in the cement and therefore main-
tain a hydrostatic overbalance for a longer period of time.

Chapter 9 Annular Formation Fluid Migration 305

Table 9-6. Solutions for Prevention of Gas Migration
Root Causes of Gas Migration Pressure applied
Annular Pressure Space for Path for to slurry displacement fluid
–< Pore Pressure Entry Migration
Primary
Immediate Fluid density na† na cement
Short term Right-angle-set Low-porosity Packers
cements cements

Sandwich Low-porosity Sandwich

squeeze cements squeeze

Compressible Compressible Low- High radial effective

Inflation stress applied to
cement cement permeability cement
cements element/formation contact

Fluid density Compressible Surfactants

cement

Thixotropic Low-fluid-loss Thixotropic Inflatable formation

cements cements cements packers, 20- and 30-ft
lengths
Low-fluid-loss Low-fluid-loss Thixotropic
cements cements cements

Back pressure Zero-free- Low-

water permeability
cements filtercake

Annular pressure Zero-free- Low-

pulses water permeability
cements filtercake Bottom plug and
solid wiper plug
Long Term na na Packers

na na Compressible
cements Fig. 9-17. Use of ECPs (from Suman, 1984).

na na Expansive
cements

na na Flexible
annulus shortly after cement placement, thus reducing
cements the hydrostatic overbalance across gas zones below the
ECP. Slurry volume reduction below the packer, from
na na Mud removal fluid loss or chemical contraction, can then result in gas
† na = not applicable invasion of the cement in this interval at an even earlier
time. This could permit undesirable crossflow between
zones located below the packer.
ECPs (Fig. 9-17), inflated by mud or cement slurry, The technique of reducing the cement column height
control gas migration by forming a positive barrier in the stems originally from the work of Levine et al. (1979).
annulus (Suman, 1984). However, ECPs require a com- Viewing the mix-water hydrostatic-pressure gradient as
petent formation to seal against, and they complicate a natural step in the pressure reduction, they proposed
the job execution. Because of the small clearance minimizing the cement column height above the gas
between the uninflated ECP element and the borehole, zone using a very simple graphical method (Fig. 9-18).
such tools have been known to suffer mechanical The job would be designed such that the pressure sum of
damage while casing is being run or during circulation at an equivalent height of water plus the hydrostatic pres-
high rates. Also, it is not uncommon for the packers to sure above the cement would always exceed the forma-
set prematurely because of unexpected pressure fluctu- tion pressure. There is little doubt that this approach
ations during the course of the job. Parcevaux (1984a) can help the design process in a gross sense; e.g., severe
pointed out that ECPs can exacerbate some problems, risks of underbalance may be avoided. It has indeed
because they effectively isolate the lower portion of the been applied with success across some depleted sands,

306 Well Cementing

 0 (Fig. 9-19). Therefore, their ability to compensate for
volume reduction during the transition state is probably
1,000 more effective in situations close to the surface, where
11 lbm/gal mud gas expansion is more significant. In most cases it is
Casing important to maintain the foam quality, or volume of gas
2,000
at 2,000 ft contained, below about 30%; otherwise, the permeability
of the set cement may be sufficiently high to allow gas
3,000 Case 1: Top of migration.
cement at 3,000 ft
The in situ gas generators are designed to maintain
16
Depth 4,000 .6 cement pore pressure by virtue of chemical reactions
lbm
(ft) /ga that produce gas downhole. The produced gases may be
lc
5,000 Formation
em
en hydrogen (Bulatov, 1970; Sutton, 1982) or nitrogen
pore pressure t
Case 2: Top of (Richardson, 1982; Burkhalter et al., 1984). To the
cement at 6,000 ft author’s knowledge, the field application of nitrogen to
6,000
control gas migration has not been reported. Hydrogen-
8.3

generating agents such as aluminum powder have been

4l
bm

7,000 used in the former Soviet Union (Kucyn et al., 1977) and
/ga
lw

elsewhere (Tinsley et al., 1979; Watters and Sabins,

ate
r

8,000 1980). It is important to note that the gas-generating

417 psi Casing at 9,000 ft agents alone cannot prevent gas migration. Fluid-loss
0 1,000 2,000 3,000 4,000 5,000 6,000 control agents and dispersants are necessary to mini-
Pressure (psi) mize interstitial water leakoff.
The principal drawback of these systems, other than
Fig. 9-18. Comparison of cement column height adjustments (from the potential safety hazard from those that generate
Levine et al., 1979). Reprinted with permission of SPE. hydrogen, is the inability of a gas at typical downhole
pressures to achieve the 4% to 6% volumetric expansion
necessary to maintain pore pressure. Strictly applying
but it is clearly not robust enough. As noted in the same Boyle’s law, the volume of gas required to offset just the
paper, as cement changes from liquid to solid, the hydro- chemical contraction would be excessive at high pres-
static pressure falls to values far below that of water sure. Gas-generating systems must also be carefully sta-
because of fluid loss and chemical contraction. bilized; otherwise, gas bubbles may coalesce and create
An elastomeric seal ring, described by Bol et al. channels for formation gas to follow. These limitations
(1986), presents an additional line of defense for inter- notwithstanding, it is clear that this technology has been
facial migration. The success rate may be improved in used with success.
wells in which downhole stresses, such as density
changes or thermal cycling, induce casing deformation.
However, it is important to note that this device cannot 1
solve the problem of gas flow through the cement matrix; 0.9 fgas = 0
thus, it should be used in concert with other techniques. 0.8 fgas = 0.292
0.7
0.6 fgas = 0.488
9-6.2 Compressible cements Volume 0.5 fgas = 0.6
Compressible cement slurries have been developed to (dimensionless) 0.4
maintain the cement pore pressure above the formation 0.3
gas pressure. In theory, this should prevent any move- 0.2
ment of gas from the formation into the cemented annu- 0.1 fgas = 1
lus. Compressible cements fall into two main categories: 0
foamed cements and in situ gas generators. It is impor- 0 100 200 300 400 500 600
tant to draw a clear distinction between the two. Pressure (psi)
Foamed cements (Chapter 7) become nearly incom-
pressible at high pressures because of the relative Fig. 9-19. Compression of foamed cement slurries. fgas is the volu-
incompressibility of gases under such conditions metric fraction of gas in the slurry at surface conditions.

Chapter 9 Annular Formation Fluid Migration 307

9-6.3 Low-permeability cement slurries
Reducing the matrix permeability of a cement system
during the critical liquid-to-solid transition time
described earlier can prevent gas migration. Several
methods have been developed.
The first approach involved the use of water-soluble
polymers to viscosify the interstitial water of the cement
slurry. Because at least a part of gas migration involves
the displacement of cement pore water, viscosification
of the water tends to limit gas mobility. This approach
is also appropriate for fluid-loss control; unfortunately,
viscosification of the cement slurry is a major side effect
of this technique, with resultant mixing difficulties,
greater displacement pressures, and increased risk to
weak formations.
Cheung and Beirute (1982) described the use of low-
permeability cement that operates by immobilizing the Fig. 9-20. Latex film in cement after coalescence.
fluids within the pore spaces of the cement. Because the
cement mix water cannot be displaced, gas cannot move
within the pore spaces of the cement slurry. According to Blomberg et al. (1986) described another technique
Williams et al. (1983), the system is composed of a com- that uses fine mineral particulates to prepare low-den-
bination of bridging agents and polymers. Such systems sity, low-permeability cements. The preferred particu-
have been applied throughout the 140° to 350°F [60° to late in this application is silica fume (also called
180°C] bottomhole static temperature range (Cheung microsilica), a byproduct of the production of silicon and
and Myrick, 1983). ferrosilicon. The average particle size of this material is
Latex additives to prevent gas migration were first 1 μm; consequently, it is able to fill pore spaces and plug
described in a 1982 patent application by Parcevaux et pore throats. Field success has been reported (Grinrod
al. (issued 1985). Subsequent refinements of this tech- et al., 1988) for shallow, low-pressure gas.
nology (Bannister et al., 1983; Parcevaux and Sault, In the late 1990s (Al-Buraik et al., 1998; Moulin,
1984) have extended its applicability to a wide range of 2001), new polymeric microgels were developed to con-
well conditions, and its field application is well estab- trol gas migration at temperatures less than 160°F
lished (Evans, 1984; Peralta, 1984; Matthews and [71°C], at which latexes were not totally effective. The
Copeland, 1986; Rae, 1987; Drecq and Parcevaux, 1988). gas migration control is primarily attributed to the abil-
Latexes are aqueous dispersions of solid polymer par- ity of the microgels to rapidly plug pore throats in the
ticles, including surfactants and protective colloids, cement filtercake, forming an impermeable barrier
which impart stability to the dispersion. Most latexes across the gas zone. Gas migration control is also attrib-
have film-forming capabilities; thus, when contacted by uted to the fact that the microgels are smaller and less
a gas, or when the particle concentration exceeds a dense than cement particles, and they help to reduce
given threshold value, latex particles coalesce to form an pore-throat continuity (and therefore permeability) in
impermeable polymer barrier. In a wellbore, the gas first the cement matrix. As a solid structure develops in the
invades the portion of the cemented annulus across the setting cement, the smaller pore throats reduce the size
gas zone and contacts the dispersed latex particles in of the gas bubbles that can enter the cement, slowing
the slurry. As shown in Fig. 9-20, the latex coalesces their subsequent migration.
within the pore spaces, blocking further progress up the Designing cement systems with a multimodal parti-
annulus. cle-size distribution can result in cement slurries with
Latexes have a number of other beneficial properties considerably less porosity and permeability than a con-
when used in cement slurries (Parcevaux, 1987). The ventional system. The concept, first described by Moulin
small, spherical latex particles act as lubricants, impart- et al. (1997), is based upon maximizing the particle
ing excellent rheological properties. Fluid-loss control is volume fraction in the dry blend by using three or more
provided by a mechanical plugging mechanism. Civil engi- distinct particle sizes. This leads to an increased solids
neers have long recognized the shrinkage-compensating content in the cement slurry and thus a reduced poros-
and bonding actions of latexes. Such attributes translate ity and permeability. More information on cements with
to improved shear-bond strength and elastic deformabil- an engineered particle size distribution can be found in
ity in well cements (Parcevaux and Sault, 1984). Chapter 7.

308 Well Cementing

9-6.4 Fluid loss and free-fluid control additives (Chapter 7). Some degree of fluid-loss control
Fluid loss and free fluid, discussed previously in Sections for thixotropic slurries can also be obtained by the use of
9-3.3.1 and 9-3.3.5, promote the occurrence of gas migra- low-fluid-loss spacer fluids (Bannister, 1978).
tion. To minimize the effect of these parameters, both Successful field results have been obtained in shallow,
must be reduced to fairly low levels. Fluid-loss rates of 50 low-temperature applications (Sepos and Cart, 1985).
mL/30 min or less, and free-water values of 0.25% or less, Stehle et al. (1985) reported good results at greater tem-
have been reported in the literature as requirements for peratures (250° to 280°F [120° to 140°C]) when cement-
slurries used in gas migration situations (Baret, 1988; ing liners and long strings.
Webster and Eikerts, 1979).
Latexes, anionic synthetic polymers, and some cellu- 9-6.6 Surfactants
losic derivatives (at low temperature) are able to pro-
The use of surfactants in cement slurries and preflushes
vide low fluid-loss rates without inducing free-water sep-
was first described by Marrast et al. (1975). Surfactants
aration. However, many of them may affect other
may, under the right circumstances, entrain invading gas
cement-slurry properties, including gel strength devel-
downhole to create stable foam. This foam presents sig-
opment and thickening time, in a deleterious fashion.
nificant resistance to flow of discrete gas bubbles,
Defossé (1985) described a series of metallic salts that
thereby limiting their upward migration through the
depress free-water development, yet are not antagonis-
cement column. Stewart and Schouten (1986) reported
tic to other aspects of slurry performance. A detailed dis-
the technique to be effective, particularly when com-
cussion of fluid-loss additives is presented in Chapter 7.
bined with the use of elastomeric seal rings, described
earlier.
9-6.5 Thixotropic cement slurries
Carter and Slagle (1970) identified slurry gelation as 9-6.7 Right-angle-set cements
a major potential cause of gas migration. However, the
Right-angle set (RAS) cement slurries can be defined as
work of Sabins et al. (1982) indicated that high gel
well-dispersed systems that show no progressive gelation
strength may help resist gas percolation; for this reason,
tendency yet set almost instantaneously because of
they proposed thixotropic and high-gel strength cements.
rapid hydration kinetics. Such systems are capable of
As discussed in Chapter 7, thixotropic cements may
maintaining a full hydrostatic load on the gas zone up to
be prepared by a number of methods, including the addi-
the commencement of cement setting, at which time
tion of bentonite, certain sulfate salts, or crosslinkable
they develop a low-permeability matrix with sufficient
polymers to a Portland cement slurry. In all cases, the
speed to prevent significant gas intrusion. It should be
transmitted hydrostatic pressure of a thixotropic system
pointed out that the transition time involved here is not
should revert to that of its interstitial water and remain
the same as that described by Sabins et al. (1982), nor is
as such until the setting period begins. Therefore,
the mechanism similar to that of high–gel strength sys-
thixotropic systems are unlikely to be effective in situa-
tems (Kieffer and Rae, 1987). Unlike high–gel strength
tions in which the gas-zone pressure exceeds the water
systems, RAS systems undergo a true set involving the
gradient, unless additional backpressure is held on the
deposition and recrystallization of mineral hydrates.
annulus.
RAS slurries are sometimes characterized as such by
The high gel strength of thixotropic cements can offer
standard high-pressure, high-temperature (HPHT)
considerable resistance to physical deformation and per-
thickening time tests, as shown by Drecq and Parcevaux
colation by a large gas bubble. However, as discussed
(1988). A RAS slurry maintains a low consistency until
earlier, the bubbles may often be smaller than the pore
setting, when the slurry viscosity increases to more than
spaces in the setting cement. Under such circumstances,
100 Bc within a few minutes. The increase in consistency
gas migration may occur without slurry deformation, and
is accompanied by a temperature increase resulting
gel strength is no longer an effective barrier.
from the exothermic cement hydration reactions taking
Thixotropic cement slurries tend to have high fluid-
place (Fig. 9-21).
loss rates; therefore, the risk of dehydration and bridg-
It is difficult to design RAS systems for temperatures
ing must be considered. Sykes and Logan (1987) found
below 250°F [120°C] because of the slower cement
the influence of fluid loss to be preponderant to that of
hydration kinetics at lower temperatures. It must also be
gel strength immediately after placement, and they rec-
noted that the shear imparted to the slurry during the
ommended designing the slurry to be well dispersed
API thickening time test varies significantly from that
until after the bulk of fluid-loss volume reduction has
which occurs downhole during a cementing operation,
occurred. The fluid-loss problem has been reduced
making the set profile on a thickening time test merely
somewhat by the development of improved fluid-loss
a qualitative measure.

Chapter 9 Annular Formation Fluid Migration 309

 9-6.9 Flexible cements
Exothermic
308°F reaction
Cements with improved flexibility have been shown to
100 reduce the potential for stress-induced cement-sheath
Viscosity cracking that may lead to long-term gas migration.
(Bc) Thiercelin et al. (1997) showed that the risk of failure is
50 related to the ratio of the cement tensile strength to the
Right-angle
set property Young’s modulus. The larger this ratio (i.e., cement with
relatively high tensile strength and low Young’s modu-
lus), the more resistant the cement is to failure.
1 2 3 4 5
Onan et al. (1993) reported the use of elastomeric
Time (hr) polymers to improve set-cement flexibility. Le Roy-
Fig. 9-21. Pressurized consistometer output from an RAS cement
Delage et al. (2000) incorporated flexible particles such
system (from Drecq and Parcevaux, 1988). as rubber, thermoplastics, or latexes in cement systems
as a method for improving flexibility. Foamed cement
has also been shown to improve flexibility (Deeg et al.,
1999). More detailed information about these methods
9-6.8 Expanding cements can be found in Chapter 7.
Cements that exhibit bulk expansion have been advo-
cated in places in which a microannulus is the gas
migration pathway, and successful field results have 9-7 Laboratory testing
been reported (Seidel and Greene, 1985). As discussed At this writing, neither API nor the International
in Chapter 7, there are two principal techniques for Organization for Standardization have published a stan-
inducing bulk expansion in Portland cement: crystal dard method or procedure for testing gas migration con-
growth and gas generation. The latter operates on the trol in cement slurries; therefore, there is no recognized
same principle as the compressible cements mentioned standard laboratory procedure to characterize the ability
above, with the exception that the concentration of gas- of a cement system to prevent or reduce gas migration.
generating material (typically aluminum) is reduced A significant number of nonstandard laboratory tests
(Sutton and Prather, 1986). The former relies upon the for gas migration has been developed by the industry. A
nucleation and growth of certain mineral species within wide variety of experimental prototypes that attempt to
the set-cement matrix. The bulk volumetric expansion simulate the gas migration process are described in the
that can be attained through crystal growth is usually literature. Two main types of experimental simulators
controlled to be less than 1% (Griffin et al., 1997). The exist: large-scale pilot devices, which attempt to repro-
bulk expansion provided by either of these mechanisms duce the process as it occurs in the wellbore, and small-
does not eliminate internal chemical shrinkage that has scale, benchtop models, which can be used to derive the
previously been discussed as a cause for gas migration. It fundamental laws of a particular physical process under
only acts to increase the bulk or dimensional volume of investigation. To date, none of the simulators described
the cement. in the literature permits the derivation of a physical
Experimental work performed by Beirute et al. model that quantitatively describes gas migration over a
(1992) found that expanding cements are not effective wide range of conditions. Recently, many investigators
at sealing microannuli along the casing/cement inter- have begun to rely upon measuring individual slurry
face if the adjacent formation does not impose an ade- properties such as gel strength development and shrink-
quate confining stress. Their laboratory studies indi- age. These data are then used either as input into math-
cated that the cement sheath can actually expand away ematical simulators or for comparing the performance
from the casing when confined by soft formations. differences between various slurries.
Further experimental work by Baumgarte et al. (1999)
and Le Roy-Delage et al. (2000) confirmed this phenom-
enon, leading to recommendations to use expanding
cements only across relatively hard formations that can
resist the expansion and impose stresses in the cement
sheath that force it to expand towards the casing.

310 Well Cementing

9-7.1 Large-scale testing Computer Gas-pressure Gas
Typically, large-scale models have been constructed in an system flowmeter inlet
effort to better understand gas migration phenomena
and are not practical for conducting routine experiments Gas Heated water out
for individual well designs. The earliest large-scale simu- Cement out
Top
lation was performed by Carter and Slagle (1970) and Flexible
later upgraded by Carter et al. (1973). In 1976, Garcia polymer plug
and Clark constructed a large-scale device specifically to Pressure and (optional)
temperature
study the influence of uneven cement setting. Levine et transducers Cement
al. (1979) described a simulator for studying hydrostatic
pressure profiles within a cement column at rest
Middle 10 ft
(Fig. 9-8). The apparatus built by Tinsley et al. (1979)
investigated the influence of fluid loss and compared dif-
ferent cement systems (Fig. 9-22). Finally, the equip- Water jacket
ment described by Bannister et al. (1983) evaluated the
influence of filtercake formation from cement fluid loss
and the conductivity to gas of setting cement (Fig. 9-23).
Cement in
Bottom
Heated water in
To constant-pressure 200-mesh screen
water source 300-mL void
48.3 cm Gas-pressure
Chamber charged with Gas
water to 3,447.5 kPa or flowmeter inlet
6,895 kPa
Fig. 9-23. Annular gas flow laboratory testing apparatus (from
Slurry-fill 2.1 m
Fluid-loss vent Bannister et al., 1983). Reprinted with permission of SPE.
vent valve
holes when using
permeable section

Permeable or 45.7 cm
9-7.2 Small-scale gas migration testers
nonpermeable
section Various bench-scale or benchtop devices for characteriz-
ing gas migration have been described in the literature.
The first, described by Cheung and Beirute (1982), used
Check valve 1.0 m a modified API fluid-loss cell to investigate the hydro-
Hot-water jacket static pressure decrease and subsequent gas migration
Gas-entry line
from volume in a setting cement column (Fig. 9-24). Adaptations of
325-mesh screen measurement this model have recently been made commercially avail-
device
able and are known as gas migration analyzers or gas
Rubber diaphragm
Slurry 49.5 cm flow analyzers (Fig. 9-25). The test cell is a modified API
fill line HPHT fluid-loss cell with a hollow hydraulic piston at the
To pressure
recorder Thermocouple top of the cell. The piston is pressurized with mineral oil
To temperature recorder to simulate the hydrostatic overbalance. Fluid loss can
occur at both the top and the bottom of the cell, either
Fig. 9-22. Schematic diagram of test fixture used to study gas through standard 325-mesh fluid-loss screens or through
leakage (from Tinsley et al., 1979). Reprinted with permission of SPE.
actual formation cores. Once the cement slurry has
developed a predetermined gel strength, gas is injected
into the bottom of the cell at an appropriate differential
pressure. Gas flowmeters and pressure transducers mea-
sure any gas migration through the slurry. A separate gel
strength–development test (described in Section 9-7.3)
must be performed on the slurry before performing this
gas migration test.

Chapter 9 Annular Formation Fluid Migration 311

 While this device has been adapted for routine use, it Stewart and Schouten (1986) investigated gas migra-
is important to note that, at the scale of this test, three tion in set and hard cement using a U-tube apparatus
factors can unduly affect the gas migration process. like that shown in Fig. 9-26 (Richardson, 1982).
Fluid loss could result in the formation of an impenetra- Parcevaux (1984a) and Drecq and Parcevaux (1988)
ble filtercake at the gas inlet or outlet. Free-water devel- described a small-scale simulator that eliminated some
opment could artificially delay the pore pressure of the limitations of earlier devices. As illustrated in
decrease by reabsorption during hydration. Finally, in Fig. 9-27, the artificial effects of fluid loss and free water
view of the length of the cement column versus the were eliminated, and the external curing pressure was
external applied pressure, such an experiment can only computer-controlled to maintain a differential pressure
consider gas migration across a short interval. close to zero between the top and bottom of the cell.

Nitrogen gas
Top valve Backpressure receiver
Nitrogen gas

Pressure To temperature controller

Mineral
regulator
oil tank Piston
O-ring Thermocouple

Pressure transducer

325-mesh screen
Slurry
To recorder

Nitrogen gas
Bottom valve Gas Pressure
flowmeter regulator

Figure 9-24. Gas flow simulator (from Cheung and Beirute, 1982). Reprinted with permission of SPE.

Fig. 9-25. Gas flow apparatus (photo courtesy Chandler

Engineering LLC).

312 Well Cementing

 Gas Another bench-scale device called the cement hydra-
source tion analyzer (CHA) was first described in a 1997 API
Pressure
gauges Technical Report (TR) 10TR2, Technical Report on
V Shrinkage and Expansion in Oilwell Cements. This
device was developed to measure the evolution of criti-
V Oil cal parameters during cement hydration and to better
discharge
Standpipe understand the mechanisms of gas entry downhole.
U-tube Under fixed-pressure conditions, the CHA measures
Gas hydration rate, shrinkage (with regard to chemical con-
column Oil column traction or development of porosity as the slurry
Water hydrates), and mechanical properties.
column The gas migration cell is an 8-in. long, 2-in. diameter
Cement slurry disposable cylinder (Fig. 9-28). The top of the cell is
closed and contains a pressure transducer and a tem-
V V perature probe. At the bottom of the cell, a sliding piston
(or diaphragm), connected to a gas source, is used to
Water p pressurize the slurry. The rate of the gas entering the
discharge cell and the rate of the water pushing the piston as
shrinkage develops are carefully monitored. The cell is
placed in a thermostatic oven until thermal equilibrium
is reached. As long as the cement slurry remains liquid,
Water the pressure measured at the top of the cell stays con-
reservoir stant. The upward movement of the piston compensates
Fig. 9-26. U-tube gas migration tester (from Richardson, 1982). for any slurry shrinkage. As the slurry hydrates, the tem-
perature in the cell increases and shrinkage continues.
The piston movement is eventually stopped by the
Valve Bleedoff Thermocouple increasing wall shear stress of the slurry, and the pres-
N2 pressure
source sure at the top of the cell decreases. When the pressure
at the top of the cell falls to a predetermined value, the
gas valve connected to the piston is opened. The gas is
Valve
allowed to enter the cell, driven by the pressure
Head- decrease and the continuing shrinkage of the cement.
pressure The permeability of the slurry to gas can be deter-
regulator
Hydrostatic mined by analyzing the pressure and gas flow rate
Cement pressure (Fig. 9-29). If the cement is permeable to gas, the pres-
columns cell sure in the cell will reach equilibrium with the gas-inlet
Dynamic
permeability
pressure once the gas valve is opened, and gas flow can
cell Differential be measured by the gas flowmeter. The amount of gas
pressure that enters the cell gives a good indication of the amount
transducer of shrinkage. A visual observation of the cement after
the test can indicate the manner by which gas migrated
Backpressure Thermostatic oven (bubbles, micropercolation, or fracture). If the cement
regulator Pressure
transducer is impermeable to gas, the pressure in the cell will con-
tinue to decrease and no gas flow will be observed.
Flowmeter N2 pressure source

Fig. 9-27. Dynamic permeability apparatus (from Parcevaux, 1984a).

Reprinted with permission from Elsevier.

Chapter 9 Annular Formation Fluid Migration 313

 Thermocouple
Pressure
gauge

Cement
slurry

Water flow

Gas and water

pressure

Fig. 9-28. Cement hydration analyzer. Reprinted with permission from Elsevier.

100 100

80 95
Pressure

60 90
Pressure Temperature
(bar) Temperature (C°)
40 85
Gas valve
opened
20 80
Shrinkage
Gas flow rate
0 75
0 4 8 12 16 20 24
Time (hr)
100 100

80 95

60 90
Gas rate Temperature
Temperature
(scf/min) (C°)
40 85

Gas valve
20 80
opened
Shrinkage
0 75
0 4 8 12 16 20 24
Time (hr)

Fig. 9-29. Output from the CHA. Cement that is permeable to gas (top) will exhibit a stable pressure
in the cell equal to the gas-inlet pressure, and gas flow will be measured with the gas flowmeter,
while cement that is impermeable to gas (bottom) will exhibit a continued pressure decrease
caused by shrinkage and no gas flow will be measured.

314 Well Cementing

9-7.3 Gel strength testing A similar modification of this method is known as the
Numerous methods for measuring the static gel strength shearometer tube. This device consists of a thin-walled
(SGS) of cement slurries have been developed. Today tube that is placed in a sample of cement slurry and
there are four common methods: using pressure-drop allowed to rest statically for a given period of time.
tubes, using vane rheometers, using rotating-paddle Weights are then applied to the top of the tube until it
devices, and employing acoustic techniques. It is impor- begins to move, and the SGS is calculated based upon
tant to note that comparative testing between these the force imparted by the weights and the surface area
devices has shown significant variations in the measured of the tube. This test can only be carried out at atmos-
SGS. As noted by Prohaska et al. (1993), gel strength pheric pressure.
development can be significantly impacted by the
amount of shear that is applied to the slurry before com- 9-7.3.2 Vane rheometers
mencing measurement and the pressure and tempera-
Early SGS testing was attempted using standard rota-
ture that the slurry encounters. These factors, along with
tional viscometers by measuring the maximum deflec-
the inherent differences in the devices, account for the
tion of the bob while rotating at 3 rpm after the slurry
variability in results between the various methods.
had been static for a designated period of time (normally
1 min or 10 min) (Appendix B). Slippage at the interface
9-7.3.1 Pressure drop tubes between the bob and the cement slurry, coupled with the
One of the simplest means of measuring SGS is by mea- relatively high minimum shear rates (5.1 sec–1), gave
suring the pressure drop across a length of tubing unreliable results. Furthermore, this test could only be
(Fig. 9-30). Cement slurry is placed in a small-diameter carried out under ambient conditions. In an effort to
tube and pressurized with water. A sensitive pressure reduce interfacial slippage, Haimoni and Hannant
transducer measures the pressure drop in the cement (1988) replaced the cylindrical bob with a vane. In the
column as it undergoes gelation. Equation 9-20 can then late 1990s, the vane geometry was further advanced. The
be used to calculate the SGS. One limitation of this test improved device allowed testing at elevated tempera-
is that it is not able to differentiate between pressure ture and pressure. In addition, the device allowed one to
drop caused by gelation and pressure drop caused by impart stress into the static slurry without shearing it
chemical shrinkage. and disturbing the evolving gel structure (Fig. 9-31).

9-7.3.3 Rotating-paddle devices

Constant Sabins et al. (1980) developed a device designed specif-
pressure ically to test SGS at elevated temperature and pressure.
source
The device, similar to a HPHT consistometer, used a spe-
cially designed paddle placed inside a pressurized slurry
container. The slurry could be stirred with a low-friction
Cement-filled tube
magnetic drive to simulate placement. After stirring, a
cord-and-weight bucket was attached to the magnetic
Δp drive, and the weight required to rotate the paddle 10°
Temperature bath
was measured at specific time intervals (normally 5 to
15 min). The SGS could then be calculated from the
geometry of the paddle and the weight required to move
it. Further improvements to this device, which included
direct torque measurement, were patented by Moon et
al. (Fig. 9-32) (1986).
Screw pump

Fig. 9-30. Schematic diagram of pressure-drop tube apparatus for

measuring SGS.

Chapter 9 Annular Formation Fluid Migration 315

Fig. 9-31. Pressurized vane rheometer to measure SGS.

Pulley rotates 0.50–2°/min

paddle rotational speed
Transducer Acoustic
pulsing waveform
Force electronics capture
transducer
Cement
slurry
Digital
Stepping signal
Sample motor processing
shearing
Torque
Measurement
Calculation
of signal
of SGS value
amplitude
Time
Fig. 9-32. Rotating-paddle SGS measuring device (from Moon and Fig. 9-33. Static gel strength analyzer block diagram (from Sabins
Wang, 1999). Reprinted with permission of SPE. and Maki, 1999). Reprinted with permission of SPE.

9-7.3.4 Acoustic devices 9-8 Conclusions

As a cement sample develops SGS, the amplitude of an Gas migration is a complex phenomenon involving fluid-
acoustic signal passing through the sample increases. density control, mud removal, cement-slurry properties,
Correlations can be developed between the change in cement hydration, and interactions between the
amplitude of the acoustic signal and the SGS of the cement, casing, and formation. It has been the subject of
slurry. Using this concept, Sabins and Maki (1999) pro- a substantial amount of research over the past 50 years,
duced a device similar to the ultrasonic cement analyzer yet it remains one of the most serious cementing prob-
(Appendix B) that could measure the SGS (Fig. 9-33). lems the industry faces. Uncontrolled, gas migration can
An apparent advantage of this method is that the lead to loss of production, environmental and regulatory
slurry is not mechanically sheared; therefore, the SGS is problems, and ultimately risk to equipment and person-
measured at zero shear rate. However, it is important to nel. The cost of remediating a gas migration problem can
note that, similar to that of the ultrasonic cement ana- far exceed the costs of active prevention.
lyzer, the gel strength is not a direct measurement, but Gas migration can be categorized according to when
only a correlation that has been established by compari- it occurs in relationship to the cement job. Gas migra-
son of acoustic signal attenuation and testing with pres- tion that occurs during cementing can be referred to as
sure-drop tubes and shearometers. immediate gas migration; gas migration that occurs after

316 Well Cementing

the cement is in place but before it has set is referred to 9-9 Acronym list
as short-term gas migration; and gas migration that API American Petroleum Institute
occurs after the cement has set is referred to as long-
term gas migration. Regardless of when it occurs, there CHA Cement hydration analyzer
are three distinct root causes that must occur for annu- CHP Critical hydration period
lar gas to migrate. ECP External casing packer
1. The pressure in the annulus must fall to a value that FP Formation parameter
is less than or equal to the pore pressure of a gas- GFP Gas flow potential
bearing zone.
HF Hydrostatic factor
2. Space must be available in the annulus for gas to
occupy in order for it to enter the annular column. HPHT High pressure, high temperature
3. A path must be present in the annulus through which MRF Mud removal factor
the gas can migrate. MPR Maximum pressure restriction
Successful strategies, both theoretical and practical, MRP Mud removal parameter
for controlling gas migration rely on minimizing the PDLP Pressure decay limit parameter
impact of one or more of these root causes. Particular RAS Right-angle set
factors that should be addressed include fluid loss, gel RP Recommended practice
strength development, cement shrinkage, cement per-
meability, free fluid, cement hydration kinetics, mechan- SGS Static gel strength
ical properties of the set cement, and last, but not least, SPN Slurry performance number
proper mud removal. TR Technical report
Because the factors surrounding gas migration are so
complex, it must be understood that there is no single
solution for all cases. Numerous predictive models
designed to understand and assess the risk of gas migra-
tion for a given set of conditions have been developed.
These models, coupled with the appropriate laboratory
techniques, aid the engineering and design of cementing
programs that will significantly minimize the likelihood
of annular gas migration.

Chapter 9 Annular Formation Fluid Migration 317

 Erik B. Nelson and Véronique Barlet-Gouédard—Schlumberger

10-1 Introduction
High-temperature wells present special cement system
design challenges. The physical and chemical behavior
Thermal Cements

10-2 High-temperature chemistry

of Portland cement
10
As discussed in Chapter 2, Portland cement is essentially
of well cements changes significantly at elevated tem- a calcium silicate material, the most abundant compo-
peratures and pressures. One must also pay close atten- nents being tricalcium silicate (C3S) and dicalcium sili-
tion to the chemical and physical properties of the for- cate (C2S). Upon addition of water, both hydrate to form
mations in contact with the cement. Corrosive water a gelatinous calcium silicate hydrate called “C-S-H
zones and very weak formations are not uncommon in phase,” which is responsible for the strength and dimen-
thermal wells. Without careful slurry design, the sional stability of the set cement at ordinary tempera-
integrity of the set cement may deteriorate, potentially tures. The reaction also liberates a substantial amount
resulting in the loss of zonal isolation. of calcium hydroxide (CH).
Thermal cementing encompasses three principal C-S-H phase is the early hydration product even at ele-
types of wells: deep oil and gas wells, geothermal wells, vated temperatures and pressures, and it is an excellent
and thermal recovery wells. In this chapter, each sce- binding material at well temperatures less than 230°F
nario is discussed separately, because the cement [110°C]. At higher temperatures, C-S-H phase is subject
system design parameters can differ significantly. to metamorphism, which usually results in decreased
During the last 50 years the most commonly used compressive strength and increased permeability of the
cements to complete thermal wells have been Portland set cement. In the petroleum literature, Swayze (1954)
cement, silica-lime systems, and high-alumina cement. described this phenomenon as “strength retrogression.”
More recently, other cement systems have been devel- At temperatures above 230°F [110°C], C-S-H phase
oped specifically for thermal-well environments. Before often converts to a phase called alpha dicalcium silicate
discussing the cement system design for the various hydrate (α-C2SH). α-C2SH is highly crystalline and much
types of thermal wells, the hydrothermal chemistry of more dense than C-S-H phase. As a result, matrix shrink-
cements used to complete thermal wells is presented. In age occurs that can be deleterious to the set-cement
this chapter, the special chemical notation for cement integrity. This effect is illustrated in Fig. 10-1, which
compounds is used. The reader is referred to Chapter 2 depicts the compressive-strength and water-permeabil-
for an explanation of the customary abbreviations. ity behavior of conventional Portland cement systems

10
40
4
1
30 3 1

Compressive 2 Water 2
strength 20 permeability 0.1
(MPa) 1 (mD) 3
10 0.01

4
0 0.001
0 1 0 1
Curing time (months) Curing time (months)

Fig. 10-1. Compressive strength and water permeability behavior of Portland cement at elevated
temperatures (from Nelson and Eilers, 1985).

Well Cementing ■ Chapter 10 Thermal Cements 319

cured at 446°F [230°C]. Significant loss of compressive the conditions for the formation of various calcium sili-
strength occurred within 1 month; nevertheless, the cate compounds, many of which occur geologically
levels to which the compressive strength fell are suffi- (Taylor, 1964). The C/S ratio is plotted versus the curing
cient to support casing in a well (Suman and Ellis, 1977). temperature. C-S-H phase has a variable C/S ratio, aver-
The real problem lies in the severe permeability aging about 1.5. The conversion to α-C2SH at 230°F
increases. To prevent interzonal communication, the [110°C] can be prevented by the addition of 35% to 40%
permeability of well cements to water should be no more silica (by weight of cement [BWOC]), reducing the C/S
than 0.1 mD. Within 1 month, the water permeabilities ratio to about 1.0. At this level, a mineral known as tober-
of the normal-density Class G systems portrayed in Fig. morite (C5S6H5) is formed; fortunately, this mineral pre-
10-1 (Curves 1 and 2) were 10 to 100 times higher than serves high compressive strength and low permeability.
the recommended limit. The permeability of the high- As the curing temperature increases to about 300°F
density Class H system (Curve 3) was barely acceptable. [150°C], tobermorite normally converts to xonotlite
The deterioration of the lower-density extended cement (C6S6H) and a smaller amount of gyrolite (C6S3H2) with
(Curve 4) was much more severe. minimal deterioration of cement performance.
The strength retrogression problem can be prevented Tobermorite sometimes persists to 480°F [250°C] in
by reducing the bulk lime-to-silica ratio (C/S ratio) in Portland cement systems because of aluminum substitu-
the cement (Menzel, 1935; Kalousek, 1952; Carter and tion in the lattice structure (Kalousek and Chow, 1976).
Smith, 1958). To accomplish this, the Portland cement is The improved performance of “silica-stabilized”
partially replaced by ground quartz, usually as fine silica Portland cements at elevated temperatures is illustrated
sand or silica flour. In some regions, special cements are in Fig. 10-3. Normal-density Class G cements, stabilized
available in which quartz has been ground with Portland with silica sand or silica flour, were cured at 446° and
cement clinker (Italcementi, 1977). Figure 10-2 depicts 608°F [230° and 320°C].

Mole fraction CaO/CaO + SiO2 of starting material

0 0.3 0.4 0.5 0.6 0.7 0.8 1.0
1,000 Rankinite α-CaSiO4
α-CaSiO3
Quartz

800 β-CaSiO3 Kilchoanite

(+ silica)
600 Calcio-
chondrodite
400 Foshagite Phase Y
Cristobalite

300 Tricalcium
Truscottite Xonolite
Calcium hydroxide
Temperature silicate
(°C, log scale) γ-C2S hydrate hydrate
200
Gyrolite Hillebrandite
150
Z-phase 11 Å Tobermorite α-C2S hydrate
100 Afwillite
Hydrous

14 Å Tobermorite
silica

75
Ca(H3SiO4)2 C-S-H(I) C-S-H(II)
50
0 0.5 0.6 0.7 0.8 1.0 1.3 1.5 2.0 2.5 3.0
CaO/SiO2 mole ratio of starting material

Fig. 10-2. Formation conditions for various calcium silicates (from Taylor, 1964).
Reprinted with permission from Elsevier.

320 Well Cementing

 Silica sand (230°C)
Silica flour (230°C)
Silica flour (320°C)
60 1

40 0.1
Compressive Water
strength permeability
(MPa) (mD)
20 00.01

0 0.001
0 1 3 6 12 24 0 1 3 6 12 24

Fig. 10-3. Compressive strength and permeability behavior of 16.0-lbm/gal Class G systems stabilized
with 35% silica (from Nelson and Eilers, 1985).

At 480°F [250°C] the phase truscottite (C7S12H3) The discussion so far has been limited to the behavior
begins to appear. As the curing temperature approaches of the silicate hydration products. To the authors’ knowl-
750°F [400°C], both xonotlite and truscottite are near edge, the hydrothermal behavior of the aluminates and
their maximum stable temperatures, and dehydration of aluminoferrites has not been specifically described in
the residual CH to C occurs. At higher temperatures, the the literature. The common hydrated aluminate and alu-
xonotlite and truscottite dehydrate, resulting in the dis- minoferrite hydrates (Chapter 2) are not typically
integration of the set cement. observed when Portland cements are cured hydrother-
In addition to the compounds cited above, other mally. Ettringite is not stable in hydrothermal condi-
–
phases such as pectolite (NC4S6H), scawtite (C7S6CH2), tions, and is not normally detected. Some of the Al3+,
reyerite (KC14S24H5), kilchoanite (C3S2H approxi- Fe3+, and SO42– ions from ettringite are incorporated
mately), and calciochondrodite (C5S2H approximately) into the silicate phases.
may appear in Portland cement systems cured at ele- The preceding discussion illustrates the complexity of
vated temperatures. These phases can affect the perfor- the hydrothermal behavior of calcium silicate hydrates.
mance of the set cement, even when present in small The performance of the set cement depends not only on
quantities. the downhole temperature, but also on the presence of
Cements containing significant amounts of truscot- subterranean brines and other minerals. As a result, the
tite are usually characterized by low permeability standard conditions for equilibrium transformations
(Gallus et al., 1978). The formation of pectolite, a that are reported in the literature are not always
sodium calcium silicate hydrate, is accompanied by observed downhole (Langton et al., 1980). Therefore,
cement expansion (Nelson and Eilers, 1982); in addition, the set cement must be considered to be metastable,
pectolite appears to render cements more resistant to because its composition can evolve as downhole condi-
corrosion by highly saline brines (Nelson and Kalousek, tions change.
1977; Nelson et al., 1981). Scawtite has been shown to
enhance cement compressive strength when present in
minor amounts (Eilers et al., 1983; Bell et al., 1989). In
general, set cements that consist predominantly of cal-
cium silicate hydrates with C/S ratios less than or equal
to 1.0 tend to have higher compressive strengths and
lower water permeabilities.

Chapter 10 Thermal Cements 321

10-3 High-alumina cement 50
High-alumina cement is a special material manufac-
tured primarily for applications in which a refractory 40
binder is required (Robson, 1962; Scrivener and Compressive
30
Capmas, 2001). In wells, it is used where the in situ strength
combustion process is employed for fireflooding (MPa)
20
(Section 10-8) and is also useful for cementing across
permafrost zones (Chapter 7). The primary cementitious 10
constituent is monocalcium aluminate (CA). As illus-
0
trated in Fig. 10-4, three initial metastable calcium 200 400 600 800 1,000 1,200
aluminate hydrates occur when water is added to CA:
Temperature (°C)
CAH10, C2AH8, and C4AH13. They ultimately convert to
C3AH6 (Quon and Malhotra, 1979). In addition to C3AH6, Fig. 10-5. Compressive strength of high-alumina cement/crushed
aluminum hydroxide (AH3) is often present as a binder firebrick concrete after 4 months’ exposure from 68° to 2,190°F [20°
phase. Unlike Portland cement, set calcium aluminate to 1,200°C] (from Heindl and Post, 1954).
cement does not contain calcium hydroxide.

C12A7 crystals intergrow and form a tightly bonded

27°C ceramic network. In thermal wells, temperatures above
CA + H20 CAH10, C2AH8, C4AH13 C3AH6 1,830°F [1,000°C] are not generally attained; thus, it is
(metastable) important to ensure that the minimum compressive
225°C
550°C to strength is sufficient for maintenance of well integrity.
to 275°C The strength and durability of high-alumina cements
950°C between 440° and 1,830°F [225° and 1,000°C] are pri-
C +C12 A7 C3 AH1.5 marily controlled by the initial water-to-cement ratio.
Depending upon the application, the amount of added
Fig. 10-4. Sequence of reactions of high-alumina cement at various
curing temperatures (from Nelson and Eilers, 1985). Reprinted with water should be the minimum required to prepare a
permission from the Petroleum Society of CIM. pumpable slurry. The use of dispersants is particularly
helpful for pumpability. A higher proportion of cement
relative to aggregate extender is also necessary. For
C3AH6 is probably the only stable hydrated calcium most applications, at least 50% of the solids should be
aluminate at temperatures below 437°F [225°C]. At cement.
higher temperatures, the water content begins to drop, A variety of materials may be used to extend calcium
and at 527°F [275°C], C3AH1.5 is found. As the temper- aluminate cement slurries, provided they have suitable
ature continues to increase, decomposition of C3AH1.5 stability at high temperatures and do not decompose or
occurs with the liberation of C. Between 1,022°F show anomalous thermal expansions or inversions. Silica
[550°C] and 1,742°F [950°C], a recrystallization occurs, sand should not be used if temperatures exceeding
ultimately resulting in C and C12A7. 572°F [300°C] are anticipated. Because of changes in
It is important to realize that, in ultrahigh-tempera- the crystalline structure, thermal expansion of quartz is
ture wells with temperatures up to 930°F [500°C], the relatively high at these temperatures, and thermal
proportional strength loss of high-alumina cement sys- cycling could eventually disrupt the cement. The most
tems is often greater than that experienced by unstabi- commonly used extender for high-alumina cements is
lized Portland cements. High-alumina cement is used crushed aluminosilicate firebrick. Other suitable materi-
because it can withstand wide-ranging temperature fluc- als include calcined bauxite, certain fly ashes, diatoma-
tuations, owing mainly to the absence of calcium hydrox- ceous earth, and perlite.
ide. Figure 10-5 illustrates the effect of curing tempera-
ture upon a high-alumina cement extended with 70%
crushed firebrick (Heindl and Post, 1954). The initial 10-4 Class J cement
strength loss between room temperature and 212°F Class J cement was developed in the early 1970s for
[100°C] is primarily caused by the conversion of the ini- cementing wells with static temperatures exceeding
tial hexagonal calcium aluminate hydrates to the cubic 260°F [126°C] (Maravilla, 1974; Degouy and Martin,
C3AH6. With further heating, the strength continues to 1993; Bensted, 1995). It was recently dropped from the
drop because of dehydration and the formation of C and list of American Petroleum Institute (API) cements
C12A7. Strength improves above 1,830°F [1,000°C] as the

322 Well Cementing

owing to low usage; however, it is still manufactured in environment, the stable minerals would be anorthite,
the Far East, mainly for geothermal well applications. A grossular, prehnite, epidote, and zeolite, possibly associ-
similar cement, belite-silica cement (BSC) has been ated with quartz (Frey et al., 1991). Roy et al. experi-
used in the CIS for high-temperature well cementing mented with compositions that would lead to
(Bulatov, 1985). These cements are advantageous from a the formation of the mineral anorthite. The raw materi-
logistical point of view, because the addition of silica is als were Class J cement, calcium aluminate cement, and
not required. β-C2S. The results were encouraging in terms of
Like Portland cement, Class J cement is a calcium sil- strength; however, the preparation of mixable and
icate material; however, no aluminate phases or C3S is pumpable slurries, with controllable set times, was a
present. The composition is essentially β-C2S, α-quartz, serious problem.
and CH. As discussed in Chapter 2, the hydration rate of More recently, other approaches to preparing anor-
β-C2S is relatively slow; consequently, retarders are thite cements were investigated (Barlet-Gouédard and
rarely necessary at circulating temperatures less than Goffé, 2002). To promote the formation of anorthite, var-
300°F [149°C]. The C/S ratio of Class J cement is ious alumina-rich materials were evaluated in combina-
adjusted such that tobermorite and xonotlite are tion with Class G cement and silica flour. Calcined kaoli-
obtained upon curing (Sasaki et al., 1986; Kalousek and nite was determined to be an ideal alumina source;
Nelson, 1978). Scawtite also occurs frequently. In addi- however, it must be very finely divided (≈2 μm) to pro-
tion, the sulfate resistance of Class J cement is very high vide sufficient reactivity to readily form anorthite. As a
owing to the absence of C3A. result, similar to the experience of Roy et al., Barlet-
Gouédard and Goffe initially encountered difficulties
preparing a mixable and pumpable system. Their solu-
tion involved controlling the particle sizes of the ingre-
10-5 Calcium aluminosilicate systems dients and optimizing the packing between them. Such
In the late 1970s, Roy et al. (1979; 1980) reported work controlled-granulometry techniques are described in
to develop cement systems that resemble the composi- detail in Chapter 7. The anorthite system is a trimodal
tion of the formation in geothermal wells. Such cements mixture: calcined kaolinite and silica flour (fine parti-
should exhibit better chemical and thermal durability cles), Class G cement (medium-size particles), and
than the conventional calcium-silicate or calcium-alumi- ceramic microspheres (coarse particles). The micros-
nate systems. Of particular interest was the pheres are also useful for lowering slurry density.
CaO • Al2O3 • SiO2 system (Fig. 10-6). In a hydrothermal

Si
100 Anorthite system (1.49 g/cm3)
10 Class G cement
90 Anorthite
20 Xonotlite
80 Garnet
30 Kilchoanite/afwillite
70 Pectolite
40 Prehnite
60 Wairakite
50 Epidote
50 Quartz
60 Diopside
40 Hauyne
70 Margarite
30
80
20
90
10
100

100 90 80 70 60 50 40 30 20 10
Al Ca

Fig. 10-6. Ternary diagram showing compositions of minerals in the CaO • Al2O3 • SiO2 system.

Chapter 10 Thermal Cements 323

 A complete solid-solution series extends from anor- calcium aluminate cement as the base reactant. The
thite, Ca(Al2Si2O8), to albite, Na(AlSi3O8). As a result, principal cementitious phase is hydroxyapatite
anorthite can accommodate the substitution of various [Ca5OH(PO4)]. Such systems normally set and harden
ions into its crystal lattice. This property makes anor- within a few minutes at ambient temperature. Citric
thite adaptable to aggressive chemical environments. acid is an effective set retarder at circulating tempera-
The typical properties and performance of an anorthite tures up to around 180°F [82°C] (Brothers et al., 1999).
cement exposed to a carbon dioxide–laden synthetic Calcium phosphate cement normally has a density in
geothermal brine are given in Table 10-1. the range of 15–17 lbm/gal [1,800–2,040 kg/m3]. To reduce
the cost or the density of the slurry, extenders are usually
added. The most common extenders include ASTM Class
F fly ash or nitrogen, which is used to prepare a foamed
10-6 Calcium phosphate systems cement system (Brothers et al., 2001). The composition
Calcium phosphate cement systems were originally and compressive-strength performance of foamed cal-
developed for applications in dentistry. Recently such cium phosphate systems are shown in Table 10-2.
cements have found application as thermal cements The carbon dioxide resistance of calcium phosphate
(Sugama and Carciello, 1992; 1993; 1995), particularly cement, compared to conventional silica-stabilized
for geothermal wells in which high concentrations of Portland cement, is shown in Table 10-3. The cement sys-
carbon dioxide are encountered (Weber et al., 1998; tems were cured in 1-wt% Na2CO3 solution at two tem-
Brothers et al., 2001). Portland cement systems are sub- peratures: 200° and 500°F [93° and 260°C]. The solution
ject to attack by CO2, resulting in the loss of cementi- pH was 2.0–2.7. The weight loss of the cement specimens
tious material from the cement matrix (Chapter 7). The was measured at curing times of up to 32 days. The
consequences of CO2 attack are loss of compressive results clearly show that, within the 32-day test period,
strength and an increase of permeability. the calcium phosphate binder demonstrated excellent
Calcium phosphate cements can be synthesized by resistance to carbonation.
the acid-base reaction between a soluble phosphate
[usually NaH2PO4 or (NaPO3)n] as the acid solution and

Table 10–1. Typical Performance of Anorthite Cement Systems

Slurry Density Plastic Yield Point Thickening 28-D Compressive Water Permeability
(lbm/gal [g/cm3]) Viscosity (cp) (lbm/100 ft2) Time at 300°F Strength in Brine at at 28 D in Brine at
[150°C] BHCT † BHST ‡ of 572°F [300°C] BHST of 572°F [300°C]
(hr:min) (psi [MPa]) (mD)
10.9 [1.31] 75 0.6 5:00 1,400 [9.8] 0.2

12.4 [1.49] 130 6 3:00 2,320 [16.2] 0.05

13.5 [1.62] 144 6 5:40 2,200 [15.4] 0.03

14.0 [1.68] 110 4 3:30 2,570 [18.0] –

16.2 [1.95] 55 10 7:40 2,290 [16.0] –

† BHCT = bottomhole circulating temperature
‡ BHST = bottomhole static temperature

Table 10–2. Composition and Performance of Calcium Phosphate Cement Systems Cured at 600°F †
Water Calcium Sodium Fly Ash Foaming Foam Slurry Density 28-Day
(wt%) Aluminate Polyphosphate (wt%) Agent Stabilizer (lbm/gal Compressive
(wt%) (wt%) (wt%) (wt%) [kg/cm3]) Strength
(psi [MPa])
23.3 17.5 15.6 40.8 1.9 0.9 12.1 [1,450] 570 [4.0]

22.3 21.0 14.9 39.2 1.8 0.8 15.1 [1,810] 1,060 [7.4]
† from Brothers et al. (1999).

324 Well Cementing

Table 10–3. Corrosion-Test Results of Thermal Cements in 1% Na2CO3 Solution; pH = 2.0–2.7†
Slurry Slurry Test Percent Weight Loss After:
Density Temperature
(lbm/gal) (°F)
2D 9D 13 D 20 D
Calcium phosphate 15.7 200 0.1 0.1 0.1 0.1

Class G + 40% silica 16.2 200 1.6 5.7 11 21

18 D 25 D 32 D
Calcium phosphate 16.0 500 0 0 0

Class G + 40% silica 16.2 500 9.5 15 22

† from Weber et al. (1988).

10-7 Deep oil and gas wells 10-7.1 Thickening time and initial compressive
Wells with depths exceeding 15,000 ft [4,570 m], with strength development
bottomhole temperatures above 230°F [110°C], are Cement slurries for deep wells are usually designed to
common throughout the world. Since the 1970s, hun- have at least 3 to 4 hr of pumping time. However, there
dreds of wells with depths exceeding 25,000 ft [7,600 m] are several complicating factors that must be men-
have been completed (Arnold, 1980; Wooley et al., 1984). tioned.
Such wells represent a large investment of time and As the length of the casing string or liner increases,
money; therefore, obtaining a successful well completion the problem of achieving a cement seal becomes more
is of paramount importance. difficult (Suman and Ellis, 1977). In many cases, the
The procedures for cementing deep wells are basi- static-temperature differential between the top and
cally the same as those for shallower wells; however, bottom of the cement column can exceed 100°F [38°C].
because of the severe well conditions and more complex Sufficient retarder must be added to the cement slurry
well architecture, such wells are usually considered to to allow adequate placement time at the BHCT; conse-
be critical (Smith, 1987). Higher temperatures, nar- quently, such a slurry may be over-retarded at the top of
rower annuli, overpressured zones, and corrosive fluids the cement column, resulting in a very long waiting-on-
are commonly encountered. Consequently, the cement cement (WOC) time. If high-pressure gas exists behind
system design can also be complex, involving an elabo- the casing string or liner, the risk of gas invasion into the
rate array of retarders, fluid-loss additives, dispersants, cement is high (Chapter 9). In recent years, advances
silica, and weighting materials. One must be certain that have been made in retarder chemistry and cement-
the cement system can be properly placed and will main- system design that have helped to mitigate such prob-
tain zonal isolation throughout the life of the well. At lems (Chapter 3).
present, Portland cement is used in virtually all deep oil When designing cement slurries for deep, hot wells, it
and gas well completions. is very important to use accurate static and circulating
Typical casing programs and cementing procedures temperature information. Such data may be obtained
for deep wells are given in Chapter 13. Detailed informa- from drillstem tests, logs, special temperature recording
tion regarding cement additives is found in Chapter 3. In subs, or circulating temperature probes run during hole
this section, the design of appropriate cement systems conditioning (Jones, 1986). Computer simulators have
for deep high-temperature wells is presented. also been developed to better predict well temperatures
(Chapter 12). If fluids are circulated in the well for sev-
eral hours before cementing, the well temperature may
be lowered significantly. In such cases, one must be
careful not to overestimate circulating temperature and
over-retard the cement slurry.

Chapter 10 Thermal Cements 325

 The cement slurry is exposed to high pressures in 10-7.3 Cement-slurry density
deep wells and, as shown in Fig. 10-7, a significant accel- Deep wells often involve cementing across high-pressure
erating effect is observed (Bearden, 1959). Earlier com- formations. To maintain control of the well, the hydro-
pressive strength development and higher ultimate static pressure of the wellbore fluids must meet or
compressive strength are also observed as curing pres- exceed the formation pressure at all times. Con-
sure increases (Handin, 1965; Metcalf and Dresher, sequently, cement slurries with densities as high as
1978). Therefore, when designing a proper cement- 27 lbm/gal [3.24 g/cm3] are often placed. When large
slurry composition in the laboratory, performing the quantities of weighting materials are present in the
tests at the anticipated pressure is recommended slurry, slurry sedimentation can be a major concern.
(Appendix B). Maximizing the packing volume fraction of the slurry
In general, as the circulating temperature increases, solids, using a multimodal particle size distribution, can
the sensitivity of Portland cement systems to subtle mitigate such problems (Chapter 7). Silica can be incor-
chemical and physical differences between the slurry porated in one of the granulometric fractions to provide
ingredients also increases. Therefore, all laboratory tests stabilization.
should be performed with samples of the water, cement,
and additives that will be used during the job.
10-7.4 Fluid-loss control
As discussed in Chapters 3 and 6, fluid-loss control is
600 necessary to preserve the chemical and physical charac-
teristics of the cement slurry and to prevent the devel-
500 opment of a cement filtercake that could cause bridging
in the annulus. For most primary cementing operations,
400
an API/ISO fluid-loss rate of between 50 and 100 mL/30
Thickening
time 300 min is generally considered to be adequate.
(min)
200
10-7.5 Long-term performance of cements
100 for deep wells
Once the cement system is successfully placed in the
annulus, it is important to ensure that it will provide
0 5,000 15,000 25,000 35,000 adequate casing support and zonal isolation throughout
Pressure (psi) the life of the well. As discussed earlier in this chapter,
Fig. 10-7. Effect of pressure on pumpability of cement (cement: API
the most common thermal cement systems are based on
Class H with 0.3% retarder; BHCT: 200°F [93°C]) (after Smith, 1976). silica-stabilized Portland cement.
Reprinted with permission of SPE. A typical slurry composition for a deep, hot well would
consist of Class H or Class G cement, 35% to 40% silica
(BWOC), a dispersant, a fluid-loss additive, a retarder,
and a weighting agent. The long-term performance of
10-7.2 Cement-slurry rheology such cement systems would be very similar to that
The narrow annuli often associated with deep-well com- shown in Fig. 10-3.
pletions increase the difficulty of achieving a good bond When high-density slurries are unnecessary, or if
between the cement, the pipe, and the formation. The lower-density slurries are required to prevent lost circu-
small clearance between the casing and the formation lation or formation breakdown, extenders such as fly
increases the risk of cement contamination by drilling ash, diatomaceous earth, bentonite, or perlite are com-
fluid that was not displaced by chemical washes or spac- monly used. The long-term performance of typical sys-
ers. Proper casing centralization can be difficult to tems in laboratory tests is illustrated in Figs. 10-8 and
achieve. As discussed in Chapter 5, the rheology of the 10-9. All systems contained 35% silica flour (BWOC). In
fluids in the wellbore should be carefully controlled. In Fig. 10-8, the systems have been cured at 450°F [232°C]
some cases, the cement slurry is designed for turbulent- under saturated steam pressure for up to 2 years, and
flow displacement, requiring the use of dispersants. compressive strength and permeability measurements
When designing highly dispersed slurries, one must be have been performed at periods ranging from 1 day to
careful to avoid slurry sedimentation or free-water devel- 24 months. Figure 10-9 presents data for systems cured
opment. This is especially important when the borehole at 600°F [315°C]. It is important to note the nonlinear
is highly deviated (Chapter 13). time scale and the logarithmic permeability scale.

326 Well Cementing

 Fly ash: 15.6 lbm/gal [1.87 g/cm3] Fly ash: 15.6 lbm/gal [1.87 g/cm3]
Perlite/bentonite: 12.9 lbm/gal [1.55 g/cm3] Perlite/bentonite: 12.9 lbm/gal [1.55 g/cm3]
Perlite/bentonite: 11.9 lbm/gal [1.43 g/cm3] Perlite/bentonite: 11.9 lbm/gal [1.43 g/cm3]
Diatomaceous earth: 13.8 lbm/gal [1.66 g/cm3] Diatomaceous earth: 13.8 lbm/gal [1.66 g/cm3]

50 50

40 40
1
1
30 Compressive 30
Compressive
4 strength
strength 2
(MPa) 20 (MPa)
20
2
3
10 10
3 4
0 0
0.03 1 3 6 12 24 0.03 1 3 6 12 24
Curing time (months) Curing time (months)
100 100

10 10
1

1 1
Water Water
permeability 3 permeability 3
(mD) 2 (mD) 0.1
0.1
1 4 2
0.01 0.01
4
0.001 0.001
0.03 1 3 6 12 24 0.03 1 3 6 12 24
Curing time (months) Curing time (months)

Fig. 10-8. Compressive strength and permeability performance of Fig. 10-9. Compressive strength and permeability performance of
conventionally extended Portland cement slurries—450°F [232°C] conventionally extended Portland cement slurries—600°F [315°C]
after Nelson and Eilers, 1985). Reprinted with permission from the (after Nelson and Eilers, 1985). Reprinted with permission from the
Petroleum Society of CIM. Petroleum Society of CIM.

System 1 contained Type F fly ash as an extender and Systems 2 and 3 were extended with perlite and ben-
was the heaviest of the four. Despite the density advan- tonite. System 2 performed well at both 450° and 600°F
tage and the highest initial compressive strength, the [232° and 315°C] with regard to compressive strength.
performance of System 1 over a 2-year period was The permeability of System 2 varied back and forth
no better than that of lower-density systems at 450°F across the 0.1-mD line. System 3 was the least dense of
[232°C], and was the poorest of the four at 600°F the four. The compressive strength performance was
[315°C]. This delayed degradation of fly-ash-containing adequate at both curing temperatures, but the perme-
systems was probably the result of alkali contaminants abilities were too high. It is important to point out that
in the fly ash. Such contaminants can slowly react and perlite is compressible, and its extending effect
form substituted calcium silicate hydrates, notably decreases as the hydrostatic pressure in the well
reyerite, with deleterious effects (Eilers and Root, 1974). increases (Chapter 3). For this reason, perlite is rarely
It is important to mention that cement degradation asso- used today. System 4, containing diatomaceous earth,
ciated with fly ash has not been observed at curing tem- was a rather poor performer in the strength category, yet
peratures below 450°F [232°C]. had low permeability.

Chapter 10 Thermal Cements 327

 Figure 10-10 shows the typical performance of a If competent cement systems with densities less than
normal-density neat Class J system. Its behavior is simi- 12.5 lbm/gal [1.5 g/cm3] are necessary, microsphere-
lar to that observed with normal-density silica-stabilized extended, multimodal-particle-size, or foamed cements
Portland cement systems. (Chapters 3 and 7) may be appropriate. However, when
The behavior of these systems illustrates that high contemplating the use of ceramic or glass microspheres,
compressive strength and low water permeability are not one must be certain that they can withstand the hydro-
necessarily linked. Although water permeability is not as static pressure. Ceramic microspheres and most grades
convenient to determine as compressive strength of glass microspheres can withstand no more than
(Appendix B), one should do so before the application of 3,000 psi [20.7 MPa], which eliminates them from con-
a cement system in severe downhole conditions. In addi- sideration in most deep well completions. However, glass
tion, the data suggest that conventionally extended microspheres with hydrostatic crush strengths as high
Portland cement systems with densities below about 12.5 as 10,000 psi [69.0 MPa] are available. Foamed cement,
lbm/gal [1.5 g/cm3] may not be able to perform suitably occasionally used in deep high-temperature wells, is
in high-temperature wells, except perhaps as “filler” sys- more common in geothermal and steamflood wells.
tems that are not placed across producing zones.

10-8 Geothermal well cementing

6,000
Projects to extract geothermal energy exist throughout
the world. Virtually any location with thermal anomalies
is a potential site for geothermal well drilling. Some of
the more notable geothermal projects are located in
4,000 California, Utah, and New Mexico, United States;
Mexico; Central America; The Philippines; Indonesia;
Compressive Slurry density New Zealand; Kenya; Iceland; and Italy (Geothermal
strength 16 lbm/gal Education Office, 2001).
(psi) [1.92 g/cm3] Most geothermal plants are configured as shown in
2,000 Fig. 10-11. Superheated formation water that lies above
the geothermal formation is produced to the surface,
whereupon it is “flashed” into steam. The steam is used
to power turbines that generate electrical power. The
spent water is injected back into the reservoir, not for
0 replenishment, but for environmental reasons. The for-
0 1 3 6 12 24
mation waters are often highly saline and corrosive and
Curing time (months) contain toxic heavy metals. Another plant design, called
1
“hot dry rock,” is employed in regions where no geother-
mal formation waters exist. Two intersecting wells are
drilled into the hot formation. Water from the surface is
pumped down one well and becomes superheated. The
superheated water is then produced out of the other well
0.1 and flashed into steam.
Geothermal wells are usually completed in much the
Water same manner as conventional oil and gas wells; however,
permeability the environment with which the cements must contend
(mD)
is frequently much more severe. The bottomhole tem-
0.01 perature in a geothermal well can be as high as 700°F
[370°C]. The failure of wells in several geothermal fields
has been directly attributed to cement failure (Radenti
and Ghiringelli, 1972; Berra et al., 1998); as a result,
research has been conducted to identify cement formu-
0.001
0 1 3 6 12 24
lations that perform suitably under such conditions.
Curing time (months)

Fig. 10-10. Compressive strength and permeability behavior of Class

J cement at 450°F [232°C].

328 Well Cementing

 Injection well

Production well

Fig. 10-11. Geothermal power plant (from Geothermal Education Office, 2001). Drawing courtesy
of Geothermal Education Office, Tiburon, California, USA.

10-8.1 Well conditions associated with An economical geothermal reservoir requires that
geothermal wells large quantities of hot water or steam must be produced
With the exception of hot, dry rock completions with cir- from each well. Therefore, the reservoirs are usually nat-
culating temperatures as high as 500°F [260°C] (Carden urally fractured and have effective permeabilities that
et al., 1983, Duchane, 1994), most geothermal wells are are probably greater than 1 D. The integrity of the for-
not cemented under “geothermal” conditions, because mations ranges from poorly consolidated to highly frac-
the fluids circulated during drilling cool the formation. tured, and the fracture gradients tend to be low.
The maximum circulating temperatures during the Consequently, lost circulation is the most serious obsta-
cement job seldom exceed 240°F [116°C]; therefore, the cle to successfully cementing geothermal wells (per-
design of cement systems with adequate thickening sonal communication, Weber, 2003). It is not uncommon
times is usually not a problem. Most geothermal wells to have losses in the casing strings set above the target
are less than 10,000 ft [3,050 m] in depth. Downhole reservoir, and in many cases total losses occur before the
pressures are seldom above the water gradient. intended setting point for the intermediate string. For
The drilling programs for geothermal wells usually these reasons, low-density cement systems are required
call for setting surface and production casing above the by most geothermal operators (Nelson et al., 1981).
reservoir. In some cases, a slotted liner is hung through Lost circulation also hampers the determination of
the producing zone, but cementing the liner is not con- cement placement temperatures. Placement tempera-
sidered critical. It is very important to cement the cas- ture simulation and modeling is essential to formulate
ings to the surface; otherwise, creep or elongation will the appropriate cement system (Chapter 12).
occur because of thermal expansion when the well is
brought into production (Shryock, 1984).

Chapter 10 Thermal Cements 329

 The chemistry of the reservoir fluids varies from fresh 8,000 Mesh 325 140 50
water to brines with greater than 200,000 mg/L total dis-
solved solids. The fluids extracted from dry steam fields 6,000 50,000
contain relatively few salts and low concentrations of Compressive
noncondensible gases, the most noticeable being H2S. strength 4,000 30,000 (kPa)
The saline brines often contain significant quantities of (psi)
2,000 10,000
carbonate and sulfate.

1 4 10 40 100 400
10-8.2 Performance requirements and design
Average silica size (μm)
considerations
Geothermal wells arguably present the most severe con- 300°F [150°C]
ditions to which well cements are exposed. As a result, 450°F [232°C]
617°F [325°C]
the performance requirements are among the most
stringent. Geothermal well cements are usually designed
to provide at least 1,000 psi [7.0 MPa] compressive
strength, and no more than 0.1-mD water permeability Mesh 325 140 70
100
(API Task Group on Cements for Geothermal Wells, 80
1985). In addition, the set cement often must be resis-
Crystalline 60
tant to degradation by saline brines. composition
Silica-stabilized Portland cement compositions are 40
(%)
almost exclusively used to complete geothermal wells; 20
however, their dominance is being challenged by sys- 0
tems that offer better resistance to the severe chemical 20 40 60 80 100 120 140 160 175
environments. Each is described in this section. Average silica size (μm)
Scawtite Kilchoanite
10-8.2.1 Portland cement–based geothermal well Xonotlite Quartz
cement compositions
When Portland cement–based cement systems are
Mesh 325 140 50
expected to contact highly saline and corrosive geother- 10.0
mal brines, the particle size of the added silica is an 4.0
important consideration. As explained in Chapter 3, 1.0
there are two forms of silica commonly used in well Water 0.4
permeability 0.1
cementing: silica sand, with a particle size of approxi- 0.04
(mD)
mately 175–200 μm, and silica flour, with an average par- 0.01
ticle size of approximately 15 μm. Field personnel usu- 0.004
ally prefer silica sand, because its lower surface area 0.001
facilitates easier slurry mixing. However, in certain geot- 1 4 10 40 100 400
hermal environments, silica sand cannot be relied upon Average silica size (μm)
to provide adequate stabilization.
Eilers and Nelson (1979) investigated the effect of Fig. 10-12. Effect of silica particle size on the performance of Class
G cement cured in geothermal brine (from Eilers and Nelson, 1979).
silica particle size on the performance of Class G cement Reprinted with permission of SPE.
formulations cured at various temperatures in a geo-
thermal brine. The salinity of the brine was 25,000 mg/L
total dissolved solids. Figure 10-12 shows the relation-
ships between the silica particle size and several para- Figure 10-13 shows that the silica particle-size effect is
meters—compressive strength, water permeability, and significantly more pronounced with lower-density
cement phase composition. The slurry density was cement compositions.
15.8 lbm/gal [1.90 g/cm3]. A decrease in compressive High concentrations of sodium chloride depress the
strength and an increase in water permeability occurred rate at which silica enters solution (personal communi-
when the average particle size of the added silica cation, R. Fournier, 1979); as a result, when the silica
exceeded about 15 μm. Xonotlite was also replaced particle size is large, the rate of dissolution of silica is
by kilchoanite as the predominant cement phase. insufficient to allow the formation of the desired calcium

330 Well Cementing

 Compressive strength Permeability
Mesh 325 140 50
4,000 10 Mesh 325 140 50
300°F [150°C]
450°F [232°C] 4
617°F [325°C] 1
3,000
0.4
Compressive
2,000 Permeability 0.1
strength
(mD)
(psi) 0.04
1,000 0.01
0.004
0.001
1 4 10 40 100 400 1 4 10 40 100 400
Average silica size (μm) Average silica size (μm)

Fig. 10-13. Effect of silica particle size on the performance of a 13.5-lbm/gal Class G-perlite-bentonite
system cured in geothermal brine (from Eilers and Nelson, 1979). Reprinted with permission of SPE.

silicate hydrates (C/S ratio < 1). The kinetics of dissolu- 40% SiO2
tion can be affected by the particle size of the solute. 70 100% silica fume
Reducing the particle size of the silica increases its sur- 100% silica flour
60 50% silica fume, 50% silica flour
face area; consequently, a sufficient supply of silica is 33% silica flour, 67% silica fume
available. 50
25% silica flour, 75% silica fume
Grabowski and Gillott (1989) and Dillenbeck et al. Compressive
(1990) studied the effects of silica “fume,” with an aver- strength 40
age particle size of approximately 0.1 μm (Chapter 3), (MPa)
upon Portland cement systems at elevated temperatures 30
and pressures. With a constant SiO2 concentration (40%
20
BWOC) and water-to-solids ratio (0.5), samples were
prepared containing silica fume, combinations of silica 10
fume and silica flour, and silica flour. Curing was per- 7 28 56 90 210 270
formed at 450°F [230°C] and 400 psi [2.75 MPa] for
Total age (days)
7 days, using samples aged under ambient conditions for
periods up to 270 days. The systems containing silica 10–1

fume developed less compressive strength, but lower

permeability, than equivalent systems containing only
silica flour (Fig. 10-14). The major phase found in all of
the samples was xonotlite (scawtite was detected in the
samples containing only silica flour); however, the Permeability
(mD)
microstructures were different. The samples containing
silica flour exhibited short parallel needles of xonotlite.
10–2
As the quantity of silica fume increased, the texture of
the xonotlite was granular. In general, the samples with
needle-shaped xonotlite crystals exhibited higher per- 5 × 10–3
meabilities.
The presence of carbonate in certain geothermal 7 14 28 270
brines presents a serious difficulty for Portland cement Total age (days)
systems (Milestone et al., 1986 and 1987). Calcium sili- Fig. 10-14. Compressive strength and permeability behavior of
cate hydrates are not stable in such a chemical environ- silica-stabilized Portland cements containing various amounts of
ment, even at ordinary temperatures (Taylor, 1964). silica fume (after Grabowski and Gillot, 1989). Ambient curing for 7
Upon exposure to carbonate solutions, calcium silicate days at 230°C, 100% real-time humidity, and 2.75 MPa compressive
hydrates are eventually converted to a mixture of cal- strength. Reprinted with permission from Elsevier.

Chapter 10 Thermal Cements 331

cium carbonate and amorphous silica. This phenomenon Table 10-4. Compositions of Typical Geothernal Cement
has been observed in well cements by numerous Systems†, ‡
researchers (Onan, 1984; Bruckdorfer, 1986; Shen and Sample Parts by Components Slurry
Pyle, 1989). High-alumina cements are also known to Code Weight Weight
degrade in the presence of carbonate (Crammond and
1 100 API Class G cement 15.1 lbm/gal
Currie, 1993). At present there appear to be no pub- (64, 2C, 21.5S, 3.9A, 3.8F) [1.81 g/cm3]
lished data regarding the behavior of high-alumina 35 Silica flour
cements in a carbonaceous environment at elevated 1 Lignin-sugar
temperatures. 54 Water
The principal, and generally successful, defense 2 100 API Class J cement 15.4 lbm/gal
against such degradation has traditionally been the (37.3C, 54.2S, 1.1A, 1.0F) [1.85 g/cm3]
placement of low-C/S-ratio cement systems with very low 0.4 Lignin-sugar
permeability. However, such systems have been shown 44 Water
to be inadequate for geothermal wells with formations 3 100 API Class F cement 15.1 lbm/gal
containing very high concentrations of CO2 (Hedenquest (63.9C, 21.1S, 3.1A, 54.F) [1.81 g/cm3]
and Stewart, 1985). Milestone et al. (1986 and 1987) 40 Silica flour
demonstrated that tobermorite and xonotlite are among 0.7 Lignin-sugar
the cement phases least resistant to carbonation, and 63 Water
their deterioration is accelerated when bentonite is 4 30 API Class J cement 13.7 lbm/gal
present in the cement. Milestone et al. discovered that 40 Pozzolan [1.65 g/cm3]
reducing the silica flour concentration from 35% to 20% 30 Blast furnace slag
(BWOC) improves the cement’s resistance to CO2. When 0.5 Carboxymethylcellulose
60 Water
less silica is present, weaker and more permeable cal-
cium silicate hydrates form; however, a substantial quan- 5 100 API Class G cement 13.5 lbm/gal
tity of calcium hydroxide also remains in the system. (64.2C, 21.5S, 2.9A, 3.8F) [1.62 g/cm3]
Upon substantial carbonation, the calcium hydroxide 35 Silica flour
8.5 Perlite
reacts to form a protective layer of calcite, the perme- 2 Bentonite
ability decreases, and further attack is inhibited. 1 Lignin-sugar
Because of the presence of weak formations and low 116 Water
fracture gradients, lower-density cements are often
required in geothermal wells. Therefore, research has 6 100 API Class G cement 14.0 lbm/gal
(64.2C, 21.5S, 3.9A, 3.8F) [1.68 g/cm3]
been performed to develop low-density systems that will 35 Silica flour
perform adequately. The typical extenders used to pre- 10 Diatomaceous earth
pare low-density geothermal cements are bentonite 1 Lignin-sugar
and diatomaceous earth. Additional silica flour, up to 91 Water
100% by weight of cement, is sometimes added in lower- 7 100 API Class G cement 15.5 lbm/gal
density systems to ensure proper stabilization (Gallus et 40 Silica flour [1.86 g/cm3]
al., 1979). 0.8 Dispersant
Ultralow-density foamed cements (Rickard, 1985; 0.8 Fluid-loss agent
Sugama et al., 1986) and microsphere-extended systems 0.4 Retarder
60.3 Water
have been used to cement geothermal wells. Such sys-
tems have been used successfully in thermal recovery 8 100 API Class G cement 13.6 lbm/gal
wells (Section 10-9.1); however, information is sparse 100 Silica flour [1.63 g/cm3]
regarding the long-term stability of these systems to cor- 0.3 Retarder
85.1 Water
rosive brines. Until sufficient data are available, it would
be prudent to restrict the use of ultralow-density sys- 9 100 API Class G cement 15.4 lbm/gal
tems to applications in which the formations fluids are 80 Silica flour [1.85 g/cm3]
not aggressive. 0.5 Fluid-loss agent
0.3 Retarder
Table 10-4 lists the compositions of both normal and 76.8 Water
low-density systems that are often used as geothermal
cements. The compressive strength and water perme- 10 100 API Class G cement 15.7 lbm/gal)
ability upon long-term exposure to actual geothermal 40 Silica flour [1.89 g/cm3]
1 Retarder
conditions are shown in Figs. 10-15 and 10-16, respec- 59.2 Water
tively. † fromAPI Task Group on Geothermal Well Cements, 1985. Reprinted with permission from
Oil & Gas Journal.
‡A = AI2O3, C = CaO, F = Fe2O3, M = MgO, S = SiO2.

332 Well Cementing

 Cement Sample Designation
1 2 3 4 5 6 7 8 9 10
1 day
12,000
3 months 80
Cubes 6 months
12 months
10,000
Cups 60
8,000
Compressive Compressive
strength strength
(psi) 6,000 40 (MPa)

4,000
20
2,000
1,000 psi

0 0
Compressive strength of cement cube and sandstone cup samples after aging periods of 1 day and 3, 6,
and 12 months. Cup samples were cured and aged downhole. Cubes were laboratory cured under water
at 392°F [200°C] for 1 day, then exposed downhole for 3, 6, and 12 months in the Cerro Prieto geothermal
field, Mexico. The downhole temperature was 417°F [214°C].

Fig. 10-15. Compressive strength performance of typical geothermal well cements under
actual conditions (from API Task Force on Geothermal Well Cements, 1985). Reprinted
with permission from Oil & Gas Journal.

Cement Sample Designation

1 2 3 4 5 6 7 8 9 10

2 5
1 day
3 months
1 4
6 months
12 months
0 3
0.1 mD
Log10 –1 2 Log10
permeability permeability
(log mD) –2 1 (log nm2)

–3 0

–4 –1

–5 –2

–6 –3
Water permeabilities of cement samples taken from slurry-filled sandstone cup holders after
curing 1 day and 3, 6, and 12 months downhole in the Cerro Prieto geothermal field, Mexico.
The downhole temperature was 417°F [214°C].

Fig. 10-16. Permeability performance of typical geothermal well cements under actual
conditions (from API Task Force on Geothermal Well Cements, 1985). Reprinted with
permission from Oil & Gas Journal.

Chapter 10 Thermal Cements 333

10-8.2.2 Alternate geothermal well cement common methods practiced today. These techniques
compositions have been the salvation of many oil fields with high-vis-
More recently, calcium phosphate and calcium alumi- cosity crudes, and essentially involve the trading of heat
nosilicate systems were developed specifically for use in for viscosity reduction (Kastrop, 1965; Butler, 1991).
geothermal wells, and both show superior resistance to As in geothermal wells, the formations associated
CO2. The chemistries of each system are described ear- with steam recovery and fireflood wells are frequently
lier in this chapter. problematic. Weak and unconsolidated zones with low
Calcium phosphate systems (Section 10-6) have been fracture pressures and high permeability are often pre-
used since 1997 to successfully cement geothermal wells sent; as a result, severe lost circulation and fluid-loss
in Japan and Indonesia, and service life is estimated problems are common.
to be about 20 years (Weber et al., 1998; Brookhaven Thermal recovery wells are usually less than 3,000 ft
National Laboratory, 2000). The cement blends consist [915 m] in depth and are frequently deviated (30° to
of fly ash, calcium aluminate cement, sodium polyphos- horizontal). The circulating temperatures during pri-
phate, and water in compositions that vary with the mary cementing operations are often less than 104°F
depth at which the cement will be used. [40°C], and accelerators such as calcium chloride or
Calcium aluminosilicate systems (Section 10-5) were sodium chloride are often added to promote early
developed more recently (Barlet-Gouédard and Vidick, cement strength development.
2001; Barlet-Gouédard and Goffé, 2002). At this writing, Thermal recovery wells are always cemented to sur-
the performance of these systems in actual wells has not face. When heat is initially supplied, the temperature
been published. rise should be controlled to prevent undue thermal
Placing cements that are chemically inert to corro- shock to the casing and cement. Nevertheless, because
sive geothermal brines would be an attractive strategy. of thermal expansion, high levels of stress are built up in
Such systems, commonly referred to as “synthetic the pipe and the cement sheath (Carter et al., 1966);
cements,” are used routinely to complete wells for therefore, the strongest possible cement-to-pipe and
CO2-flooding projects or chemical waste disposal cement-to-formation bonds are necessary. Failure of the
(Chapter 7). Epoxy-based polymer systems are most bonds could allow interzonal communication and pipe
commonly used for such applications; unfortunately, expansion. The ultimate result would be casing failure
they would suffer thermal degradation at the tempera- by buckling or telescoping (Humphrey, 1960). A sub-
tures encountered in geothermal reservoirs. However, stantial amount of work has been performed to devise
epoxy cements are often used near the surface, because cementing techniques that minimize the effects of ther-
they can withstand the temperature cycling associated mal expansion. Such methods include the placement of
with geothermal well production. thermal packers (Smith, 1966) and the inclusion of a
Polymers that are stable to high temperatures have sliding sleeve in the casing string that can move freely in
been examined by various researchers. Zeldin and response to thermal stress (Greer and Shryock, 1967). A
Kukacka (1980) developed an organosiloxane polymer third procedure involves holding the casing in tension
cement that was proven suitable as a geothermal during the cement job to minimize the expansion when
cement in an API study. A coal-filled furfuryl alcohol- thermal stress is eventually applied (Farouq Ali and
base cement system for geothermal wells was proposed Meldau, 1979).
by Pettit (1979) and Eilers (1985). Degouy and Martin The cement must also be able to withstand the ele-
(1993) demonstrated that phenolic resins, with fillers vated temperature exposure and thermal cycling associ-
such as calcium carbonate and sand, provided accept- ated with steamflood and fireflood wells. To maximize
able performance at curing temperatures up to 330°F the delivery of heat to the pay zones, an insulating
[200°C]. No commercial use of these technologies has cement is desirable in thermal recovery wells; however,
been reported. the presence of such cements places additional thermal
stress on the casing (Leutwyler, 1966). Thermal conduc-
tivity is more dependent upon the cement density than
cement composition (Nelson, 1986). Typical laboratory
10-9 Thermal recovery wells data are shown in Fig. 10-17. At equivalent density, the
The application of heat to stimulate heavy oil production thermal conductivity of foamed cement is only margin-
has been practiced for more than 50 years. Methods such ally different from that of conventionally extended
as in situ combustion (fireflood), downhole heaters, hot cement.
fluid injection, and steam stimulation have been used. In
situ combustion and steam injection are the most

334 Well Cementing

 Cement density (g/cm3)

0.9 0.7 0.9 1.1 1.3 1.5 1.7 1.9

0.8

Steam
0.7

Oil
0.6
Thermal
0.5
conductivity
BTU/hr ft °F 0.4
0.3
0.2
0.1
5.8 7.5 9.1 10.8 12.5 14.1 15.8
Cement density (lbm/gal)

Fig. 10-17. Typical cement density/thermal conductivity relation-

ship (from Nelson, 1986). Reprinted with permision from Oil & Gas
Journal. Caprock
Sand

10-9.1 Steam recovery wells

Enhanced oil recovery may be accomplished by steam-
flooding or cyclic steam stimulation (Gates and Holmes,
1967; Curtis et al., 2002). Steamflooding consists of intro- Steam injection
ducing steam into an injection well and sending the Heated heavy oil
steam through the formation to a production well. Cyclic Shale
flows to well
steam stimulation of production wells involves the injec-
tion of steam into the production well for a short period Fig. 10-18. SAGD process (from Curtis et al., 2002).
of time and returning the well to production
(Earlougher, 1968). Steam recovery techniques are prac-
ticed extensively throughout the world (Chu, 1983).
During the last decade, the steam-assisted gravity permeability is required, and normal- to high-density
drainage (SAGD) method for recovering heavy crudes slurries are best at providing these qualities.
has been extensively developed (Butler, 1998 and 2001). Unfortunately, because of the lost circulation and ther-
The process uses twin horizontal wells drilled and mal conductivity considerations, such slurries are gen-
extended into the base of a reservoir with the horizontal erally unsuitable. Therefore, much research has been
steam injector placed directly above the horizontal pro- performed to devise low-density slurries with high com-
duction well (Fig. 10-18). In an ideal SAGD process, a pressive strength and low permeability.
growing steam chamber forms around the horizontal Conventionally extended Portland cement systems,
injector, and steam flows continuously to the perimeter containing perlite, bentonite, diatomaceous earth, etc.,
of the chamber, where it condenses and heats the sur- generally perform adequately in steamflood wells, pro-
rounding oil. As the viscosity of the oil decreases, it vided the slurry density is above 12.5 lbm/gal [1.5 g/cm3].
drains to the horizontal production well underneath. Their long-term performance is very similar to that
Thus, the use of gravity increases the efficiency of oil exhibited by such systems in deep wells (Fig. 10-19).
production. The formations in steamflood wells are often so
The most important steamflood fields are located in incompetent that cement systems with densities less
central and southern California, United States; Alberta than 12.5 lbm/gal [1.5 g/cm3] are required to avoid lost
and Saskatchewan, Canada; Venezuela; The Netherlands; circulation or formation damage. Thus, silica-stabilized
West Germany; and Indonesia. Reservoir temperatures foamed cements (Smith, 1983) and microsphere-
seldom exceed 600°F [315°C]; therefore, Portland cement extended systems (Ripley et al., 1980) are very common
is used in virtually all steamflood well completions. in steamflood well completions today. Previously, multi-
The characteristics of steamflood wells and the asso- stage cementing was necessary to successfully complete
ciated performance requirements of cementing materi- these wells.
als are often at cross-purposes. A strong cement with low

Chapter 10 Thermal Cements 335

 12 lbm/gal [1.44 g/cm3]
11 lbm/gal [1.32 g/cm3]
10 lbm/gal [1.20 g/cm3]

Cured at 450°F [232°C] Cured at 600°F [315°C]

6

4
Compressive
strength 3
(thousand
psi)
2

100

10

Water
permeability 1
(mD)

0.1

0.01
1 3 6 12 24 1 3 6 12 24
Time (months) Time (months)
Fig. 10-19. Long-term performance of glass microsphere systems cured at elevated temperatures.

Typical slurries using glass or ceramic microspheres peratures (unpublished data, Nelson, 1987). X-ray dif-
are prepared with a silica-stabilized Portland cement– fraction analysis of the systems revealed the coincident
based slurry. The long-term performance of glass micro- appearance of reyerite and certain aluminosilicate
sphere systems cured at 450° and 600°F [232° and 315°C] hydrate phases. Ceramic microspheres are derived from
is shown in Fig. 10-19. The slurry densities vary from 10.0 fly ashes, and the delayed (reyerite-related) deteriora-
to 12.0 lbm/gal [1.20 to 1.45 g/cm3]. tion of normal-density fly ash cement systems has been
The performance of silica-stabilized ceramic micros- discussed earlier in this chapter.
phere systems at 450° and 600°F [232° and 315°C] is Typical foamed cement systems for thermal wells are
shown in Fig. 10-20. Initially, these systems were gener- prepared from a normal-density base slurry of Portland
ally stronger and less permeable than their glass micros- cement, at least 35% silica flour, a surfactant, and a
phere counterparts. However, between 1 and 2 years of foam stabilizer. The long-term performance at 450° and
curing, significant deterioration was noted at both tem- 600°F [232° and 315°C] of three foamed cement systems

336 Well Cementing

 12 lbm/gal [1.44 g/cm3]
11 lbm/gal [1.32 g/cm3]
10 lbm/gal [1.20 g/cm3]

Cured at 450°F [232°C] Cured at 600°F [315°C]

6

4
Compressive
strength 3
(thousand
psi)
2

100

10

Water
permeability 1
(mD)

0.1

0.01
1 3 6 12 24 1 3 6 12 24
Time (months) Time (months)

Fig. 10-20. Long-term performance of ceramic microsphere systems cured at elevated temperatures.

with densities ranging from 9.0 to 12.0 lbm/gal [1.08 to Febus, 1984). Compressive strength and permeability
1.44 g/cm3] is shown in Fig. 10-21. Comparison of the data for systems cycled between 550° and 100°F [288°
foamed cement data with those of equal-density micros- and 94°C] are shown in Table 10-5. More recently, ther-
phere systems reveals the foams to have significantly mal cements containing additives that impart flexibility
higher compressive strength. The water permeabilities (Chapter 7) have been successfully introduced for
of the foamed cements are also higher (>0.1 mD), and steamflood applications (Stiles and Hollies, 2002; Stiles,
more variable with curing time. 2006).
Foamed cements have also been shown to resist
repetitive thermal cycling, which occurs when the cyclic
steam stimulation technique is applied (Harms and

Chapter 10 Thermal Cements 337

 12 lbm/gal [1.44 g/cm3]
11 lbm/gal [1.32 g/cm3]
9 lbm/gal [1.08 g/cm3]

Cured at 450°F [232°C] Cured at 600°F [315°C]

6

4
Compressive
strength 3
(thousand
psi)
2

100

10

Water
permeability 1
(mD)

0.1

0.01
1 3 6 12 24 1 3 6 12 24
Time (months) Time (months)

Fig. 10-21. Long-term performance of foamed cement systems cured at elevated temperatures.

Table 10-5. Effect of Thermal Cycling on Performance of 10-9.2 In situ combustion wells
Foamed Cements for Steamflood Conditions† In situ combustion recovery, or fireflood, consists of ini-
Properties of Foamed Cement Density tiating combustion in an injection well and then propa-
Foamed Cement gating the combustion front by the injection of air or
10 lbm/gal 11.5 lbm/gal 13 lbm/gal
oxygen through the reservoir to the production wells
Compressive strength 1,210 psi 1,680 psi 2,260 psi (Chu, 1981; Petit et al., 1992). In such wells, the cement
after 20 days at 550°F
is exposed to maximum temperatures between 700° and
Compressive strength 1,630 psi 1,550 psi 2,440 psi 1,700°F [371°and 926°C] near the burning zone. Such
after 100 days at 550°F‡ temperatures exceed the stable range of Portland
Compressive strength 1,240 psi 2,020 psi 2,430 psi
cement; therefore, high-alumina cement is necessary.
after 160 days at 550°F§ Fireflood wells are physically similar to and are usu-
ally found in the same locations as steam injection wells.
Air permeability after 2.4 mD 1.0 mD 0.9 mD Thus, the formation conditions and cement performance
100 days requirements are basically the same. Usually, most of the
† Surface slurry: 15.4 lbm/gal Class G, 40% silica flour, 3% lime (from Harms and Febus, 1984).
Reprinted with permission of SPE. casing is cemented with Portland cement systems, with
‡ Cycled to 100°F twice.
calcium aluminate cement placed opposite and about
§ Cycled to 100°F three times.

338 Well Cementing

100 ft [31 m] above the pay zone as a tail slurry. However, because of the previously described conversion of the
this practice is not without risk. Calcium aluminate initial aluminate hydrates to C3AH6. The water-perme-
cement is a strong accelerator of Portland cement, and ability values are extremely low as well.
setting can occur within a few minutes after the two The performance of foamed calcium aluminate
slurries commingle in the annulus. For complete safety, cements has also been investigated (Nelson and Eilers,
calcium aluminate cement should be the only system 1985). Figure 10-23 shows the compressive strength and
pumped. water permeability of three systems cured for 7 and
The performance of two normal-density calcium alu- 28 days at 1,250°F [677°C] in a refractory furnace. Two
minate cement systems is depicted in Fig. 10-22. Data foams, with densities of 11.0 and 9.0 lbm/gal [1.32 and
are given for systems cured at 100° and 220°F [38° and 1.08 g/cm3] were prepared from a neat calcium alumi-
93°C] and heated in a refractory furnace at 600°, 1,000°, nate cement-based slurry. Another foam, with a density
and 1,500°F [315°, 538°, and 815°C]. The compressive of 11.0 lbm/gal [1.32 g/cm3], contained fly ash. The com-
strengths of the aluminate systems at the lower temper- pressive strength was adequate; however, the water
atures are adequate, yet considerably lower than similar- permeabilities were excessive.
density Portland cement systems. This is primarily

538°C 538°C
38°C 93°C 315°C 815°C 38°C 93°C 315°C 815°C
15 100

10

10
1
Compressive Water
strength permeability
(MPa) (mD)
0.1
5

0.01

0 0.001
1 3 7 1 7 3 1 7
Curing time (days) Curing time (days)

100% ground firebrick (BWOC)–15.6 lbm/gal [1.87 g/cm3]

200% ground firebrick (BWOC)–15.7 lbm/gal [1.87 g/cm3]

Fig. 10-22. Compressive strength and permeability performance of calcium aluminate cement systems
at various temperatures (from Nelson and Eilers, 1985). Reprinted with permission from the Petroleum
Society of CIM.

Chapter 10 Thermal Cements 339

 15 10
Neat: 11.0 lbm/gal [1.32 g/cm3]
Fly ash: 11.0 lbm/gal [1.23 g/cm3]
Neat: 9 lbm/gal [1.08 g/cm3]

10 1

Compressive Water
strength permeability
(MPa) (mD)

5 0.1

0 0.01
7 28 7 28
Curing time (days) Curing time (days)

Fig. 10-23. Performance of foamed calcium aluminate cement systems at 1,250°F [677°C] (from Nelson
and Eilers, 1985). Reprinted with permission from the Petroleum Society of CIM.

■
10-10 Conclusion Microsphere cement systems can be used in thermal
The preceding discussion has demonstrated that ther- wells, provided the base slurry is stabilized to high
mal cements encompass a wide variety of wellbore con- temperatures, and the collapse pressure (usually
ditions and complex chemical processes. Many factors 3,000 psi or 20.7 MPa) is not exceeded.
■ Foamed cement, made from a stabilized base slurry,
must be considered to determine the optimum cement
composition for a particular situation. Nevertheless, can be used with confidence in most thermal wells. In
there are several basic points that the engineer should geothermal wells, in which corrosive fluids are pro-
remember when contemplating this problem. duced, the long-term stability of foamed cements has
■ When static temperatures exceed 230°F [110°C], 35%
not been proven.
■ If the cement will be exposed to temperatures
to 40% silica BWOC must be added to Portland
cements; otherwise, strength retrogression will occur. exceeding 750°F [400°C], Portland cement should
■ If saline geothermal brines are present, fine silica
not be used. High-alumina cement is suitable.
■ Silica is deleterious to the stability of high-alumina
flour (less than 15-μm particle size) should be added
to Portland cement as a stabilizer. Silica sand does cements at temperatures exceeding 572°F [300°C].
not reliably provide adequate protection. Crushed aluminosilicate firebrick or fly ash is suit-
■ If high concentrations of CO2 are present, using cal-
able.
■ During laboratory testing, accurate static and circu-
cium aluminosilicate or calcium phosphate cements
is recommended. If Portland cement is used, degra- lating temperatures must be used to obtain an opti-
dation can be inhibited by reducing the silica con- mal thickening time and compressive strength at the
centration to 20% BWOC. wellsite.
■ Most common cement extenders are compatible with
thermal cements; however, if the static temperature
exceeds 450°F [232°C], fly ash should not be used in
Portland or Class J cement systems. Bentonite, per-
lite, and diatomaceous earth are suitable.

340 Well Cementing

■ High pressure strongly affects the behavior of thermal 10-11 Acronym list
cement systems; therefore, laboratory testing must be API American Petroleum Institute
performed at the anticipated bottomhole pressure.
■
ASTM ASTM International, formerly the American
Thermal cements are sensitive to subtle chemical Society for Testing and Materials
changes; therefore, laboratory testing should always
be performed with samples of the cement, additives, BHCT Bottomhole circulating temperature
and location water that will be used during the job. BHST Bottomhole static temperature
■ The common assumption that high compressive BSC Belite-silicate cement
strength is automatically linked with low permeabil- BWOC By weight of cement
ity is false. Permeability should be measured in the C/S Bulk lime-to-silica (ratio)
laboratory before a cement system is placed in a ther-
mal well. CA Monocalcium aluminate
CH Calcium hydroxide
C-S-H Calcium silicate hydrate
SAGD Steam-assisted gravity drainage
WOC waiting on cement

Chapter 10 Thermal Cements 341

 Cementing Equipment and Casing Hardware
Ed Leugemors, James Metson, and Jean-Louis Pessin—Schlumberger

11-1 Cementing materials

R.L. Colvard—Weatherford International
C. Dennis Krauss and Mark Plante—Baker Oil Tools

Before describing the design and function of cementing

equipment, one must be familiar with the physical and
11
(94 lbm) or SI (50 kg) sacks, so-called “big bags” (1 to
1.5 SI tons generally), or larger quantities (truck, rail-
way car, or ship).
chemical properties of the various cementing materials.
A thorough discussion is presented in Chapters 2 and 3; 11-1.2 Water
however, a rapid review of the principal points is useful Fresh water is normally used for cementing onshore
for this discussion. wells and seawater for offshore locations. However, one
must be aware that fresh waters are often not very
11-1.1 Cement “fresh.” Inorganic salts and organic residues from vege-
As discussed in Chapter 2, Portland cement is used tation are frequently present in significant quantities.
during almost all well cementing operations. It is a Such materials are known to affect the performance of
finely divided and highly reactive powder. The vigorous Portland cement systems.
hydration of Portland cement during initial mixing, as
well as the changeable slurry properties during place- 11-1.3 Dry cement additives
ment, complicate the design of cement mixing and Table 11-1 summarizes the relevant properties of dry
pumping equipment. additives with regard to cementing equipment and
Portland cement is usually stored in silos at a central logistics.
storage location. Alternatively, it may be packaged in U.S.

Table 11-1. Important Properties of Dry Additives with Respect to Cementing Logistics
Material Chemically Inert Chemically Active
Form Insoluble powder or finely cut material Soluble powder or finely cut material

Examples (function) Powdered coal (extender) Lignosulfonate (retarder)

Hematite (weighting agent) Sodium silicate (extender)
Barite (weighting agent) Calcium chloride (accelerator)

Maximum concentration 100% (BWOC) 10% (BWOC)

(order of magnitude)

Influence of accuracy in Concentration acts directly on system Materials may have secondary effects.
concentration on slurry quality density; no unexpected effect.

Handling The additives are blended with the dry The additives are normally dry-blended with the
cement in a special blender at the central cement. They are sometimes added to mix water
storage location up to several days before on location in an open horizontal tank, just prior
the pumping job. The blended material is to the job.
then transported to the wellsite. (If the
amount required is small and the material
easily scattered in the water, it is treated
as a soluble material.)

Well Cementing ■ Chapter 11 Cementing Equipment and Casing Hardware 343

11-1.4 Liquid cement additives Table 11-2. Important Properties of Liquid Additives
Offshore, liquid additives are usually preferred. Such with Respect to Cementing Logistics
materials are more compatible logistically, because their Type of material All liquid additives are chemically active.
mixing requires less space. Because liquid additives are Frequently, they are water solutions of
preblended with the mix water, the resultant slurry the corresponding dry additive.
tends to be more homogeneous than one mixed from a Examples Lignosulfonate solution (retarder)
dry-blended powder mixture. Table 11-2 is a summary of (function) Naphthalene sulfonate solution (dispersant)
the relevant points. Latex (gas migration prevention)

Concentration (order Up to 25 L/100 kg cement

of magnitude) (3.0 gal/U.S. sack)
11-2 Basic equipment
Influence of Materials may have detrimental
Figure 11-1 is a schematic flow diagram of cement-slurry accuracy in secondary effects.
preparation that indicates the steps performed at the concentration
central storage location and at the wellsite. Each func- on slurry quality
tion in the flow diagram also represents a major piece Handling Additives are blended with the mix water
of equipment. Some functions may be combined into a on location in a horizontal open tank,
multipurpose “cementing unit.” In the following discus- shortly before the job. They also can
sion, the design and operation of each class of equip- be blended on the fly during the job.
ment are presented in detail. The second method is preferred.

Localization Mainly offshore

11-2.1 Cement and dry additive blending
Figure 11-2 is an overview of the delivery and blending
processes that occur at the central storage location. pneumatic loading bottles (Fig. 11-3), mechanical screw
Cement is delivered to the central storage location. elevators, or combined systems (for unloading from
Upon delivery in bags, the cement is usually combined. ships). These transfer systems may also be used to load
Equipment that may be used in the transfer includes dry additives.

Central storage location Transfer Wellsite

Storage Storage

Dry
cement (1)
Dry (3) Blend Surge
additive (3) tank
blending
HP steel
Dry
flow hoses
additives (2)
Water
Liquid Cement
Cement
additive slurry
mixer
mixing pumper
Liquid (2) Liquid
additives (2) additives (4)

Well casing
(1): Usually bulk, possibly U.S. (94-lbm) or metric (50-kg) sacks, or “big bags” (0.5 to 1.5 metric tons) topped with
(2): Manufacturer‘s packaging the cement
(3): Bulk (for very small and/or unplanned jobs, cement is sometimes stocked in paper sacks) head
(4): Bulk in special containers, except if mixing is done in an open tank

Fig. 11-1. Typical cementing process.

344 Well Cementing

 Cement from cement mill delivered
to the central bulk station

1.

2.

3.

4. 6.
1. Standard sacks or large bags
2. Bulk (notice the absence 5. Cement from the bulk
of surge tank on trailer) station sent to the
3. Sacks, large bags, or bulk rig location
4. Pneumatic silo 7.
5. Pneumatic loading bottle
6. Dry-additive blender 8.
7. Air-compressor plant
8. Horizontal tank trailer with
9.
merge tank
9. Twin vertical tank trailer
with merge tank
10. Supply boat, cementing 10.
barge, or vessel

Fig. 11-2. Delivery, storage, and distribution of cement and dry additives.

Loading The bulk cement is stored in pneumatic or atmos-

(unsacked dry pheric silos. Transfer systems are available to move the
materials)
cement from one silo to another or to a blender, road
transport unit, or supply boat. When the transfer system
is pneumatic, several silos are connected permanently to
Hopper save time and labor. In humid climates, an air dryer may
be installed in the system.
The cement and dry additives are usually combined in
Transfer to Bottle a pneumatic blending tank (10- to 20-ton capacity) at
dry-additive the central bulk material plant (Fig. 11-4). Bulk materi-
blender, bulk als are usually air-blown, and sacked materials are
storage silo, or
bulk transport Pressurized poured into the tank through a hopper located on top.
air The sacked additives may also be poured into pneumatic
additive bottles (1- or 2-ton capacity) and then blown
into the blending tank. Pressurized air is supplied by one
Fig. 11-3. Pneumatic dry-material loading bottle.
or more air compressor units.
The bulk materials are loaded into the blending tank
first for weighing purposes. A weigh cell is permanently
integrated into the tank frame. The sacked additives are
loaded last. Air, pressurized to about 35 psi [2.5 bar], is
injected through nozzles into the mass of the materials
until thorough blending is accomplished. Then the blend
is transferred pneumatically to a bulk material transport
container for delivery to the wellsite. To obtain the
required amount of cementing materials, more than one
batch must frequently be prepared.

Chapter 11 Cementing Equipment and Casing Hardware 345

 rough (e.g., corrugated desert track). This problem is
Dust magnified if the blended materials have significantly
collector different particle sizes or densities. If small quanti-
ties of soluble or easily dispersible additives are
needed, mixing the dry additives with water at the rig
Hopper
Transportation site is often preferred.
Vent skid ■ If the cement job is canceled or postponed for a long
Pressure
period after the blend has been prepared, a storage or
filling lines reuse problem may arise. Adding liquid additives only
during the job prevents such situations because shelf
Air-jetting Vent from other life problems are eliminated.
system tank (if needed†)

Delivery Delivery 11-2.2 Transportation of bulk materials or blends

†Used
to the wellsite
if the other pressurized tanks
in the bulk station vent out only The equipment used to deliver the cementing materials
through this dust collector to the wellsite varies according to the location. The var-
Fig. 11-4. Dry-additive blender. ious types of transports are discussed below.
Land rigs: Trucks or semitrailer transports are gen-
erally used for land operations. As shown in Fig. 11-5,
truck-mounted vertical tanks or semitrailer-mounted
Although very popular, this method has two draw- vertical or horizontal tanks are the most common.
backs. Limited access locations: Cementing materials in
■ Some particle segregation (caused by shaking) can sacks or big bags can be transported to remote locations
take place during transportation to the rig site, espe- by truck, cargo plane, barge, ship, or helicopter. The con-
cially if the distance is long and the road surface is

Air compressor
Surge
tank

Fill Vent

High-pressure Low-pressure
delivery delivery

Air compressor Surge

tank
High-
pressure
delivery

Low-
pressure
Vent delivery
Fill

High-pressure delivery: Material transferred to storage tank or to mixing unit surge tank
Low-pressure delivery: Material air-blown to surge tank on rear of unit and dumped into cement mixer hopper

Fig. 11-5. Bulk material transports.

346 Well Cementing

tainers are made of durable, water-resistant materials. Side view Back view
Sometimes, the containers are placed in stackable Vent
crates for added protection. Cement
Offshore rigs: Supply boats or cementing vessels are
used for offshore locations. They normally have built-in
tanks, pneumatic unloading equipment, and a supply of
hoses. Sometimes, mobile skid-mounted tanks with a low
center of gravity and mobile unloading equipment are

Back
Face
used. It is important to note that the pneumatic equip-
ment must be sufficiently powerful to blow heavy mate-
rials such as barite (specific gravity of 4.33) up to the
drilling rig tanks, a vertical distance of 130 to 200 ft [40 Porous
to 60 m]. The air compressors used for this task typically material
deliver 250 to 350 ft3/min [7,100 to 9,900 L/min] with a Air box
Valve Fill Vent
pressure rating of 28 to 44 psi [2 to 3 bar]. Sock
Air box
Delivery into
11-2.3 Wellsite storage of cement or blends Aeration and cement mixer
fluidization air hopper
As discussed above, pneumatic bulk trucks or trailers
transport neat or preblended dry cement to the wellsite Fig. 11-6. Atmospheric transportable bulk tank (typical piping
from the central storage and blending plant. Neat arrangement).
cement can also arrive directly from the cement mill.
The material is then transferred pneumatically to trans-
portable tanks that are either brought to the rig site for Safety valve
the cement job or are a permanent part of the drilling rig
equipment. Such tanks are similar to those used at cen-
tral storage locations, but their dimensions allow trans-
port on standard or specially designed (with a built-in
hydraulic laying/raising system) trailers. When empty,
the tanks must not exceed the weight limits specified by
various countries. A large variety of storage tanks for
road travel exists within two principal categories—
atmospheric and pressurized. Both are equipped with a
set of skids for proper installation on imperfectly leveled
ground and for easy winching onto trailers.
The atmospheric tank is always operated in a vertical
position. Air at low pressure (about 3 psi [0.2 bar]) is
blown into a gutter fixed to the slanted bottom of the 8 jets (1-in.)
2-in. pressurizing line
tank. The roof of the gutter is made of a porous material. 5-in. bleedoff/vent line
5-in. fill
The air passes through the porous partition and fluidizes 5-in. material delivery
Quick manhole
the cement blend. The cement blend glides along the 50-in. ID 6 jets (1-in.)
slanted bottom to a chute gate and then to the hopper of 2-in. air (letting) Bronze porous floor
a slurry mixing system. As illustrated in Fig. 11-6, atmos- 2-in. air (aeration)
pheric tanks are made in the shape of a parallelepiped.
Pressurized tanks use air at about 44-psi [3-bar] pres- Fig. 11-7. Pressurized bulk tank (typical piping arrangement).
sure and can operate horizontally or vertically.
Figure 11-7 is a schematic diagram of a typical unit. As
shown in Fig. 11-8, the vertical tanks are generally cylin- mixer. For versatility, some vertical pressurized tanks
droconical in shape, while horizontal models are more are also equipped to release the cement directly to a
complex. In the first stage, pressure-reduced air is blown hopper at atmospheric pressure (Fig. 11-9).
from the bottom through the mass of cement for aera- The bulk trailers are sometimes used for additional
tion and fluidization. Then air at 44 psi [3 bar] is storage. Indeed, they can serve all storage needs on the
injected into the tank, and the cement flows out through rig site, provided they are equipped with their own surge
a discharge line to a surge tank, which feeds the cement tanks, described later.

Chapter 11 Cementing Equipment and Casing Hardware 347

 Horizontal (supply boat)

4-in. vent line

5-in. delivery line
Fill line
3-in. air line
Vertical, single Vertical, twin

Fig. 11-8. Pressurized bulk tanks (various configurations).

Delivery to surge tank

(in high-pressure mode)

Bulk cement
(or blend)

Valve‡
Porous
Enlarged material Annular
to show air box
detail Valve† Valve†
Sock
Aeration and
fluidization air
Delivery to mixer hopper
Open in atmospheric mode, closed in high-pressure mode
†

Closed in atmospheric mode, open in high-pressure mode

‡ (in atmospheric mode)

Fig. 11-9. Pressurized bulk tank (typical piping arrangement).

11-2.4 Metering of water freshwater or seawater distribution system is used. Each

Contrary to what one might think, the simple method of batch of mix water is then used successively to feed the
employing a flowmeter is not used for water metering. A cement mixer.
set of twin 10-bbl (or sometimes 20-bbl) tanks is pre- The additives may be preblended with the water in
ferred. A “displacement tank” (Fig. 11-10), which is the storage tanks or may be blended while the water is
divided equally by a partition, is also used. passing through the displacement tank. In the second
Both sides of the partitioned tank are filled with mix case, a liquid-additive metering system (described later)
water from the rig storage. If the job is offshore, the is required.

348 Well Cementing

 Additives (from exist on the rig site, especially offshore. If the job is can-
liquid additives celed or postponed, the costly solution may have to be
metering system) thrown away. Also, if a larger-than-expected volume of
slurry becomes necessary during the job, the volume of
Water the premixed additive solution may be inadequate. Thus,
methods that allow continuous (“on-the-fly”) mixing are
often preferred. On-the-fly methods employ a semiman-
ual or automatic metering system that delivers the cor-
Closed
Open rect amount of additives to each side of the displace-
Partition ment tank.
Water
pipe

Level scale
11-2.5.1 Liquid additive metering system with
metering tanks
All liquid-additive metering systems consist of two prin-
cipal parts—a storage and transfer unit and a metering
unit.
The storage/transfer unit generally includes four stor-
age tanks of various capacities (usually between 6.2 and
Drain
Open
25 bbl [1,000 and 4,000 L]). The storage and transfer
pipe Closed unit allows the independent metering of additives
according to the requirements of a particular job. This is
Mixing water
(to the mixing
convenient because well cement slurries typically con-
pump) tain two or three additives.
Each storage tank is equipped with its own air-oper-
Fig. 11-10. Displacement tank system. ated diaphragm pump and agitation system (recircula-
tion, as illustrated in Fig. 11-11, or air-operated stirrer)
to avoid segregation of the additive components.
For precise placement of the slurry in the wellbore, Therefore the operation of the unit requires a source of
the volume of the displacement fluid must be accurately clean and dry air at 120 to 145 psi [8 to 10 bar]. The con-
measured. After the cement slurry has passed through figuration of the unit varies depending on whether it is
the mixing system, the displacement fluid usually passes designed for use on land (skid or trailer-mounted) or off-
through the displacement tanks for volume measure- shore (containerized).
ment and is pumped by the cementing unit instead of The metering unit generally consists of a set of three
the rig’s mud pumps. (Fig. 11-11) or four 25-gal or 10-L tanks, with visible level
scales. To prepare a batch (10 bbl or 20 bbl [1.6 m3 or
3.18 m3] according to the displacement tank), the
11-2.5 Liquid additive storage and mixing
proper amounts of the selected additives are introduced
The simplest method of mixing liquid additives (and dry into the metering tanks. The additives are then released
additives at less than 3% by weight of cement) with into one of the two displacement-tank sections that is
water consists of pouring the required amount of each being filled with water. Finally, the mixture is agitated to
additive into a tank of water. One should measure the obtain a homogeneous solution. The same operation is
additives and water accurately to obtain the correct con- repeated for the following batch in the other half of the
centration; the preparation of a slight excess of solution displacement tank and so on. The repetitions of the
is also advisable. The mixing can be achieved with a operation may be automatically or semiautomatically
paddle mixer, circulation pump, jetting system, or a com- controlled.
bination of these.
The premix method has several disadvantages.
Premixing requires an extra tank, which must be clean
and sufficiently large. Extra tanks are not always avail-
able, and sufficient space to accommodate them may not

Chapter 11 Cementing Equipment and Casing Hardware 349

 A1 A2 A3
Additive
metering
tank(s)

Open Closed Air-powered

Closed Open stirrer

Water
Additive
storage
tank(s)

Air-powered
diaphragm
pump(s)

Mixing water

Fig. 11-11. Liquid-additive metering system (with metering tanks).

11-2.5.2 Liquid-additive metering system without 11-2.6 Surge tanks

metering tanks For smooth cement mixer operation, the supply of
The liquid-additive-system metering rack (Fig. 11-12) is cement (or blend) should be steady, and the pressure at
used to provide accurate (±2%) delivery of up to four the mixer bowl should remain constant. The bulk
additives into displacement tanks. The operator cement is moved from the storage tank toward the
depresses a button to initiate the delivery. The metering cement mixer, driven by the differential pressure cre-
rack behaves as four “smart valves,” installed between ated between the tank and the end of the line. If the line
the additive pumps and the displacement tanks. The is longer than approximately 23 ft [7 m], the cement
valves are controlled by a microcomputer using data tends to separate from the conveying air into slugs, cre-
from electromagnetic flowmeters. ating pulsating flow. To smooth the flow and allow for
operational requirements, such as changing from one
storage tank to another, a surge tank is used.
Window As shown in Fig. 11-13, the surge tank has a cylindro-
Control cabinet conical shape. It has a capacity of about 12.5 bbl
[2,000 L] and connects the end of the transfer line to the
top of the mixer bowl. This device maintains the pres-
A4 sure above the mixer bowl at atmospheric level plus the
To cementing unit A3 hydrostatic head of the material. A dust separator is
displacement tank From also installed where the conveying air is vented to the
A2 additive atmosphere.
A1 storage In some cases, the surge tank is maintained at higher
tanks
than atmospheric pressure. The advantage is a higher
On/off box delivery rate to the cement mixer, but the drawback is a
higher required pressure from the bulk system to feed
the surge tank. A pressurized surge tank is also used
Fig. 11-12. Metering rack hardware.

350 Well Cementing

with certain types of recirculating cement mixers to 11-2.7 Cement mixing
maintain a steady delivery rate. The tank is pressurized The cement mixer is a device in which a flow of pressur-
by controlling the flow or air in the vent line. This type ized water (possibly containing additives) meets a flow
of surge tank is used almost exclusively offshore, where of cement (possibly containing additives), and a cement
the bulk system is required to transport the cement long slurry is formed at a prescribed rate. Several types of
distances. mixing systems exist; they are described individually
below.

Vent
Dust separator 11-2.7.1 Conventional jet mixer
Air-forced The conventional jet mixer consists of a hopper, a mixing
cement bowl, a discharge gooseneck, and a slurry tub. The max-
Windows imum slurry-generating capacity of the conventional jet
(for watching mixer, evaluated in rate of dry material, is slightly higher
Air-jetting cement level) than 2,200 lbm/min (1 SI ton/min). Figure 11-14 shows a
system
(not shown) configuration for sacked cement, and a system for pneu-
matically delivered cement is illustrated in Fig. 11-15.
The cement is delivered to the hopper. The water is
Valve injected into the bowl through jets for mixing with the
cement and into the gooseneck for adjusting the slurry
Sock density. The jets are chosen according to the operating
pressure, slurry fabrication rate, and type of dry materi-
Delivery als. The movement of cement down through the hopper
(into mixer
hopper) is assisted by the high-pressure flow of water through the
jets. The resulting pressure drop pulls the dry cement
into the stream of water. To reinforce this effect, the
Fig. 11-13. Surge tank. gooseneck can be given a venturi tube profile. Further
along at the gooseneck, turbulent flow mixes the cement
particles with the water, and the result is a cement
slurry.

Mixing water
(from the mixing Slurry (to the
manifold) displacement
pump(s)

Additional
Knife water
Sack
Suction
Gooseneck pipe
Hopper Dry Grating
cement
Two or three jets
y
Slurr
Cutting Bowl
table

Mixing water Slurry tub (200 L)

Fig. 11-14. Conventional jet mixer (sacked cement).

Chapter 11 Cementing Equipment and Casing Hardware 351

 Mixing water
Surge tank or (from the mixing Slurry [to the
atmospheric manifold) displacement
tank pump(s)]

Additional
water
Butterfly valve
Suction
Sock pipe
Dry cement Gooseneck
Hopper Grating
Two or three jets

Bowl

Mixing water Slurry tub (200 L)

Fig. 11-15. Conventional jet mixer (bulk cement).

The slurry density is adjusted by using the bypass ■ The slurry density is adjusted by operating the sliding
system to change the water-to-cement ratio. As the gate.
bypass is opened, the suction effect decreases and ■ The slurry is removed from the slurry tub by a recir-
reduces the amount of cement drawn out of the hopper. culation jet, fed by a centrifugal pump. The centrifu-
At the same time, the water bypassing the jets enters the gal pump force feeds the displacement pumps and
slurry. The combined effect is a decrease in slurry den- recirculates some slurry through the mixing system.
sity. Conversely, if the bypass is closed, the density ■ Water is always injected ahead of the recirculation
increases.
jet.
The conventional jet mixer can be operated at low
(175 to 200 psi [12 to 14 bar]) or high (880 to 1,180 psi Recirculation through the mixer heart and the tub
[60 to 80 bar]) water pressure. In the first case, the mix- improves the homogeneity and rheology of the slurry.
water pump is a centrifugal pump. In the latter case, it Adjustment of the slurry density is also easier.
is a reciprocating pump, usually identical (except per-
haps in plunger size) to the displacement pump. The
“double high-pressure pump cementing units,” which 11-2.7.3 Recirculation mixer without conventional
are the most widely used throughout the world, are jets
equipped to mix at either low or high pressure. The low- Available equipment includes a variety of mixers without
pressure method is preferred for two main reasons. Less conventional jets (Fig. 11-17). The maximum capacity of
horsepower is required and, because both high-pressure most mixers, evaluated by rate of dry material, is close to
pumps are available to displace the slurry, higher mixing 4,400 lbm/min [2 SI tons/min]. They all consist of the fol-
and displacement rates are possible. With the high-pres- lowing.
sure method, the jets and the bowl-and-gooseneck ■ A sophisticated metering system to mix cement with
assembly are less apt to become plugged with dirty water and a device to mix the resulting slurry with
mixing water or poor-quality cement. previously mixed slurry from the mixing tub
■ A centrifugal pump or similar device (located at the

11-2.7.2 Recirculation jet mixer bottom of the tub) to improve the initial mixing by
The maximum capacity of the recirculation jet mixer shearing, ensure recirculation through the mixer, and
(Fig. 11-16) is slightly more than 4,400 lbm/min feed pressurized slurry to the downhole pump
(2 SI ton/min). The recirculation jet mixer differs from ■ A mixing tub that can be divided into two sections,

the conventional type in several ways. each of which can be equipped with a stirrer to
■ A remotely controlled sliding gate is present between
improve mixing, allowing a film-like flow over the
the hopper and the bowl. common partition that assists the release of
entrapped air

352 Well Cementing

 Depending on the model, the density of the slurry is cement is delivered to the mixer. Normally, the cement
remotely controlled by metering the cement and/or is transferred directly from a pressurized tank without
water. The water rate is usually kept constant, and the passing through a surge tank.
slurry density is controlled by altering the rate at which

Mixing water
(from the mixing
manifold) Slurry [to
Centrifugal the displacement
pump pumps(s)]

Dry cement

Suction
Remotely controlled pipe
sliding gate Grating
Tub
Jet(s) recirculating
line

Mixer
Slurry tub recirculating
line

Fig. 11-16. Recirculation jet mixer.

Bulk cement

Cement
mixing valve
Mixing water
Water metering
valve (annular) Slurry

Recirculation
line

Centrifugal
pump

Slurry (to the

displacement Deaeration Mixing tub
pumps) partition
(optional)

Fig. 11-17. Recirculation jet mixer (without jets).

Chapter 11 Cementing Equipment and Casing Hardware 353

11-2.7.4 Cement mixing units Front view
As discussed above, the batch-blending system and the Vent
liquid-additive metering system have been designed to Cement Missing water
solve the proportioning problems encountered with Surge (from the mixing
or blend tank
cementing materials. However, the slurry properties are manifold on the
affected not only by the proportions of cement, water, cementing unit
and additives, but also by the shearing that occurs
during mixing.
New
Proper operation of a mixing unit should help blend slurry
the appropriate quantities of the cement blend and the
mix water. The correct ratio will provide the expected Cement
mixer
slurry density and other properties. Continual verifica- Holding
tion of slurry density is essential; however, some density tank
fluctuations during slurry mixing are unavoidable. The
longer the mixing time and the larger the slurry volume, Recirculating
the better the homogeneity of the resulting slurry. pump
Finally, the slurry should be given the proper amount
of shearing, which is a function of mixing energy and
mixing time. Because a centrifugal pump is an ideal Top view
shearing device, it is advisable to increase the volume of
slurry being recirculated so that such a pump can be Slurry [to the
downhole Recirculation
used. pump(s)] slurry
Recirculating mixers are available in a variety of con-
figurations (skid-, truck-, or trailer-mounted, diesel- or
electric-powered, sometimes soundproofed) and sizes.
They all have certain common features (Figs. 11-18 and Fig. 11-18. Single 6-bbl tank mixing unit (with conventional jet mixer).
11-19).

Cement
From mixing or blend
manifold Surge
Mixing tank
water

Holding Recirculation Holding

Tank No. 1 jet mixer Tank No. 2
Stirrer Stirrer

Densitometer

Slurry Slurry
To downhole To downhole
pump Centrifugal Centrifugal pump
Pump No. 1 Pump No. 2
Drain Drain

Available Available
Fig. 11-19. Twin-tank mixing unit (with recirculation jet mixer).

354 Well Cementing

■ A surge tank with a capacity of 9.4 to 25.2 bbl [1,500
to 4,000 L] Discharge
■ A conventional or recirculation jet mixer port
■ One or two holding tanks with a capacity ranging Discharge
from 6.3 bbl [1 × 1,000 L] to 50 bbl [2 × 8,000 L] (size valve
is limited by transportability concerns)
■ Two recirculating centrifugal pumps (only one on the
smallest units), with a maximum displacement rate Plunger
of 25 bbl/min [4,000 L/min], to either circulate the
slurry in the holding tank(s) to improve shearing and
homogeneity or to feed the slurry to the downhole Suction
pump stroke
Suction
■ A pair of paddle stirrers, hydraulically or electrically valve
driven, to maintain homogeneity
■ A manifold, sufficiently versatile to be used in a vari-
ety of combinations
In particular cases such as very small jobs, or when
the proportions of additives and the slurry density are
very critical, the total volume of slurry needed to com-
plete the job (including the usual excess) is prepared
before pumping downhole. The liquid additives are not
metered as described earlier; instead, they are released Plunger
directly into the tank or added through a jet mixer. No Discharge
special practice exists regarding dry additives. The vol- stroke
umes or weights of dry additives are usually measured in
the conventional manner.
Suction manifold
(not shown)
11-2.8 High-pressure pumps
All current cementing pumps are of the reciprocating Fig. 11-20. Reciprocating plunger pump fluid end (section).
type, mostly with three plungers (triplex) and spring-
loaded suction and discharge valves (Fig. 11-20). The
transformation of the rotating motion of the input shaft pressure and flow ratings without modifying the maxi-
into the reciprocating motion of the plungers is gener- mum available horsepower. The plungers used in
ally accomplished either by a system of a crankshaft and cementing usually have a diameter between 3 and 6 in.
connecting rods or by a swash plate and connecting rods [7.6 and 15.2 cm].
system. These pumps include an internal fixed-ratio
speed reducer. Depending on the make and the model,
the plunger stroke varies from 5 to 10 in. [12.5 to 25 cm]. 11-2.8.2 Hydraulic horsepower
The global efficiency of triplex pumps is 85% to 90%. Depending on the make and the model of the pump, the
If adequately pressurized, the volumetric efficiency can maximum horsepower varies between 200 and 500 hhp
reach 98% with water at 80% of the maximum speed. The [150 kW to 370 kW].
construction is particularly rugged, allowing the pumps
to handle the heaviest and most abrasive slurries.
11-2.8.3 Versatility
Heavy-duty triplex pumps, which can handle gravel-
11-2.8.1 Convertibility laden fluids, can also perform hydraulic fracturing
Depending upon the manufacturer, the “size” of a pump treatments. In the 200-hhp to 500-hhp [150-kW to
can be altered by changing either the fluid-end assembly 370-kW] range, the same pumps are used for cementing
or the plungers and packing system, using adapters to and stimulation.
convert the fluid-end body. Size alteration changes the

Chapter 11 Cementing Equipment and Casing Hardware 355

11-2.8.4 Maximum flow rate and pressure Slurry density: The slurry density is traditionally
Available makes, models, and sizes offer a large variety measured manually by a pressurized mud balance
of specifications. One should bear in mind that, during (Fig. 11-21). More sophisticated systems determine the
most cement jobs, only one pump is used to pump slurry density with a mass flowmeter that employs the
downhole. Usually the maximum rate is 8 bbl/min Coriolis effect (Fig. 11-22) or with a radioactive densito-
[1.3 m3/min]. This limitation is based upon the maxi- meter connected to a central computer-based recording
mum allowable rate for the 2-in. treating line most com- unit (Fig. 11-23).
monly used for cementing.
The pump pressure usually does not exceed 1,030 psi
[70 bar] (cement squeezes excepted). As a matter of
fact, if the density of the cement slurry is equal to that
of the drilling mud, the pumping pressure is simply a
consequence of the friction losses in the surface equip-
ment (steel flow hoses and cement head) and below the
surface.
Fig. 11-21. Pressurized mud balance (photo courtesy of Chandler
11-2.8.5 Drive Engineering, LLC).

The pumps on mobile units are driven by a diesel engine

and may have either automatic or manual transmissions.
Those permanently installed on an offshore rig are fre-
quently driven electrically (usually with a directly cou-
pled DC motor).

11-2.9 Controls and instruments

At or before the start of a job, some control devices on
the mixer are selected (e.g., chokes on a jet mixer) or set
in position (e.g., a mix-water valve on most recirculating
mixers) according to the composition and density of the
slurry and the desired injection rate. During the job, the
final adjustment is made with either the cement (blend)
or the mix-water metering valve, depending upon the
type of equipment. Adjustment of the downhole pump
rate may also be necessary to maintain a constant level
in the slurry tank and maintain the pumping pressure
within fixed limits (e.g., squeeze jobs).
Cement jobs require the measurement of many para-
meters. A discussion of the various types of measuring
equipment is given below.
Mix water: The volumes of water are measured in the
displacement tanks.
Cement (blend) and slurry: The volumes of mixed
slurry and dry cement are calculated from measure-
ments of the mix-water volume and the slurry density. Fig. 11-22. Coriolis effect mass flowmeter for slurry-density
Flow rate: The slurry flow rate is observed at the measurement.
downhole pump-stroke counter. A flowmeter is used if
job parameters are being recorded continuously.
Pressure: The pumping pressure is read at a gauge or
mechanical recorder. An electronic pressure transducer
is used if the various parameters are recorded by a cen-
tral unit.

356 Well Cementing

 Pipe Three key parameters are monitored: water flow rate,
Radioactive
mixing-tub level, and slurry-flow rate. The water- and
source slurry-flow rates are measured with electromagnetic
Amplifier Panel
To power flowmeters. The mixing-tub level is monitored by an
supply and ultrasonic-guided wave sensor. The solids fraction can be
Detecting processor calculated by computing the difference between the
cell slurry-flow rate and the water flow rate. The mixing-tub
level is important because the volume of slurry in the tub
Gamma can fluctuate, and the system must be able to instantly
rays compensate for such fluctuations as it calculates the
solids volume fraction.
Slurry rheology: Rheology measurements are not
Fig. 11-23. Schematic diagram of radioactive densitometer. routinely performed at the wellsite. However, a couette
viscometer (Appendix B) can be used at the wellsite for
ambient-temperature rheology measurements.
Compressive strength: This measurement is rarely
Solids fraction: With the advent of engineered parti- made at the wellsite because of the long curing times.
cle size cement systems, it has become possible to Slurry samples are normally taken to a central labora-
design high-performance cement systems at slurry den- tory for postjob tests.
sities approaching that of water (Chapter 7). Under Computer-based central recording units are available
these circumstances, the densities of the mix water, the to continuously record vital pumping parameters
cement blend, and the final slurry are essentially identi- (Fig. 11-25). The recorders significantly improve onsite
cal. Therefore, the mixing process cannot be monitored job monitoring while simultaneously storing data for
or controlled by conventional slurry-density measure- postjob evaluation. The stored data can be used to pro-
ments. To overcome this problem, a different quality- duce job-data sheets and enhanced graphics for in-depth
control principle based on solids fraction measurement analysis. Slurry density, flow rate, and pressure can be
was developed (Vigneaux et al., 2003). monitored for various configurations of pumping equip-
The solids fraction of a slurry is the percentage of ment. The microprocessor calculates and displays cumu-
solids contained in slurry sample. It can be expressed in lative volumes, and the total volume is printed on the log
terms of volume or mass. Expressed in terms of volume, in increments as short as one second.
the solids fraction is 1 minus the inverse of slurry poros- One can imagine that the continuous monitoring of
ity (Chapter 7). key pumping parameters could be used to automatically
A schematic diagram of equipment to monitor the control slurry mixing. As of this writing, equipment
solids volume fraction is shown in Fig. 11-24. applying such a principle is beginning to appear in the

Surge
can

Tub level

Water
flowmeter
Slurry
flowmeter
Mixing
tub
Mixer Mixing water

To triplex
pump

Fig. 11-24. Schematic diagram of solids fraction monitoring equipment (Vigneaux et al., 2003). Reprinted with
permission of SPE.

Chapter 11 Cementing Equipment and Casing Hardware 357

Fig. 11-25. Computer-based monitoring system for cement jobs.

field. In the near future, automatic cement mixing will Articulated (loop) section
undoubtedly become a routine procedure.

Three Two
11-2.10 Steel flowhoses and cement head swivels swivels
A “cement head” (Section 11-5.14) is screwed into the
top casing collar or landing joint, depending on the type
of cement job. The discharge side of the downhole pump Half union Half union
and the cement head are connected by a series of artic- (male) (female)
ulated or straight sections of high-pressure steel pipe,
also known as “treating iron” (Fig. 11-26).

Straight section

Fig. 11-26. Steel flow hoses.

358 Well Cementing

11-3 Cementing units
The various components of cementing units, which fab-
ricate and inject the cement slurry, have been described
individually in Section 11-2. Figure 11-27 illustrates the
combination of the components to assemble a basic
cementing unit. A variety of configurations and composi-
tions exists, tailored to the type of rig to be serviced and
the redundancy, versatility, and mobility required.
The various configurations are described below,
according to the type of rig to be serviced.
Skid-mounted units: Illustrated in Fig. 11-28, skid-
mounted units are most applicable to isolated land rigs,
offshore rigs, cementing barges (lakes and rivers), and
open-sea cementing vessels.
Truck-mounted units: Shown in Fig. 11-29, such units
are suitable for almost any land rig. However, the chassis
must be adapted to the type of surface upon which the
unit will travel. The “standard” unit is designed to travel
on roads and must conform to local road regulations. The
“off-road” unit is built for more difficult terrain. The
“desert” unit can be driven over soft surfaces, even sand
dunes.
Fig. 11-28. Typical cementing skid and control console.

Liquid additives from Cement and

the liquid additive Basic cementing unit dry additives
metering system
3 6

2
From rig
4 storage

Slurry
Water from to well
rig storage 9
1 5
10

7 8
1. Centrifugal water supply pump 6. Mixing water manifold
2. Water distributor 7. Cement mixer (conventional jet mixer shown)
3. Additive distributor 8. Slurry tub
4. Displacement tank system 9. Centrifugal pressurizing pump
5. Mixing water pump (centrifugal—low-pressure 10. Reciprocating displacement (downhole) pump(s)
mixing; reciprocating—high-pressure mixing)

Fig. 11-27. Mixing and pumping equipment on rig site (typical setup).

Chapter 11 Cementing Equipment and Casing Hardware 359

 However, single-pump units now exist. One such unit
is shown in Fig 11-31. This unit provides a cost-efficient
solution for less-critical, lower-pressure well casing jobs.
High-pressure treating iron is replaced with a flexible
hose for faster wellsite rig-up.

Fig. 11-29. Typical cementing truck.

Semitrailer-mounted units: Like the truck-mounted

units, semitrailer-mounted units as shown in Fig. 11-30
are appropriate for almost any land rig. They can be
drawn by many types of tractors, providing a logistical
advantage. A heavy tractor-drawn unit with five axles has Fig. 11-31. Single-pump cementing unit.
a better weight distribution ability than the correspond-
ing truck with only three. The maximum authorized pay-
load is greater than that of the truck, which allows the For economic reasons, a cementing unit is designed
loading of more equipment on the same chassis. to meet the requirements (including road regulations)
Helicopter units: Helicopter units are intended for of as many locations as possible. In Europe, for example,
rigs totally inaccessible by land or water. The units, the the required specifications vary from one country to
mixing equipment, and the cement silos are designed to another, and the unit must conform to the most stringent
be transported by helicopter. They can be dismantled regulations. In addition, soundproofing is more fre-
into smaller components, incorporating lifting frames, quently demanded because of the proximity of wells to
and are often made of lighter materials to reduce weight. residential areas (Fig. 11-32).
Traditionally, a cementing unit contains two of each Cementing-unit designs are developed with special
vital item. This redundancy is necessary because a well attention to safety and environmental requirements.
can be severely damaged or lost if it becomes impossible Environmental release of wastes (liquid or dry chemi-
to complete a job after it has commenced. The extra cals, cement slurries, or pump or engine oil spillages) is
equipment serves as an “insurance policy” to protect the prevented by better equipment design and the use of
operator’s investment. recovery receptacles.

Fig. 11-30. Semitrailer-mounted cementing unit.

360 Well Cementing

 Surge tank Control cabin Roof Soundproof power
unit (diesel)

Cement

Recirculation
jet mixer
(not shown)

Pressurizing
pump Slurry tub Low-pressure Two high-
mixing pump pressure pumps

Water Water supply pump Double displacement tank

Water distributor

Mix water manifold

(inside cabin/not shown) Slurry

Fig. 11-32. Semitrailer-mounted cementing unit (Europe).

The safety requirements with which the equipment 1. Special water-cooled manifold rated to cool exhaust
should comply depend upon the location and are espe- gas to 200°C (392°F) maximum, and with a surface
cially dependent upon possible sources of flammable or temperature not exceeding 200°C at any point.
explosive gases. Whenever the unit can be placed more 2. Oversized radiator.
than 98 ft [30 m] away from the well (as on most land rig
3. Inlet air combustion, slam-shut valve.
sites) there are no special requirements. Standard
equipment can often be used without modification. This 4. Inlet air flame trap.
distance condition is often difficult to satisfy on offshore 5. Exhaust gas spark arrestor, DNV type approved.
rigs, where every compartment or deck location is clas- 6. Overspeed valve, which closes the engine blower
sified according to the potential risk of explosion or fire. flapper valve when speed exceeds the normal maxi-
The classification is made by official regulatory bodies mum by 10%.
according to standards that may vary slightly from one
7. High-water-coolant temperature valve, which shuts
country to another; however, operators usually adhere to
down the engine when water temperature exceeds
the most stringent regulations.
95°C (204°F). Fuel rack actuated.
For example, the following is a summary of the Det
Norske Veritas (DNV) requirements for diesel engines to 8. Low-water-coolant level valve, which shuts down the
be located in a hazardous area, classified as “Zone 2,” in engine.
which an explosive gas mixture may exist for a short time 9. High-exhaust-gas temperature valve, which shuts
only under abnormal conditions. Diesel engines are ban- down the engine when the gas temperature exceeds
ished from Zones 0 and 1, which are more sensitive 200°C.
areas. The DNV is the Norwegian certification body, and 10. Special control panel.
its standards serve as a reference in the North Sea.

Chapter 11 Cementing Equipment and Casing Hardware 361

 Diesel engines must often be adapted further to meet Years ago, most casing hardware was manufactured
fire-protection standards. The equipment required to and sold by the cementing company, and the operator
adapt an engine is sophisticated, entirely made of high- depended upon the service companies to develop new
quality stainless steel, and bulky. It is extremely expen- tools. Today, a multitude of engineering groups and man-
sive. Electric motors for Zone 2 areas are typically con- ufacturers supports and supplies the cementing compa-
fined in a closed shelter, which is pressurized with air nies with most casing hardware.
taken from a safe area. An overpressure is maintained so Some cementing tools are retrievable devices that
that no gas from the hazardous area can enter the shel- may require some form of operation from the surface.
ter. When a cementing unit is to be operated offshore in These tools are usually operated by specialists or tool
a nonhazardous area, the drilling and service companies operators within the cementing companies and are typi-
often opt for “protected” diesel engines, which provide cally associated with remedial cementing operations
increased security at a more reasonable cost. Following such as squeeze jobs or pressure testing. Packers, bridge
is a list of the devices that should be installed on a stan- plugs, and retainers are examples of cementing tools
dard diesel engine to ensure this protection. that are commonly used for squeeze or plugback
1. Overspeed valve, which closes the engine blower flap- cementing.
per valve when speed exceeds the normal maximum
by 10%.
2. High-water-coolant temperature valve, which shuts 11-5 Casing hardware
down the engine when water temperature exceeds A typical application of casing hardware for a primary
95°C. Fuel rack actuated. cement job of moderate depth is shown in Fig. 11-33. The
3. Low-oil-pressure valve, which shuts down the engine lower end of the casing is protected by a guide shoe or
when engine oil pressure is below a value to be set- float shoe. A float collar or autofill collar is placed one
tled with the manufacturer. Fuel rack actuated. or two joints above the shoe to provide a seat to land
cement plugs and to halt cement displacement. The
4. High-oil-temperature valve, which shuts down the
short section of casing between the shoe and float collar
engine when oil temperature exceeds 130°C (266°F).
is called the shoe joint and is provided as a buffer within
Fuel rack actuated.
the casing to contain fluid contamination that may build
5. Special control panel. up ahead of the top wiper plug. The length of the shoe
joint may be as few as 1 joint (40 ft [12 m]) or as many
as 10 joints, but most operators use 2 joints. The length
11-4 Introduction to casing hardware of the shoe track is sometimes adjusted on the rig to
accommodate a lack of confidence in the accuracy of the
This section focuses on equipment used on or within the
rig pump efficiency. Rig pumps are often used to displace
casing string to enhance casing placement and cement-
the cement and plug(s). Plugs act as physical barriers to
ing operations. Within the petroleum industry, this
separate the cement slurry from the drilling mud and
equipment is commonly referred to as casing and liner
displacement fluids.
hardware and cementing tools. These tools fall into five
Centralizers are placed in critical sections to prevent
basic categories.
sticking while the casing is lowered into the well. Then
■ Float equipment
they keep the casing in the center of the borehole to aid
■ Centralization and flow enhancement tools mud removal and help ensure placement of a uniform
■ Elastomers—casing or liner wiper tools and cement sheath around the casing in both the openhole
inflatables and cased hole sections.
■ Stage tools

■ Torque and drag reduction tools 11-5.1 Guide shoes and float shoes
An entire textbook could be written about these tools; Guide shoes and float shoes are tapered, commonly
for this textbook, the discussion will be limited to the bullet-nosed devices that are installed at the bottom of
most common or basic types, with the emphasis placed the casing string. They guide the casing toward the
on application, principles of operation, and basic design center of the hole to minimize hitting rock ledges or
characteristics. washouts as the casing is run into the well. The outer
portions of these shoes are usually made from steel, gen-
erally matching the casing in size and threads. The

362 Well Cementing

 Rigid centralizer

Liner hanger

Casing packer

Spring-bow centralizer

Tandem-rise centralizer Low-drag roller centralizer

Stage tool

Centralizer sub
Semirigid centralizer
Reamer shoe

Fig. 11-33. Typical application of casing hardware for a primary cement job (drawing courtesy of
Weatherford International).

inside (including the taper) is generally made of con-

crete or thermoplastic, because this material must be
drilled out if the well is to be deepened beyond the
casing point. Guide shoes and float shoes are made in a
wide size range, from 31⁄2 to 24 in. [8.9 to 61 cm] in diam-
eter, and they are generally used in surface-to-interme-
diate casing strings at shallow depths.
The guide shoe differs from the float shoe in that it
lacks a check valve. The check valve in a float shoe can
prevent reverse flow, or U-tubing, of cement slurry from
the annulus into the casing. The float shoe also reduces
hook weight, because the check valve increases the
buoyancy of the casing string by preventing backflow of
fluid as the casing is lowered into the well.
A basic guide shoe is shown in Fig. 11-34. The steel
Fig. 11-34. Basic guide shoe (drawing courtesy of Weatherford
shell is molded into a rounded nose with a central orifice. International).

Chapter 11 Cementing Equipment and Casing Hardware 363

 A variation is the “down jet” guide shoe. This modifi-
cation incorporates some side jets or ports to divert all
or part of the fluid (Fig. 11-35). The purpose of the jets
or ports is to promote high displacement efficiency by
distributing the fluids evenly around the annulus. The
jets come in a wide variety of sizes and shapes. The con-
crete nose is fairly strong, yet can be easily drilled with
polycrystalline diamond compact (PDC), insert, and
roller cone bits.

Fig. 11-36. Aluminum nose float shoe (drawing courtesy of

Weatherford International).

Fig. 11-35. Cement nose guide shoe with down jet option.

11-5.1.1 Nose materials and shapes

Guide shoes and float shoes are available with various
nose designs, each built for a specific purpose. Most
common among these is the cement nose guide shoe.
The slightly rounded nose is usually made with high-
strength concrete that has a compressive strength of
8,000 and 10,000 psi [5.5 and 6.9 MPa]. The rounded
nose allows the casing to be deflected off of ledges and
obstructions as the casing is lowered. The shells of guide
shoes and most float equipment have grooves into which
the cement flows when poured. As a result, the cement
is not held in place by shear bond alone, and it has the
strength to withstand the stresses associated with casing
placement and the subsequent cement job.
Aluminum nose shoes are another option; however, Fig. 11-37. Eccentric composite-nose float shoe (drawing courtesy
their popularity has been diminished by the introduction of Weatherford International).
of concrete and composite materials. Aluminum noses
can be cast into a wide variety of shapes. Lengthened
noses can ease movement past bridges and obstructions
typical 9 5⁄8-in. eccentric nose float shoe can withstand
downhole (Fig. 11-36).
end loadings greater than 100,000 lbm [45 SI ton] at
Composite nose guide shoes are the newest option,
temperatures exceeding 250°F [121°C]. Their perfor-
made possible by the advent of composite materials that
mance exceeds the American Petroleum Institute (API)
can withstand high load weights and resist abrasion, ero-
specifications, and they can withstand running into well-
sion, and high temperatures (Fig. 11-37). For example, a
bores up to and beyond 90° inclination.

364 Well Cementing

 Other types of guide shoes include the Texas pattern 11-5.2 Float equipment
(beveled nose) and the sawtooth pattern (Fig. 11-38). The primary purpose of float equipment is to be able to
The noses of these guide shoes are formed or cut out of pump cement slurries into the well that are heavier then
the steel shell. The beveled nose shoe has the same out- the mud. Check valves prevent the cement from U-
side diameter (OD) as the couplings and an inside diam- tubing back into the casing or liner string (Fig. 11-39).
eter (ID) that is the same as the casing or drillout diam- Float equipment consists of specialized casing shoes
eter. The bevel is advantageous in situations in which it and collars with check valves to prevent wellbore fluids
is necessary to pull back the bottomhole assembly from entering. As the casing is lowered, the hook load or
through the previous casing. The beveled guide reduces hanging weight is reduced by the weight of fluid dis-
the probability of sticking. placed. The casing is filled from the surface, and the
The beveled nose shoe is most often run in shallow, hook load or amount of buoyancy is controlled by moni-
vertical wells. Such bevels are also incorporated into toring the weight indicator. The frequency of filling is
most float shoes. generally once every 5 to 10 joints; however, some large-
In recent years, the price of float shoes has diameter or thin-wall casings may require more frequent
decreased, and many operators use them routinely to filling to prevent casing collapse. In addition to proper
have the option of an additional valve in the casing filling, the casing should be lowered at a slow, steady
string. However, in deepwater applications, guide shoes rate to prevent pressure-surge damage.
are making a comeback. The guide shoes leave an unre-
stricted path for tripping devices to pass through.

Texas pattern

Fig. 11-39. Typical float shoe (drawing courtesy of Davis-

Sawtooth pattern Lynch, Inc.).
Fig. 11-38. Texas pattern and sawtooth pattern guide shoes
(drawing courtesy of Weatherford International).

Chapter 11 Cementing Equipment and Casing Hardware 365

 As the demand for larger and heavier casings has Circulating at least one hole volume of mud is typi-
increased, so has the concern for derrick stress and cally required before performing the primary cement
fatigue. Float equipment can be used to reduce derrick job; however, to optimize hole and mud conditions for
stress by inducing flotation or increased casing buoyancy cementing, some drilling programs call for many hours
by simply not filling the casing or liner while running into of circulating (Chapter 5). Most float equipment valves
the wellbore. After landing, the casing is filled and cir- have been rated (after testing according to Recom-
culation is slowly established to begin hole conditioning. mended Practice for Performance Testing of Cementing
Flotation devices must be carefully deactivated, slowly Float Equipment [API RP 10F]) to withstand more than
releasing trapped air inside the casing; otherwise, casing 24 hr of solids-laden fluid flow at 10 bbl/min and to hold
collapse may occur. A specialist from the vendor supply- 5,000 psi of backpressure after circulation is complete.
ing the flotation device is usually present to monitor its After the cement is displaced, the float valve must
use. A typical casing flotation system is shown in prevent backflow into the casing. Should the float valve
Fig. 11-40. The operating sequence is shown in Fig. 11-41. fail, surface pressure must be applied until the cement
Owing to the risk of casing collapse during the flota- hardens to prevent U-tubing. Applying surface pressure
tion of larger strings, some operators and consulting is undesirable because it expands the casing while the
groups are studying casing drag more closely. In shal- cement hardens. When the pressure is released, the
lower, high-angle wellbores, casing drag and differential casing relaxes, potentially creating a microannulus
sticking frequently impede pipe placement to total between the casing and cement.
depth (TD). Roller centralizers are being proposed as a
safer method to reduce friction (Section 11-5.11).

Mud- Special Special

filled Flotation application application
casing collar float collar float shoe

Air-filled Semirigid
casing centralizer
Mud

Fig. 11-40. Typical casing flotation system (drawing courtesy of Davis-Lynch, Inc.)

366 Well Cementing

 Bottom
Opening plug
sleeve
Shear
pins

Seals

Circulating ports
Bottom sleeve

1. Bottom portion of casing is run dry 2. Casing pressure is increased until 3. Bottom cementing plug is launched
(not filled with fluid), with flotation collar the opening sleeve shifts down to ahead of cement. After landing on the
installed at desired depth. Casing above permit fluid and air to swap. After a bottom sleeve, it pushes both sleeves
the collar is filled with drilling fluid as fluid stabilization period, the casing ahead of the cement to the float collar
casing run continues to desired depth. is filled with drilling fluid. below.

Top plug

Casing
Casing
Bottom
plug
Bottom
plug

Float collar

Float collar

4. Bottom cementing plug and sleeves land and seal on 5. Top cementing plug seals and locks on bottom
the float collar. Bottom cementing plug ruptures, and cementing plug/collar assembly at the float collar.
cement is pumped through and out of the float equipment.

Fig. 11-41. Operating sequence of casing flotation system (drawing courtesy of Davis-Lynch, Inc.).

Chapter 11 Cementing Equipment and Casing Hardware 367

11-5.3 Valve materials and types
The three most common types of check valves used
in casing hardware are the ball, flapper, and poppet
valves. They have been adapted for floating and autofill
applications and are commercially available from sev-
eral suppliers. There are many specialized variations,
the latest of which is the large-bore double-flapper valve,
used for surge reduction in deepwater applications. The
large bores offer less resistance during the autofill
process.
API RP 10F provides standard methods for equipment
testing and a classification system based on perfor-
mance capabilities. On scales of I, II, and III and A, B,
and C, the best equipment rating is III-C.
Materials used in valves vary depending on require-
ments for drillability, compatibility with bits, and tem-
perature stability. The major valve components are gen-
erally made of phenolic resin, composites, or aluminum. Ball valves are still used Float collar sized for
Nitrile rubber is used for valve seals and coatings rated primarily in smaller float ball valve tubing.
for service at temperatures up to 300°F [149°C]. Beyond equipment, like this tubing
float shoe.
that temperature, viton rubber is used. Most large
springs are made from a phosphorus bronze, beryllium, Fig. 11-42. Tubing-size ball valves (drawings courtesy of
or aluminum alloy. Steel parts such as springs and pins Weatherford International).
are generally too small to pose a problem; nevertheless,
they are always tested to verify that they are drillable
with PDC bits. Plastic parts are typically made from als used and the construction of the valve. The use of ball
thermoset materials such as phenolic resins that are valves as the primary valve for cement placement is
stable to temperatures as high as 450°F [232°C] and waning. Poppet valves (described later) are more popu-
inert to most oils and synthetic mud components. They lar today because of their superior erosion resistance.
are easily drilled and are compatible with most bits. Flapper valves are composed of a spring-loaded flap-
Aluminum is easily drilled, but it has a tendency to foul per hinged to a plate with an integral seat (Fig. 11-43).
or smear over some diamond or fine-cutter bits when The flapper and plate are generally made of aluminum or
found in large masses unless the PDC bit is modified to composite materials. The spring-loaded seating action is
drill soft metals. virtually unaffected by hole deviation. The flapper opens
Ball valves consist of the typical ball-and-cage proportionally to the rate at which fluid is being pumped
arrangement (Fig. 11-42). The ball is generally made and the strength of the spring in the flapper. Because of
from weighted phenolic resin and may be rubber-coated. the irregular surfaces behind the flapper and spring,
The cage may be made of plastic or aluminum. The ball smaller valves tend to induce a large amount of turbu-
is buoyant in many drilling fluids and cement slurries; lence during flow. This increased turbulence can easily
however, when the hole is deviated, reverse flow may be cause erosion and seal damage when the flapper is not
required to return the ball to its seat. Ball check valves completely open, leading to a short valve life.
may not be effective in preventing a slight reverse flow A variety of flapper valves is available to fit within
or gas migration in highly deviated holes. most API threads such as 8-round or buttress as well as
The minimum clearance between the ball and the cemented versions of conventional shoes and collars.
cage may be one-half inch or smaller. Solids such as mud The new generation of composite large-bore flappers
scale or cuttings should be avoided to prevent wedging of (diameters greater than 3 in. [7.6 cm]) opens to provide
the ball or plugging the valve. Likewise, ball valves are a large, unobstructed flow path that passes tripping balls
not recommended for use in muds containing heavy con- and extraneous materials without blockage or harm to
centrations of lost circulation materials (LCM). The the valve. The temperature rating of composite flapper
backpressure resistance of the ball valves is more than valves is less than 300°F [149°C]. Aluminum valves are
adequate for most applications. The temperature stabil- available for service at higher temperatures.
ity and wear resistance depend greatly upon the materi-

368 Well Cementing

 Flow path during normal Backflow path blocked by
circulation, in which spring-loaded poppet (cone).
hydraulic force against
the poppet pushes it
down and compresses
the spring.
Aluminum nose flapper Flapper valve float collar with
float shoe with autofill autofill tube installed and Fig. 11-44. Poppet valve (drawing courtesy of Weatherford
tube installed and conversion ball shown above. International).
conversion ball shown
above.
the poppet should be avoided. The spring-loaded seating
action is virtually unaffected by hole deviation. The tem-
perature resistance depends on the type of materials
used and should be specified by the supplier. The pres-
sure resistance typically exceeds that of the flapper and
ball valves. Wear resistance is also generally good but
should be verified before using poppet valves in
Insert flapper equipment as used with extended-circulating-time applications.
8-round and buttress casing. The poppet spring is usually a nonferrous material
Fig. 11-43. Flapper valves (drawings courtesy of Weatherford such as phosphorus bronze or beryllium copper.
International). Generally, softer alloys are used for the spring materials
and are chosen according to the temperatures they will
experience. It is widely assumed that the springs in
Bump pressure ratings are based on the way the valve poppet valves are very strong, but this is far from true.
is cemented in place or the number of threads in the The purpose of the spring is to simply lift the poppet into
valve body. Backpressure resistance is highly dependent its seat. The differential pressure of the fluid trying to
on the valve body and is generally more than adequate for reenter the pipe holds the poppet in place.
most conditions. The wear resistance varies with size and There are two types of poppets on the market today.
manufacturer; it should be specified for circulating appli- The more common type, shown in Fig. 11-44, has the
cations longer than 8 hr. head on the top of the stem. Another version has the
Poppet valves are composed of a spring-loaded head or sealing area on the bottom of the stem. This gen-
poppet (also referred to as a plunger or cone) housed in erally results in a shorter piece of equipment for the
a cage much like that of a ball valve (Fig. 11-44). float collar, but may pose a problem on the float shoe (for
Phenolic resin or aluminum is used for the cage, and the clearance).
poppet is often made from nitrile rubber or is rubber- Table 11-3 is a summary of the advantages and disad-
coated. Like the ball valve, the poppet valve has a vantages associated with the various types of valves.
restricted flow path. Solid debris that may plug or bind

Chapter 11 Cementing Equipment and Casing Hardware 369

Table 11-3. Advantages and Disadvantages of Various Types of Valves
Types of Valves Flapper Ball Phenolic Poppet or Cone
Advantages Inexpensive Simple Self-closing
Least restrictive Easy to use Erosion-resistant
Self-closing Easy to drill out 0–90° inclination
Easy to drill out

Disadvantages High erosion susceptibility Dependent on flow to close Low tolerance to some LCM (e.g., fibers)
Easily blocked Poor performance in inclined wellbores
Difficult to drill out High erosion susceptibility
Easily damaged

11-5.4 Surge pressure and float equipment

Pressure surges are generated each time the casing is
raised and lowered; they are the product of inertia and
flow resistance of the displaced fluid. Pressure surges
combined with hydrostatic-pressure differentials may
exceed the casing-collapse or the formation-fracture
pressures, causing loss of mud or permanent formation
damage. As the trend continues towards reducing the
number of casing strings in a well, many operators are
running reduced-clearance combinations such as
■ 17 7⁄8 -in. to 18-in. casing through 20-in. hole

■ 16-in. to 17 7⁄8-in. casing through 18-in. hole

■ 13 3⁄8-in. to 13 5⁄8-in. casing through 16-in. hole

■ 113⁄4-in. to 11 7⁄8-in. through 13 3⁄8-in. to 135⁄8-in. hole

Closed Open Autofill Autofill with
■ 9 5⁄8-in. to 9 7⁄8-in. casing through 113⁄4-in. to 117⁄8-in. hole diverter
■ 7-in. to 7 5⁄8-in. casing through 9 5⁄8-in. to 9 7⁄8-in. hole. Fig. 11-45. Different types of surge models (drawing courtesy of
Pegasus Vertex).
External attachments such as centralizers and recip-
rocating scratchers may increase the flow resistance and
should be considered when determining a safe lowering
speed. Fig. 11-45 shows four ways to run casing or liners
into a well. Pipe ending conditions include
■ closed pipe

■ fully open pipe

■ pipe with autofill restriction v = v–

■ pipe with circulating sub (flow diverter). v=0

In all of these scenarios, the surge-pressure calcula- v = –vpipe

tions are based on many assumptions. Many drilling
fluids are compressible; therefore, surge pressures
should be reduced compared to conventional steady-flow
models. Unsteady flow conditions are reduced to a sta-
tionary flow regime model by considering an improved
mud clinging constant (Fig. 11-46). Clinging constants vpipe
were addressed in earlier works by Burkhardt (1961)
and Fontenot and Clark (1974) and are incorporated
Fig. 11-46. Graphic modeling of clinging characteristics of fluids on
into much of the mud displacement software used today. moving pipe (drawing courtesy of Pegasus Software).

370 Well Cementing