See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/296638638

Properties of oil well cement with high dosage of metakaolin

Article  in  Construction and Building Materials · June 2016

DOI: 10.1016/j.conbuildmat.2016.02.173

CITATIONS READS

14 1,185

5 authors, including:

Yuhuan Bu Jiapei Du
China University of Petroleum - Beijing China University of Petroleum - Huadong
32 PUBLICATIONS   149 CITATIONS    9 PUBLICATIONS   33 CITATIONS   

SEE PROFILE SEE PROFILE

Shenglai Guo
China University of Petroleum
21 PUBLICATIONS   64 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Interfacial properties and nanostructural characteristics of epoxy resin in cement matrix View project

All content following this page was uploaded by Jiapei Du on 11 October 2017.

The user has requested enhancement of the downloaded file.

 Construction and Building Materials 112 (2016) 39–48

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Properties of oil well cement with high dosage of metakaolin

Yuhuan Bu a, Jiapei Du a,⇑, Shenglai Guo a, Huajie Liu a, Chenxing Huang b
a
College of Petroleum Engineering, China University of Petroleum, 266580 Qingdao, China
b
College of Pipeline and Civil Engineering, China University of Petroleum, 266580 Qingdao, China

h i g h l i g h t s

At 75 °C, the early-term strength of cement decreased with the increasing MK content.

At 75 °C, the set cement with more MK showed higher long-term strength.

Beyond 150 °C, samples M40 to M60 presented good strength in all test ages.

Hydration products and structure would influence strength and porosity of cement.

High dosage of MK cement is suitable for corrosion and high temperature conditions.

a r t i c l e i n f o a b s t r a c t

Article history: This research work was carried out to evaluate the properties of oil well cement with high dosage of
Received 23 October 2015 metakaolin. The fresh pastes were made using an initial water/solid ratio of 0.5 by weight and then
Received in revised form 4 February 2016 hydrated for periods up to 56 days. The mechanical performance was assessed from compressive
Accepted 23 February 2016
strength. X-ray diffraction (XRD) and scanning electron microscopy (SEM) were applied in order to study
the hydration products and qualitative of microstructure. Feasibility as oil well cement of metakaolin
replaced cement was assessed by thickening time, corrosive fluid exposure, density and rheology test.
Keywords:
Test results confirmed that the early-term strength decreased with the increasing metakaolin content
Metakaolin
Compressive strength
and the set cement with more metakaolin showed higher long-term strength at 75 °C. The cement pastes
Hydration products with 40–60%wt metakaolin presented good strength both in early-term and long-term curing at 150 °C
Microstructure and above. The metakaolin replaced cement pastes demonstrated low porosity and good strength due
Feasibility to stable hydration products and dense microstructure. Meanwhile, the filling effect of metakaolin par-
Oil well cement tials reduced porosity and improved corrosion resistance of hardened pastes. The MK replaced cement
pastes are applicative for the cementation of deep wells, acid rich wells and geothermal wells.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction fluids migrate in cracks and structure pores in cement sheath,

Portland cement has been used as well cement for many years.
However, Portland cement is not suitable for some critical circum-
stances, especially in deep oil wells, geothermal wells, and acid
rich wells. It goes through strength reduction, increased porosity,
and there are durability concerns in high temperature and corro-
sion environments where Portland well cement could lose its seal-
ing property in a short period of time [1–4]. Elevated temperatures
lead to gradual damage of connections between admixture and
cement paste while causing the generation of mineral with poor
structure, resulting in bad mechanical behavior [5]. Corrosive
⇑ Corresponding author at: College of Petroleum Engineering for Oil Well Cement
Research and Testing, Economic and Technological Development Zone, Changjiang
West Road 66, 266580 Qingdao, China.
E-mail address: jiapeidu@163.com (J. Du).
which lead to the increase of permeability and porosity. As a result,
an increase of area and contact time between the cement and cor-
rosive fluids occurs, accelerating the corrosion procedure [6].
Presently, pozzolanic material had already proved to be a mate-
rial which leads to high strength in early age, low porosity, high
temperature resistance and good chemical corrosion resistance
[7–9]. These materials react with Ca(OH)2 which is released in
the hydration of cement and convert Ca(OH)2 into high crys-
tallinity CSH (calcium silicate hydrates) phases according to poz-
zolanic reaction. The most common pozzolanic materials are fly
ash (FA), silica fume (SF) and metakaolin (MK). Due to the advan-
tageous capabilities, the pozzolanic material is a prospective sub-
stitution to ordinary Portland well cement. However, the stability
of performance and market price should be considered in cement-
ing operation. The properties of FA are distinct in different batches

http://dx.doi.org/10.1016/j.conbuildmat.2016.02.173
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
40 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48

Table 1 equal or even better than low MK dosage cement, the high dosage
Chemical composition and physical properties of OPC. of MK replaced cement will be a promising candidate as an alterna-
Oxides Wt% tive to oil well cement. Brooks et al. [18] reported that setting time
CaO 64.20 significant decrease as the dosage of MK increase. They explained
SiO2 19.40 this phenomenon is caused by the increasing water demand. This
Fe2O3 5.50 may be the bottleneck reason for not using high dosage of MK in
Al2O3 4.50 the industry. In a recent research by the authors, the setting time
SO3 2.80
MgO 2.00
issue with high MK dosage was solved by adding specific retarder
Na2O 0.10 and adjusting water-solid ratio. So it is possible to use high dosage
K2O 0.60 of MK replaced cement in the industry. Therefore, this study aimed
Loss on ignition 0.49 to produce cement mixture containing high dosage of MK which is
Density (g/cm3) 3.15
suitable for deep well, geothermal well and acid rich well cemen-
Specific surface area (m2/kg) 336.00
tation. The properties of Portland cement pastes containing high
dosage of MK were studied. This was done according to determina-
tion of compressive strength, porosity, hydration products and
Table 2
microstructure test. Moreover, the corrosion-resistance, thickening
Chemical composition and physical properties of MK.
time, rheological property and density of MK replaced cement
Oxides Wt% pastes were measured.
CaO 0.17
SiO2 55.06
2. Materials and methods
Fe2O3 0.76
Al2O3 42.12
2.1. Materials
SO3 0.15
MgO 0.06
The following raw materials were used: API class G oil well Portland cement
Na2O 0.06
(OPC) was obtained from Shandong Shengwei Enterprise Company (Linqu, China).
K2O 0.55
The chemical composition and physical properties of OPC are presented in Table 1.
Loss on ignition 1.20
This cement was used in preparing all samples. MK was obtained from Jiaozuo
Yukun Enterprise Company (Jiaozuo, China). Chemical composition and physical
properties of MK are shown in Table 2. Fluid loss control additive BXF-200L was
obtained from Company of CNPC (Tianjin, China). This additive reduces the fluid
because FA is manufactured from coal combustion [10,11]. Mean- loss. Dispersing agent FHJZ-1 obtained from Fuhai Industry Development Company
while, the price of SF which is more than 290$ per ton is not eco- (Dongying, China), was used to reduce the friction between particles. Retarder OAI,
nomical for large quantity cementing operation. Unlike other synthesized by Shenglai Guo, was used to adjust the thickening time of cement
slurry. The primary functional groups of OAI are maleic acid and itaconic acid.
pozzolans, MK, obtained by thermal treatment of kaolin, is much
more inexpensive than SF. Its price is approximately 140$ per
2.2. Methods
ton and the performance of MK is stable. In addition, the substitu-
tion of MK also can reduce the pollution of cement industry [12]. 2.2.1. Manufacturing process
Currently researches usually use MK as a cement substitute Cement slurries were mixed according to API Standard 10B-3-2004. Additives
with 5–20%wt [13–16]. Shatat et al. [13] incorporated the ternary replaced the cement in a certain proportion by weight. Mix design parameters of
pastes and their designation are given in Table 3.
blends of cement with 5–10%wt rice husk ash and 15–20%wt MK.
First, the fluid loss control additive, retarder and dispersing agent were dis-
The hardened cement blend showed better mechanical behavior solved in water and transferred into the cup of waring blender. Then, the blended
than ordinary Portland cement. Morsy et al. [14] studied the cement was added within 15 s to the aqueous solution with a stirring rate of
behavior of cement containing nano-MK at elevated temperatures. 4000 rpm and mixed for 35 s at 12,000 rpm. After being prepared, cement slurry
They indicated that replacement of cement with 5%wt nano-MK at was placed into 5 cm cubes for compressive strength test and pressurized consis-
tometer for consistency measurement.
25 °C leads to optimal mortar and 15%wt nano-MK replacement at
250 °C–800 °C gives better strength. Li et al. [17] compared MK
2.2.2. Compressive strength test
cement with alkali-activated slag cement, they indicated that
Considering the different in downhole conductions, temperature would be an
cement pastes containing MK get better strength at high important factor that affects cementing quality. Cement strength decreases when
temperatures. the temperature reaches 110 °C. In some cases, the temperature could even reach
However, the performance in well cementation of oil well 240 °C. For example, a reservoir in Villafortuna Trecate oilfield has been identified
cement pastes containing high dosage of MK (more than 20%wt) between 5800 and 6100 m depth with a temperature beyond 150 °C [19]. Another
reservoir at depths of 5641–6027 m and temperature beyond 230 °C in Well ND1 in
is rarely reported. Using greater proportion of MK could signifi-
the Jizhong oilfield has been discovered [20]. So different temperature intervals
cantly reduce the environmental impact of cement industry. Mean- were set and the molds were placed into a high temperature chamber. Cement slur-
while, if the properties of high MK dosage replaced cement are ries were cured for different time periods at 75 °C, 150 °C, 240 °C, respectively. The

Table 3
Mix design parameters and their designation.

Mixes Mix proportion (Wt%) Water/solid ratio Fluid loss additive (Wt%) Dispersing agent (Wt%)
OPC MK
M0 100 0 0.5 4 0.00
M30 70 30 0.5 4 1.50
M35 65 35 0.5 4 1.75
M40 60 40 0.5 4 2.00
M45 55 45 0.5 4 2.25
M50 50 50 0.5 4 2.50
M55 45 55 0.5 4 2.75
M60 40 60 0.5 4 3.00
 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48
Fig. 1. Core samples.
curing time of cement mortars were the same at different temperatures. Two cubes
of each kind of hardened cement paste were subjected to compressive strength test
and the average value was recorded. This was performed on wetted specimens with
a 0.5 Mpa/s loading rate by NYL-300 forcing press machine obtained from Wuxi
Building Material Equipment Factory.
2.2.3. Porosity test
The H19915 porosity apparatus was used to test the porosity of cement cores of
2.5 cm diameter and 5 cm length which cored by a coring equipment. The cores are
presented in Fig. 1.
2.2.4. Scanning electron microscope test
Scanning electron microscope S-4800 was used for carrying out the microstruc-
ture studies. The specimen cubes were cured for 56 days and finished with ion
beam milling to obtain the best possible surface for high resolution observation.
2.2.5. Thickening time test
The thickening time of cement slurry which is related to pumpability time
needs to be considered [21]. The consistency was tested at 75 °C, 150 °C, 240 °C,
respectively, considering the high temperature in deep wells and geothermal wells.
Fig. 2. Acidproof chamber.
41
The unit of consistency is Bearden (Bc). Thickening time test is ended when the
slurry achieves a consistency of 100 Bc. The slurry container which is equipped with
a stationary paddle assembly is rotated at a speed of 150r/min.
2.2.6. CO2 and acid exposure
Considering the corrosion condition in some geothermal wells and acid treat-
ment wells, CO2 and acid treatment condition should be simulated. To simulate
the CO2 within the cement samples under static circumstance, an acidproof cham-
ber was used (Fig. 2). The chamber was filled with water and CO2 was injected from
the top of the chamber. Cement stones were placed into the chamber at 75 °C for
8 days. Afterwards, compressive strength and porosity test of samples were
measured.
To simulate the acid treatment condition, samples were placed into a solution
of hydrochloric acid concentration of 12%wt at 75 °C for 8 days. The solution
replaced every two days to keep the acid concentration. And then, samples were
weighted every 2 days after drying for 2 h in a muffle furnace at 75 °C. After that,
compressive strength and porosity test of samples were measured.
2.2.7. Density and rheology test
Density of cement slurries was measured by a densimeter. Rheology was mea-
sured using a ‘‘ZNN-D6” rotary rheometer. All the rheology tests were carried out
after 5 min of mixing of the cement slurries. Maintain the test cement slurry at
75 °C with a heating chamber. During measurement, make every endeavor to pre-
vent the cement slurry from remaining static.
3. Result and discussion
3.1. Compressive strength
The compressive strength of hardened pastes made of OPC and
OPC with different dosage of MK for all test ages (2, 7, 28 and
56 days) are shown in Figs. 3–5. For all ages of pastes, the variation
pattern of compressive strength of cement pastes with MK was
almost identical at the same temperature.
At 75 °C, the M0 sample had higher strength at all ages than
cement pastes with MK replacement, it is clear that the strength
decreased with an increase in the level of MK dosage (Fig. 3). The
strength loss of M6 in 2 days was 61% in comparison to M0. This
may be due to that the reaction of MK was not completed. Poon
[22] indicated that even after reaction for 56 days at this tempera-
ture, the reaction of MK was not completed. After all, the 2-days
compressive strength of paste M60 is still 16.4 MPa which is con-
sidered to meet the basic cementation standard of 13 MPa. How-
ever, in all of the cement samples with MK replacement it was
observed that the compressive strength increased remarkably on
the 28th and 56th days due to the pozzolanic reaction. This is
attributed to the fact that the dissolution and hydration of MK is
slow at 75 °C.
Fig. 3. Compressive strength test results at 75 °C.
42 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48

Fig. 4. Compressive strength test results at 150 °C. Fig. 6. Porosity and compressive strength of paste samples M0 and M40.

2-days strength decreased. The cement paste containing 55% MK

achieved the maximum gain of 102% in 2-days compressive
strength with respect to M0 paste. The strength development of
high dosage MK samples became rapidly due to the fast decompos-
ing of MK at temperatures beyond 200 °C [24]. After 28 days
curing, the compressive strength of sample M60 showed better
performance than other samples. This may be explained by the
further hydration of unreacted MK at 240 °C.

3.2. Porosity

The results of porosity and compressive strength of paste sam-

ples M0 and M40 at 56 days are presented in Fig. 6. The values on
each column of porosity are compressive strength whose unit is
MPa. At 150 °C and 240 °C, the high compressive strength of paste
samples in accordance with low porosity. But at 75 °C, the paste
sample M30 showed lower porosity than the M0 sample, even
though the M0 sample presented better mechanical behavior. This
Fig. 5. Compressive strength test results at 240 °C.
is because the MK particles filled the void spaces inside the
skeleton of hardened structure (Fig. 7).

It is obvious that, at 150 °C, an increase in MK content from 40% to 3.3. Mineralogical analysis of hydration products and microstructure
50% had a positive influence; the 2-days compressive strength of
M30 and M35 is lower than M0, and then increased with the content Mineralogical analysis of hydration products and microstruc-
of MK increasing up to 45% (Fig. 4). The decrease of compressive ture study were carried out to explain the compressive strength
strength of M30 and M35 is attributed to dilution effect [23]. The and porosity test results.
strength enhancement is through two main reasons which are poz-
zolanic reaction with Ca(OH)2 producing additional CSH and the fil-
ler effect that increases cement strength [14]. After that, the strength
decreased again with further increasing of MK content which means
that content of OPC also plays a significant role in compressive
strength development. The excessive MK particles not only filled
the interstitial spaces inside the skeleton of hardened structure
but also appeared around the cement particles resulted in the deten-
tion of cement hydration. The maximum 2-days strength gain of 68%
occurred when cement made with 45% MK. This may be due to the
different hydration products obtained in cement pastes containing
various amount of MK. In addition, the strength of cement pastes
containing MK increased significantly on the 28th and 56th days.
The 56-days strength of all other cement samples is higher than
M0, even though the 2-days strength of M30 and M35 is lower than
M0. Even at 150 °C, the reaction of MK is still sluggish.
At 240 °C, an increase in MK content from 35% to 60% had a
positive influence; the 2-days compressive strength showed an
increase up to 55% (Fig. 5). With further addition of MK, the Fig. 7. SEM image of M30 at 75 °C.
 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48 43

Fig. 8. XRD data of M0 and M30 pastes. Peak labels are A-portlandite (Ca(OH)2), B-
Fig. 10. XRD data of M0, M45 and M60 pastes. Peak labels are A-portlandite
larnite (Ca2SiO4), C-calcium iron aluminium oxide (Ca2Fe1.28Al0.72O5), D-calcium
(Ca(OH)2), B-calcium silicate (Ca2SiO4), C-calcium iron aluminium oxide
silicate (Ca3SiO5), E-calcium iron (Ca2FeAlO5).
(Ca2Fe1.28Al0.72O5), D-tobermorite (Ca5(OH)2Si6O1.64H2O).

3.3.1. XRD and SEM of M0 and M30 at 75 °C

Fig. 8 illustrates the XRD patterns of the M0 and the investi-
gated cement paste M30 which was cured at 75 °C. The main pro-
duct obtained from hydration of M0 paste was portlandite. This is
due to part of b-C2S was converted into Ca(OH)2 and CSH gel by (1).
b  Ca2 SiO4 þ H2 O ! C—S—HðgelÞ þ CaðOHÞ2
The peak characteristics for calcium iron aluminium oxide and
larnite could also be distinguished. Fig. 9a and b shows the exis-
tence of pore and crack structure in hardened M0 paste. These pore
and crack would act as channels to fluid permeating, resulting in
high porosity (Fig. 6). On the contrary, the filling effect (Fig. 9c)
of MK particles which filled in the pores of hardened M30 paste
led to close-grained structure and low porosity. As for the com-
pressive strength of M30 (Fig. 3), the decrease in compressive
ð1Þ strength was due to two primary factors: (a) the extension in lar-
nite phases (Fig. 9d) which is deleterious for cement structure;
(b) poor crystallinity resulted in inferior strength. The XRD
patterns of M30 indicated that the intensities of the peaks

Fig. 9. SEM image of (a) M0, (b) M0, (c) M30, (d) M30 at 75 °C.
44 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48
Fig. 11. SEM image of (a) M0, (b) M45, (c) M60 at 150 °C.
characterizing portlandite with a high degree of crystallinity van-
ished and the intensities of the peaks characterizing C3S and cal-
cium iron with poor crystallinity increased.
3.3.2. XRD and SEM of M0, M45 and M60 at 150 °C
Fig. 10 shows the XRD patterns of the M0 and the investigated
cement pastes M45 and M60 cured at 150 °C. The M0 paste showed
Fig. 12. XRD data of M0 and M55 and M60 pastes. Peak labels are A-portlandite
(Ca(OH)2), B-reinhardbraunsite (Ca5(SiO4)2(OH)2), C-tobermorite
(Ca5(OH)2Si6O1.64H2O), D-calcium silicate (Ca3SiO5), E-andradite
(Ca3Fe1.5Al5SiO42.75(OH)), F-dmisteinbergite (CaAl2Si2O8), G-calcium magnesium
aluminium oxide silicate (Ca54MgAl2Si16O90).
the main peak characteristics to portlandite and C2S. At 150 °C, the
b-C2S hydrated phases almost entirely converted into a-C2S
hydrated phases [25]. The a-C2S hydrated phases are not stable,
resulting in a decrease in strength (Fig. 4). The microstructure of
M0 displays the distribution of calcium iron aluminium oxide
(CIAO). This phase formed a highly symmetrical cube - a hexagonal
crystal with 1–3 lm diameter [26]. But structure voids are formed
between the CIAO phase (Fig. 11a); so it is prone to penetration of
water and other fluid [27]. For the hardened paste of M45, the peak
characteristics for tobermorite could be distinguished as well as
calcium iron aluminium oxide. The outer layer of tobermorite con-
sists of hydroxyl silicate chain of infinite length and three Si3O9
tetrahedrons. The three tetrahedrons can be seen as two Si2O7 in
the same direction and a tetrahedron in a different direction as
bridge, whose core is filled with calcium [28]. This structure is
stable at elevated temperatures. Meanwhile, tobermorite phase is
in a layered structure. Its crystal is stacked layer upon layer and
densely filled (Fig. 11b). These factors led to compact structure
and strength increase. So it is able to inhibit the strength decrease
of cement at high temperatures (Fig. 4). XRD results obtained for
the paste M60 displayed that only C2S was identified and it is hard
to distinguish other phases because of their low degree of crys-
tallinity. This is owing to the excessive MK prevented cement
hydration and resulted in insufficient crystallization. However,
the 56-days compressive strength of M60 is the highest of all the
hardened pastes at 150 °C. Studies indicated that the net structure
is the reason why geopolymer shows excellent mechanical prop-
erty [29–31]. And the similar net structure was observed in hard-
ened M60 paste (Fig. 11c) as well. This may be the reason why
high strength was accompanied by insufficient crystallization of
M60.
3.3.3. XRD and SEM of M0, M55 and M60 at 240 °C
Fig. 12 shows the XRD patterns of the M0 and the investigated
pastes M55 and M60. The main peak of M0 is still portlandite. And
the peak characteristics for reinhardbraunsite and C3S could also
be distinguished. Reinhardbraunsite was formed by the hydration
of metastable Ca8Si5O18 which is unstable at 240 °C [25]. The
SEM image of M0 paste at 240 °C shows a loose crumb structure
(Fig. 13a) because of the water loss of cement at high tempera-
tures. This reason led to cement strength decrease and porosity
increase. The intensities of the peaks characteristic for tobermorite
 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48 45

Fig. 14. The thickening time of cement slurries.

is a kind of metastable calcium silicon aluminium garnet in transi-

tion state [32]. Because its unstable nature, cement strength loss
occurred. As for the SEM image of M60 paste (Fig. 13c), layered
structure was not presented. Instead, plenty of structure pores
formed and resulted in strength loss.

3.4. Feasibility as oil well cement

3.4.1. Thickening time

From Fig. 14 it can be seen that, when the consistency is tested
at 75 °C, the thickening time of cement slurries with MK obviously
became shorter. Compared to the test at 150 °C, the M0, M30 and
M55 pastes were reduced by 30 min, 25 min and 28 min respec-
tively. As expected, the thickening property became worse at
240 °C. Due to the high activity of MK at elevated temperatures,
the consistency of M30 and M55 tested at 240 °C were reduced
to 62 min and 56 min, respectively. It is apparent that MK addition
and the increasing in temperature had a negative influence on
paste thickening property [33]. Therefore, when MK is used in oil
well cementation, suitable retarder should be added to cement
slurries. As can be seen in Fig. 15, when the consistency of M30
paste was tested at 75 °C, 150 °C and 240 °C, respectively, the
thickening time of cement slurries with retarder OAI became
longer. The thickening time and retarder proportion show linear
relationship, which indicate that the thickening time of cement
with a high dosage of MK is adjustable.

3.4.2. CO2 and acid exposure

Portlandite becomes unstable at a pH under 12.6, so when port-
landite is formed during the exposure to acid, the set cement is lea-
ched from the surface (Eq. (2)). This is called ‘‘dissolving attack”.

2Hþ þ CaðOHÞ2 ! Ca2þ þ 2H2 O ð2Þ

As for cement corrosion due to CO2 attack, it can be described
with three steps: (1) formation of H2CO3; (2) carbonation of the
cementitious phases; (3) leaching out and deposition of CaCO3 [6].
Fig. 13. SEM image of (a) M0, (b) M55, (c) M60 at 240 °C. The results of compressive strength, porosity of paste samples
M0 and M30 at 56 days after acid and CO2 exposure are presented
in Fig. 16, respectively. The values on each column of porosity are
compressive strength whose unit is MPa. The sample M30 showed
and andradite could be distinguished for the OPC paste containing
better porosity than the sample M0 after acid and CO2 exposure,
55% MK. Owing to the closely packed structure and high stiffness,
even though the compressive strength of M30 was lower before
these phases inhibited the decrease of cement strength (Fig. 13b).
treating with corrosive fluids. As expected, weight decrease after
For the hardened paste of M60, dmisteinbergite and calcium mag-
acid exposure was both observed in M0 and M30 (Fig. 17). The
nesium aluminium oxide silicate were identified. Dmisteinbergite
weight of M0 at 2, 4, 6 and 8 days decreased by 6.2%, 12.6%,
46 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48
Fig. 15. The thickening time of cement slurry M30 with retarder OAI at (a) 75 °C, (b)
150 °C, (c) 240 °C.
17.5% and 19.9% respectively. The SEM images of M0 after acid
attack showed the structure of cement stone became loose
(Fig. 18a). More pores and cracks were connected because of the
deposition of cement structure (Fig. 18b). As for the weight of
M30, it was decreased by 2.8%, 5.8%, 7.4%, 9.3%, respectively. This
is mainly attribute to that the pozzolanic reaction consumed the
Ca(OH)2. By performed the porosity test, it came out that cement
samples with MK presented a lower porosity after acid exposure.
Fig. 16. Results for compressive strength, porosity of pastes after CO2 or acid attack
for 8 days.
Fig. 17. Acid test results.
As a result, a decrease of area and contact time between the
cement and corrosive fluids occurred, slowing down the corrosion
procedure [34].
3.4.3. Density
Fig. 19 shows the results for density of paste samples made with
different dosage of MK. The trend was that the density of the sam-
ples decreased with an increasing MK content. It is clear that the
MK can be used as light-weight-additives (LWA) in low density
cement slurries, but can also preserve the strength of these low-
density cement slurries.
3.4.4. Rheology
Because the very high surface area of MK particles leads to high
water demand and poor rheology, we used water/solid ratio of 0.5
and FHJZ-1 dispersing agent. Table 4 shows the results of rheology
properties. Where w/s represents the water solid ratio, n is the flow
behavior index and K is the consistency index. As compared to No.1
paste, the No.2 paste was more fluid and displayed a little lower n
and K values. The addition of FHJZ-1 in blended cement (No.3),
made the rheological properties better. The flow behavior index
was high and showed good shear thinning property. The consis-
tency index was low and the flowability parameter was greater
than 20 cm which was benefit to reduce the pump pressure.
 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48 47

Fig. 18. SEM images of M0 after acid attack.

(3) The hydration products of cement pastes with MK were var-

ious at different curing temperatures. The structure and dis-
tribution of the hydration products would influence the
hardened cement paste performance. The plate-like compact
deposit structure led to strength increase while the pores
and cracks would act as channels to fluid permeate and
resulted in strength decrease.
(4) Due to the pozzolanic reaction of MK, the MK replaced
cement pastes showed better performance of corrosion
resistance. The thickening time of cement slurry with MK
became shorter. MK addition and the increasing in tempera-
ture had a negative influence on paste thickening property.
But the thickening time of MK replaced cement slurry is
adjustable with proper retarder. MK can be used as light-
weight-additives (LWA) in low density cement slurries with-
out significant detrimental effect on compressive strength.
The MK replaced cement slurries showed good rheological
Fig. 19. Density test results. property with appropriate dispersing agent.
(5) In general, at temperature of 75 °C, the early-term strength
decreased with the increasing MK content while the set
Table 4 cement with more MK showed higher long-term strength.
Rheology behavior test results. At temperatures of 150 °C and above, the cement pastes
M40 to M60 presented good strength both in early-term
No. MK w/s FHJZ-1 n K/Pa sn Flowability/cm
and long-term curing. The hydration products and structure
1 45% 0.4 / 0.97 0.33 17
would influence the strength and porosity of cement pastes.
2 45% 0.5 / 0.95 0.30 20
3 45% 0.5 2.25% 0.91 0.21 22 The MK replaced cement pastes are suitable for the cemen-
tation of deep wells, acid rich wells and geothermal wells.

4. Conclusions and suggestions Acknowledgement

Based on the test results, the following conclusions can be
drawn.
(1) The variation pattern of compressive strength of cement
pastes with MK was almost identical for all ages of pastes
at each temperature interval. At 75 °C, the strength
decreased with the increase in the level of MK content at
early ages, but the cement pastes containing high dosage
of MK showed better long-term strength. Beyond 150 °C,
the cement pastes containing 40%wt–60%wt of MK pre-
sented good mechanical performance both in early age and
long-term curing.
(2) At 150 °C and 240 °C, high compressive strength of paste
samples is consistent with low porosity. At 75 °C, the paste
sample containing 30%wt of MK showed lower porosity than
the M0 sample, even though the M0 sample presented better
mechanical behavior.
This study was supported by China Postdoctoral Science Foun-
dation (2014M561981), Program for Changjiang Scholars and Inno-
vative Research Team in University (IRT1086) and National Basic
Research Program of China (2015CB251202).
References
[1] M.C.M. Nasvi, P.G. Ranjith, J. Sanjayan, A numerical study of triaxial mechanical
behaviour of geopolymer at different curing temperatures: an application for
geological sequestration wells, J. Nat. Gas Sci. Eng. 26 (2015) 1148–1160.
[2] J. Won, D. Lee, K. Na, I.-M. Lee, H. Choi, Physical properties of G-class cement
for geothermal well cementing in South Korea, Renewable Energy 80 (2015)
123–131.
[3] C. Ma, L. Chen, B. Chen, Analysis of strength development in soft clay stabilized
with cement-based stabilizer, Constr. Build. Mater. 71 (2014) 354–362.
[4] C. Ma, B. Chen, Properties of a foamed concrete with soil as filler, Constr. Build.
Mater. 76 (76) (2015) 61–69.
[5] W. Khaliq, H.A. Khan, High temperature material properties of calcium
aluminate cement concrete, Constr. Build. Mater. 94 (2015) 475–487.
48 Y. Bu et al. / Construction and Building Materials 112 (2016) 39–48
[6] A. Brandl, J. Cutler, A. Seholm, M. Sansil, G. Braun, A. Brandl, et al., Cementing
solutions for corrosive well environments, Spe Drill. Complet. 26 (2) (2011)
208–219.
[7] M.R. Shatat, Hydration behavior and mechanical properties of blended cement
containing various amounts of rice husk ash in presence of metakaolin, Arab. J.
Chem. (2013).
[8] M.A. Caldarone, K.A. Gruber, R.G. Burg, High-reactivity metakaolin: a new
generation mineral admixture, Concr. Int. 16 (1994).
[9] M.H. Zhang, V.M. Malhotra, Characteristics of a thermally activated alumino-
silicate pozzolanic material and its use in concrete, Cem. Concr. Res. 25 (8)
(1995) 1713–1725.
[10] F. Wang, S. Wang, Y. Meng, L. Zhang, Q. Wu, J. Hao, Mechanisms and roles of fly
ash compositions on the adsorption and oxidation of mercury in flue gas from
coal combustion, Fuel 163 (2016) 232–239.
[11] S.V. Vassilev, C.G. Vassileva, V.S. Vassilev, Advantages and disadvantages of
composition and properties of biomass in comparison with coal: an overview,
Fuel 158 (2015) 330–350.
[12] J.L. Provis, P. Duxson, J.S.J.V. Deventer, G.C. Lukey, The role of mathematical
modelling and gel chemistry in advancing geopolymer technology, Chem. Eng.
Res. Des. 83 (7) (2005) 853–860.
[13] M.R. Shatat, Hydration behavior and mechanical properties of blended cement
containing various amounts of rice husk ash in presence of metakaolin, Arab. J.
Chem. (2013).
[14] M.S. Morsy, Y.A. Al-Salloum, H. Abbas, S.H. Alsayed, Behavior of blended
cement mortars containing nano-metakaolin at elevated temperatures, Constr.
Build. Mater. 35 (2012) 900–905.
[15] A. Nadeem, S.A. Memon, T.Y. Lo, The performance of fly ash and metakaolin
concrete at elevated temperatures, Constr. Build. Mater. 62 (2014) 67–76.
[16] A. Nadeem, S.A. Memon, T.Y. Lo, Evaluation of fly ash and metakaolin concrete
at elevated temperatures through stiffness damage test, Constr. Build. Mater.
38 (2013) 1058–1065.
[17] C. Li, H. Sun, L. Li, A review: The comparison between alkali-activated slag (Si
+Ca) and metakaolin (Si+Al) cements, Cem. Concr. Res. 40 (9) (2010) 1341–
1349.
[18] J.J. Brooks, M.A. Megat Johari, M. Mazloom, Effect of admixtures on the setting
times of high-strength concrete, Cement Concr. Compos. 22 (4) (2000) 293–
301.
[19] C. Alimonti, E. Soldo, Study of geothermal power generation from a very deep
oil well with a wellbore heat exchanger, Renewable Energy 86 (2016) 292–
301.
[20] G. Zhu, H. Wang, N. Weng, H. Yang, K. Zhang, F. Liao, et al., Geochemistry,
origin and accumulation of continental condensate in the ultra-deep-buried
Cretaceous sandstone reservoir, Kuqa Depression, Tarim Basin, China, Mar.
Petrol. Geol. 65 (2015) 103–113.
View publication stats
[21] C. Ma, Y. Bu, B. Chen, Preparation and performance of a lignosulfonate–AMPS–
itaconic acid graft copolymer as retarder for modified phosphoaluminate
cement, Constr. Build. Mater. 60 (12) (2014) 25–32.
[22] C.S. Poon, L. Lam, S.C. Kou, Y.L. Wong, R. Wong, Rate of pozzolanic reaction of
metakaolin in high-performance cement pastes, Cem. Concr. Res. 31 (9) (2001)
1301–1306.
[23] E. Badogiannis, G. Kakali, G. Dimopoulou, E. Chaniotakis, S. Tsivilis, Metakaolin
as a main cement constituent. Exploitation of poor Greek kaolins, Cement
Concr. Compos. 27 (2) (2005) 197–203.
[24] C.-S. Poon, S. Azhar, M. Anson, Y.-L. Wong, Performance of metakaolin concrete
at elevated temperatures, Cement Concr. Compos. 25 (1) (2003) 83–89.
[25] K. Yanagisawa, X. Hu, A. Onda, K. Kajiyoshi, Hydration of b-dicalcium silicate at
high temperatures under hydrothermal conditions, Cem. Concr. Res. 36 (5)
(2006) 810–816.
[26] S. Kanehira, S. Kanamori, K. Nagashima, T. Saeki, H. Visbal, T. Fukui, et al.,
Controllable hydrogen release via aluminum powder corrosion in calcium
hydroxide solutions, J. Asian Cer. Soc. 1 (3) (2013) 296–303.
[27] N. Singh, Hydrothermal synthesis of b-dicalcium silicate (b-Ca2SiO4), Prog.
Cryst. Growth Charact. Mater. 52 (1–2) (2006) 77–83.
[28] S. Shaw, C.M.B. Henderson, B.U. Komanschek, Dehydration/recrystallization
mechanisms, energetics, and kinetics of hydrated calcium silicate minerals: an
in situ TGA/DSC and synchrotron radiation SAXS/WAXS study, Chem. Geol. 167
(1–2) (2000) 141–159.
[29] T. Phoo-ngernkham, V. Sata, S. Hanjitsuwan, C. Ridtirud, S. Hatanaka, P.
Chindaprasirt, High calcium fly ash geopolymer mortar containing Portland
cement for use as repair material, Constr. Build. Mater. 98 (2015) 482–488.
[30] P. Nath, P.K. Sarker, Effect of GGBFS on setting, workability and early strength
properties of fly ash geopolymer concrete cured in ambient condition, Constr.
Build. Mater. 66 (2014) 163–171.
[31] M. Khalifeh, A. Saasen, T. Vralstad, H. Hodne, Potential utilization of class C fly
ash-based geopolymer in oil well cementing operations, Cement Concr.
Compos. 53 (2014) 10–17.
[32] A.N. Krot, K. Nagashima, G.J. Wasserburg, G.R. Huss, D. Papanastassiou, A.M.
Davis, et al., Calcium-aluminum-rich inclusions with fractionation and
unknown nuclear effects (FUN CAIs): I. Mineralogy, petrology, and oxygen
isotopic compositions, Geochim. Cosmochim. Acta 145 (2014) 206–247.
[33] A.S. Aguilar, J.P. Melo, F.H. Olivares, Microstructural analysis of aerated cement
pastes with fly ash, metakaolin and sepiolite additions, Constr. Build. Mater. 47
(2013) 282–292.
[34] G. Batis, P. Pantazopoulou, S. Tsivilis, E. Badogiannis, The effect of metakaolin
on the corrosion behavior of cement mortars, Cement Concr. Compos. 27 (1)
(2005) 125–130.