Microstructural changes due to elevated temperature in cement based grouts

Microstructural Changes Due to Elevated Temperature in Cement Based Grouts Daniell e Palardy,* Mari a Onofrei,t and Gerard Ballivy* "Univetsite de Sherbrooke, Dep artement de Genie Civil, She rbroo ke, Quebec, Canada; and t AfCL, Whiteshell Laboratories Pinawa, Pinawa, Manit oba, Canada

The durability of type K and H cement based grouts under conditions potentially found to be in high level nuclear waste repositories was studied. Tests have been carried out to determine the effects of temperature on the hydraulic conductivity and the leaching resistance of the grouts. Mea surements of merCilry intrusion porosimetry and scanning electron microscopy with energy dispersive X-ray analysis have been used to investigate the changes in the pore structures of both grouts as f unction of leaching and permeating time. Type K and type H cement based grouts, made with low water to cemeniitious materials ratio, silica f ume, and superplasiicizer, were exposed to high temperature and pressure. Preliminaru results indicate that both hydraulic properties and leaching behavior of the grouts investigated are affected by the increase in temperature. However, data show that leaching rate and hydraulic conductivity of the grouts decrease with time. The results showed clearly that chemical reactions, presumably accelerated by tile elevated temperatures (100°C), led to the [orma tion of a precipitate ill the microcracks and on the surface of the leached specimens. This precipitate is likely the cause of the observed decrease in the hydraulic conductivity and leaching ra te. ADVAN CED CEMENT BASED MATE RIALS 1998, 8, 132-138. © 1998 Elsevier Science Ltd. KEY WORDS: Cement bas ed gro u t, High temperature, Microstructure modifications, Perm eability, Static leaching, Typ e K cem ent, Type H cement

u rrent repository design s for the disposal of radioactive w as te env isage exten sive use of cement ba sed materials (i.e., grouts, con crete ) [1-4] . The se materials ma y be used for the waste form matrix, as construction materials for wa lls and floor, as a grout to fill in gaps between wa ste pa cka ges, and to seal bor eh oles, shafts, tunnels, and natural fractures or frac tures gene rated during va ult excavation. The ce-

C

Andress correspondence (0: D r. Gerard Balli vy, Civil Engin eerin g Departm ent, Un iversit6 de Sherb rooke. Rocks M echanics L rboratory, Sherbroo ke, Qu ebec J I K 2 Rl , Ca na da. Received December 5, 199 7; Accepted M arch 20, 199 8.

© 1998 by Elsevier Science Ltd. 655 Av enue of the Americas, New York , NY 10010

ment based materials will fulfil different functions in each of these applications. They are used in these app lications due to inherent strength, durability, and versatility. Cement based grouts ha ve been identified as potential grou ting materials. The requi red performance of ceme n t based grouts includes not onl y an ability for the material to penetrate and seal ve ry fine fracture, but also a potential for lon g-term mechanical and chemical sta bility in disp osal va u lt envi ronment. Under man y cond itions, the deteriorati on of cement ba sed materi als is not attribu ted to an y sing le cause, but arises fro m the combined action of a number of potentially destructive age n ts. There ar e man y possible env ironme ntal factors and processes by which the design properties of the ceme nt based grouts may cha nge over long period s of time. These include reac tion with the en vironment through the aqueous phases, int ernal microstructural changes, temperature, action of microorganism, and radiation (i.e., radiolysis). It is comm only held that the hydraulic performance of ceme nt based grouts w ould chan ge as water passed th rou gh the pores in the material s. The cement will di ssolve in a saturated environment, and porosity and hydraulic conductivit y would inc rease w ith tim e. The rates of change would be influenced by spontaneou s cha nges in microstructure of the material. It is obvious that the hydraulic d yn amics in the vicinity of the gro u t are an important factor in estima ting the lon g-term durability of cemen t ba sed gro u ts. It is generally kno wn that th e durability of harden ed cem ent exposed to aggressive aq ueous env iro nme nts is related to permeability and, thereby, to the pore structure of the hardened ceme nt. Durability is an important par ameter; it is defined as the capacity of a cementitiou s mater ial to resist under given environmen tal con di tions [5]. Th is article presents the results of laboratory studies carr ied out to determine the effects of temperature, typ e of cement, and harden ed gro u t porosity on th e hydrauISSN 1065-7355/98/$19 .00 PII S1065-7355 i9i:l)0001 6-9

Cement Based Grouts 133

Advn Cem Bas Mat 1998;8: 132-138

TABLE 1. Composition of grouts studied

Axial load

Mix

Type of Cement

Ratio W/C

Silica Fume (%)

Super plasticizer (%)

1 2

K H

0.5 0.3

10 15

1 1

Pressure

F

(E x ter ,-,"n~_......

uppe r plate

lie conductivity and leaching behavior of cement based grouts.

Experimental Procedures Permeability Tests The hydraulic conductivity tests were carried out in a specially con structed permeameter (Figures 1 and 2) [9,lD]. It is a radial flowing permeameter and allows the study of samples under high temperature and pressure. This permeameter permits the creation of medium depth conditions. Cylindrical samples with a central cylindrical opening parallel to the longitudinal axis

Sam?l e - r od

~-fU---

t-JI-

Materials Type K cement is the most common expansive cement used today [6,7]. In this cement, the aluminate source for the ettringite formation is calcium sulfoaluminate (C4A3S ). An expansive cement normally is used in a situation where shrinkage of the cement based material cannot be tolerated [8]. In the case of cement based grout, the required exp ansion is higher than the shrinkage to create a contact force between the grout and its environment. This balance between expansion and shrinkage reduces cracking. Type H cement is an oil-well cement; specifications were established by the American Petroleum Institute. It is a slow-setting cement that is used to seal deep wells under high temperature and pressure. Type H cement can be used as a basic well cement from the surface to depths of 2500 m . Its typical fineness is about 160 m 2 /kg. This cement is a high sulfate resistance cem ent, as determined by C3A content (max imum 3%). Two different grout mixes, one with type K cement and the other with type H cement, were studied . The mixes are shown in Table 1. Silica fume and a naphthalene based superplasticizer were added to each cement to improve grout properties such as low permeability and long-term leaching resistance. All the grout components (cement, distilled water, silica fume, and superplasticizer) were mi xed for 300 s at a speed of 40 rpm with a laboratory mi xer. After the mixing, the grout specimens were cast in molds and placed in a humid chamber for 24 hours. The samples were demolded and cured under a saturated Ca (OHh water at 20°C for 28 days.

~thermocouple

r-

-

,.......:;~I~~t-NI~-- glue

I 50

msr

pl at e Pres s u r e (Intern)

"'-

FIGURE 1. Radial-typ e perrneametric cell [9,10].

wer e tested. The dimensions of the samples are 50 mm in diameter and 110 mm in length with an axial hole diameter of 13 mm. Before testing, the samples were saturated, for at least 2 days, with deionized, deaerated water. Tests were done by the convergent method, which forces the flow from the outside to the inside. In other w ords, the sample is subjected to a radial compressive stress field. The samples were tested at three different temperatures (200, 500, and lDO°C) in a sequential manner with a 10-MPa pressure differential. The temperature wa s incr eased as soon as the hydraulic conductivit y reached a constant value. In general, the test at each temperature wa s carried out for approximately 30 days.

Static Leaching Tests The full description of the experimental design and method ha s been given elsewhere [11]. The test is used sec urity vane

¢¢ 2 JJj

ret urn

ce r rs

V\l nc

securit Y ~ ~ valve ~ Debit q

pn ewnatic v ane

snocx ab s orber

__ -

Atmospheric pressure pressure of service

Water tan k

Pump

FIGURE 2. Environmental permeametcr [9,10].

134

D. Palardy, et al.

Advn Cern Bas Mat 1998;8:132-138

Analytical Method

TABLE 2. Nominal composition and pH of Whiteshell N-1 synthetic groundwater Ion

Concentration (mg/L)

Ca Na Mg Sr K

1890 1830 55 22 13

Characterization of the microstructure and the chemical properties were done before and after the permeability tests and the static leaching tests to find any changes in the cement grouts attributable to different test conditions. Any changes in porosity, chemical composition of the samples, or liquids deriving from the leaching tests were determined. Optical, chemical, and microstructural methods were used to measure these changes occurring as a result of the durability tests.

Concentration Ion

(mg/L)

Si

4 5790 1000 40 7

cr

S042HC0 3 pH

Microstructure Characterization

to investigate the behavior of cement based grouts under various conditions of temperature, water chemistries, and lengths of time in contact with the leachant. The leaching tests were carried out at 25°, 500 a n d 100°C in Teflon containers of 90-mL volume. The grout specimens were placed on Teflon grids at the bottom of each container to ensure full contact between the sample and the solution. Leachant was added to yield a sample surface area to solution volume ratio of 0.1 cm- 1 . Prismatic-shaped samples were cut to obtain a surface area of approximately 6 ern", The amount of water used in each container was about 60 mL. The leachants used in this study included deionized distilled water and Whiteshell N-l synthetic groundwater; they were not renewed during the test. Table 2 shows the composition and pH of the synthetic groundwater used. Leach times of 7, 14, 28, and 35 days were used. The grout samples were removed from the container at their predetermined leaching times.

Surface analysis of the test specimens at the termination of the hydraulic conductivity and leaching tests were performed by scanning electron microscopy (SEM). Before the analysis began, the grout specimens were impregnated with epoxy based resin and polished with increasingly finer diamond powders (15, 3, and 1 urn). The microscope used was a JEOL JMS 840A connected to a microprobe LINK 10,000. Techniques such as mercury intrusion porosimetry have also been used to investigate the changes in pore structure of both grouts. Before measurements, the grout samples were dry by placing them in acetone for 4 hours and then in an oven at 60°C for 12 hours.

Chemical Characterization Crystalline phases were determined by X-ray diffraction (XRD). Samples were crushed under a controlled

1E-10

'ii

!

25° C

1E-11

100°

50° C

r:

1E-12

~

.s; 1E-13 ~ :::I

. • Type H .-TypeK

~ 1E-14

o u

-,

.2 1E-15

_. -.. .

:::I

! 1E-16

~

-.. -

•

::I: 1E-17

Detection limit

------------- ---o

500

1000

~.

1500

2000

2500

Time (hour) FIGURE 3. Effect of percolating time and temperature on the hydraulic conductivity of the grouts studied.

Ce me nt Based Grouts

Advn Ce rn Bas Mat

135

1998;8:132-13 8

TABLE 3. Vilriation of wilter viscosity with temperature {12]

Temperature

Viscosity

(0C)

(v . 106 m 2/s)

20 60 100

0.995 0.480 0.290

(~ v =

(~v

-50%)

= - 70%)

nitrogen atmosphere to prevent carbonation and analyzed by a Rigaku DMAXB apparatus. Solution analyses were performed at th e termination of each static leaching test. Calcium was analyzed by inductively co uple plasma spectroscopy.

Results and Discussion Permeability Tests H ydraulic conductivity wa s assessed w ith grout specimen s under compression. Figure 3 p resents the effects of p ercolating time and temperature on hydraulic conductivity of both grouts under in vestigation. A me an hydraulic conductivity was calculat ed for each temperatu re. The results show that, under the test conditions, both grouts had a permeability lower than the detection limits « 10- 17 m / s) at 25°C and even at 50°C for the type H cement based g ro u t (Figure 3). Type K cement based gro u t has shown a moderat e flow at SO°C and a

more important h ydraulic conductivity at lOO°e. As a comparison, the conductivity m easured for this grout at 100°C (10- 12 m /s) corresponds to the conductivity at 25°C of sound granite. The important fact to notice is th e considerably low hydraulic conductivity of type H cemen t based grout. As exp ected , permeability increases along w ith the temperature increases. The observ ed increase in hydraulic conductivi ty m ay be attributed to changes in microstructure and th e decrea se in the vi scosity of permeated water. The viscosit y of water is lower (Tabl e 3) and it moves easily through the grout samples. A high permeability w as me asured at the initiation of the tests. Temperature can bring thermal expansion that crea tes fractures and crack s. Examination of the specimen s after the tests revealed th e presence of cra cks p arall el to their longitudinal axis . The presence of these cracks may explain the high permeability at the initiation of the tests. However, thi s inc rease in hydraulic cond uctiv ity is observed for a shor t period of time, after which the h ydraulic con d uctivi ty decreases with time. Thi s w as attributed to the self-sealing properties of the g routs. Microscopic examination of the grouts at the termination of the test s showe d th at chemical rea ction , accelerated by the ele vat ed temperature, led to the formation of a precipitate in the cracks in the gro u t speci mens . Thi s precip itate is the likel y cause of the temperature-induced d ecrease in the hydraulic conduc-

11 655

s

FIGURE 4. SEM/energy d ispersive X-ray analyses of the deposit in a crack from the type K cement based

grout.

136

Advn Cem Bas Mat 1998;8:132-1 38

D. Palardy, et al.

tivity of the grout. The mechani sm of self-sealing of cracks in the cement ba sed materials is n ot well understoo d even though the phenomenon is we ll-establishe d . Several au tho rs ha ve indicated that the mo st likely mechani sms responsibl e for the crack sea ling ar e dep osition of calcium carbo nate and/ or formation of ne w hydration products [13-15]. Under th e experimental conditions used in th is study, the results sugges t that the most likely process responsible for the crack sea ling is continued hydration reaction. The SEM / energy dispersive X-ray an alyses of the fillin g material indicated that it consisted main ly of Ca and Si (Figur e 4). The morphology and th e che mical comp osition of the filling material suggest CSH. Figure 5 show s a general view of a fracture obser ved in the typ e K cement based grout and filled up w ith rea ction product. It is a kn own fact that the por e structure of cem ent pa ste con tro ls to a lar ge extent the perm eab ility of the material. There fore, any facto rs affecting either the total p or osity or the pore size distribution of hardened grou t sh ould affect the hydraulic properties of the mat erial. Med ium siz e cap illary pore s (0.01 to 0.05 IJ-m) and large p ores (0.05 to 10 urn ) control permeab ility [16]. As shown in Figures 6 and 7, each gro u t ha s a low percentage of lar ge capillary pores due to the use of silica fum e ad mix ture . This may explain the very low results obtained for p ermeability. Howev er, the typ e K cem ent based gro ut contains a higher percentage of med ium capillary pores, w hich may expl ain the higher hyd rauli c cond uctivity in comparison wi th th e othe r one . Classification use d in these figures comes from [17]. An increase in tot al porosity is observed in both grouts rel at ive to th e in itial va lu es obtained for the control specimens: 6% for th e type K cemen t based g ro u t and 4% for th e typ e H (Fig ures 6 and 7). H owever, d espite th e fract ures ob ser ved, classes 1

l 30 ,..-- - - - - - - - - - - - - --, 5

:sD

25 20

:a

10

~ 15 + - - - - - - ~

~

(;

Q.

+-

5 +-- - - - - 0 +-- ---,-- ---,Class 1

Class 2

Class 3

Class 4

Total

FIGURE 6. Pore size dis tribution of type K grou t before and

after the permeability tests (see no menclatur e).

and 2 have not increased excessively : a slight increase was ob served in typ e K cemen t p aste. Thi s ma y be att ribu ted to th e formati on of precipitate. Class 3 shows n o change, w he reas for th e fourth one, an increase of 5% is measured (Fig ure 6). The increa se in class 4 m ay indicat e the in crease in th e CSH ge l as the result of con tin ue hydrat ion. Finally, th e type H cement based grout shows a low in crea se fo r cla sses 3 and 4 (Fig u re 7). SEM in back scattered electro n induced mod e examinat ion s revealed the presen ce of unhydrated phases in the harden ed grouts: approximately 10% of the matrix for the typ e K cem ent based grout and 25% for the type H grout. Some particles ha ve an hydrated layer on their margin, w hereas othe rs show only a remnant of the ir init ial struc tur e. The hyd ration of these unhydrated ph ases w he n in contact wi th the p ermeating wa ter may exp lain the delay before the water flowe d th rough the gro ut. In the control sam ples, the initial unhydrated phases represented approxi mately 20%, in the type K cemen t based grout and 40')10 in th e type H gro ut. The increase in classes 2 and 3 in type K grou t may be attr ibuted to ett ringite [Ca6AI2(SO.jh (OH )12 · 25H 20 ] deco m position above 60°C. The se obse rva tions have been confirmed by XRD ana lysis performed on the spec imen after the pe rmeability tests. The results sh ow the presen ce of katoite [Ca 3 Al2(SiO.j)(OH)8]' Th is min-

~

30 ..-- - - - - - - - - - - - - --,

S

D

25 20

:a

10

:s

~ 15 ~

+-- -- - - - -- -- -- - +-

_

iii

5

+-- -- - --

~

0

+-----,-- ---,-

e

Class 1

FIGURE 5. Deposit in a cra ck from the typ e K cement based grout (2,200X).

_

Class 2

Class 3

Class 4

Total

FIGURE 7. Pore size d istributi on of typ e H grout before an d after the permeability tests.

Ce me nt Based Grout s

Advn Ce rn Bas Mat 1998;8 :132 -1 38

~

70

"-

~>o

E

60

0;-

50 I

__ 25 °C

.2 30

40

--.- 50 °C ___ 100 °C

~

"04 ..>0/

s ~ os

u

'0 ~

Tempe rature

30

c 35

u

.,os

25 20 '6 15 ~ 10 0 5 0 Co 0

.,

..

10

...J

-

r--

Control K·D K 25·35

O·

0

8

16

24

32

D Class 1 • Class 2 • Class 3 o Class 4 • Total

~

e 20

.c

137

..I

K·D 100·35

K·W 25·35

--

K·W 100-35

40 FIGURE 10. Porosity distribution in the type K cement based grout after 35 days of leaching (see nomenclature),

Time (day)

FIGURE 8. Effect of leaching time and temperature on the

leach rate of calcium for the type K grout reacted with deionized water. era l and calci um sulfa tes, CaS04x H20, ap pear after the decomposi tion of ettringite.

Static Leaching Tests As for th e permeability test s, the increase in tem perature to lOODC brings more important changes in th e gro u t in compa rison to 25°C or SO°c. In fact, at lOO°e, th e resu lts show higher calcium leaching ra tes, leac hing depth, and p orosity. Fig ures 8 and 9 show th e calcium leaching ra te for th e type K an d type H gro u ts submer ged in deionized water: leaching increases at high er tem peratures. H ow ever , after a certain pe riod of time, the system tends to approach a steady sta te attributed to calcium concentration approachi ng solubility lim its and/ or due to formation of a surface protective layer , SEM examination confirmed the form ation of a precipitate layer on the surface of th e leached specimens. This layer "vas id entified, by XRD, as bein g in major p art composed of calcium carbonate. Th e sam ples leached in

~

N

70

Vl

synt he tic groundwater also show traces of Mg(OHh on the extern al surface. A solu tion with MgS0 4 or MgCl 2 leaches the calcium hyd roxide and brings th e formation of bruci te deposits [18]. TIle calcium leachin g originates mos tly from the portlandite Ca(OHh, of which the decrease can be seen semi-quantitatively by XRD. The d issolution of this mineral increases the porosity of the gro uts and may affect their lon g-term durability . Figures 10 and 11 show the pore size distribution in the type K and type H cement based gro uts. The porosity doe s not change significantly at 2SoC in compa rison with the control sample, bu t at lOO°C, a significant differe nce is observe d . The decomposition of ettri ngi te at high temperature also may contribute to th e increase in porosit y. Th is is more importan t in the type K grout, beca use it contains agglome rations of ettringite crysta ls. This m ay exp lain the in crease of class 2 at lOO°C for this grout. Furthermore, few er differences are observed in the porosity classes between 2Soand SO°C; it is at lOO DC that the cha nges are reall y importan t. At the tw o first temperatures, the h ydra tion is mor e imp ortant, w hich is rep resented by an increase for class 4. Howe ver , at lOO°e, it is class 3 (and 2) th at increases: leaching is more important th an hyd rat ion .

E 60

OJ,

.:..:

Temp era ture

'b 50 ~

___ 25 °C

40

c :J

. 'fi

os

20

os

10

.:1

~ 40

--.- 50 °C ___ 100°C

u'" 30 '-

~*

0 0

8

16

24

c 35 30 .& 25 ~ 20 '6 15 ~ 10

.g

: 32

ell

40

Time (da y)

FIGURE 9. Effect of leaching time and temperature on the leach rate of calcium for the type H grout reacted with deionized water.

s0

Q.

5 0

-i ••

Control H·D H 25-35

•• • ....

D Class 1 • Class 2 • Class 3 DClass 4 . Total

l-

H-W H-D 100-35 25·35

H-W 100·35

FIGURE 11. Porosity distribution in the type H cement based grout after 35 days of leaching.

Advn Cern Bas Mat 1998;8 :132-1 38

138 D. Palard y, et al.

Finally, SEM observations show that the leaching depths increase with temperature, especially at 100°C, and duration of test. This results in higher porosity.

2. 3.

Conclusions The study shows that both, hydraulic conductivity and leaching properties of the grouts investigated, were affected by the increase in temperature. However , the results show that both leaching and hydraulic cond uctivit y of the grouts tend to decrease with time. The observed changes were attributed to the formation of new hydration products resulting from the increase in the degree of hydration and associated reactions. The work discussed here revealed that the porosity of the grout does change with leaching time, but in the limits that depend on grout composition, temperature and initial porosity. The results confirm that the grouts investigated have the potential to self-seal.

Nomenclature Class 1 (%) Class 2 (%) Class 3 (%) Class 4 (%) K and H D and W 25 and 100

35

300-0.9 urn 0.9-0.06 j.Lm 0.06-0.009 urn 0.009-0.003 urn Cement type Water type Temperature (0C) Test duration (d ays)

Acknowledgments

4. 5.

6. 7.

8. 9. 10.

11.

12.

13.

14. 15.

The authors w ish to thank Atom ic Energy of Canada Limited for providing finan cial support du ring this proje ct.

References 1. Atkinson , A.; Hearne, J.A. An assessment of long-term

durability of concrete in radioactive waste repositories.

16. 17. 18.

AERE-R 11467, UK Atomic Energy Authority Harwell, O xfordshire, 1984. Push, R. Waste Disposal in Rock; Elsevier Science: Amsterdam, The Netherlands, 1994. Co ons, W.; Bergstorm, A.; Gnirk, P.; Gray, M .; Knecht, B.; Pu sh, R.; Steadman, R.; Stillborg, J.J.; Tokonami, M.; Vaajasa, M. State-of-the-art report on potential usef ill materials for sealing Iluclear waste repository. Stripa Project Tech n ical Rep ort 87-12, Swe dis h Nuclear Fuel and Was te Management Company, Stockholm, 1987. Roy, D.M.; Lan gt on , C A. Longevity of borehole and shaft sealing materials. ONWI-202, 1982. Mehta, P.K; Monteiro, P.J.M. Concrete, Structure Properties and Materials; Prentice-Hall: Englewood Cliffs , NJ, 1993. Folliard, KJ.; Ohta, M.; Rathje, E.; Collins, P. Cement and COIlcrete Research 1994, 24, 424-432. Kurdowski, W. In Zieme Congrcs International de la Chimie des Ciments, Paris, France. 1980; pp V.2/1-V.2 /1 1. Derucher, K.N.; Hein s, C P. Materials for Civil and Highway Engineers; Prentice-Hall: Englewood Cliffs , N], 1981. Co lin, J.C MsA thesis. Universite d e Sherbrooke , Qu ebec, 1990. Balliv y, G.; Colin, J.C ; Mnif, T. In Grouting, Soil ImproveIIlt'llt and Ceosvnthetics, GT Di v / ASCE Proceedings, New Orl eans, LA, 1992; pp 588-600. On ofrei, M.; Gray, M.N.; Breton, D.; Ballivy, G. In Scientific Basic for Nuclear Waste Management XIV, Mater. Res. Soc. Pro c., Pittsburgh, PA, 1991; pp 417-425. Welt y, J.R.; Wicks, C. E.; Wilson, R.E. Fundalllentais of Momentum, Heat and Mass Transfer, Th ird Editi on ; John Wile y & Sons: New Yor k, 1984; p 803. On ofre i, M.; Gra y, M .N.; Keil, L.D.; Push, R. In Pore Structure and Permeability of Cenieniitious Materials, Mater. Res . Soc. Proc ., Pittsburgh , PAr 1989; pp 349-358. On ofr ei, M.; Gr ay, M.N .; Coons, W.E.; Alcorn, S.R. Waste Mmlagement1992, 12, 133-154. Brodersen, K; Nil sson , K. Cement and Concrete Research 1992, 22, 405-417. Mindess, S.; Young, J.F. Concrete; Prentice-Hall: Englewood Cliffs, NJ, 1981. Revertegat, E.; Richet, C; Cegout, P. Cement and COllcrete Research 1992, 22, 259-272. Frigione, G.; Ser sale, R. Cement and COllcrete Research 1989, 19, 885-893.