Ind. Eng. Chem. Res. 2006, 45, 7781-7788
7781
MATERIALS AND INTERFACES Thermal Conductivity of Metakaolin Geopolymers Used as a First Approximation for Determining Gel Interconnectivity Peter Duxson, Grant C. Lukey, and Jannie S. J. van Deventer* Department of Chemical and Biomolecular Engineering, The UniVersity of Melbourne, Victoria 3010, Australia
The thermal conductivities of a systematic series of metakaolin-derived Na, NaK, and K geopolymers have been measured under a range of different environmental conditions, including varied relative humidity (RH) and temperature. The thermal conductivity of geopolymers is closely linked with the specific heat, with little variation in thermal diffusivity observed in different conditions. The thermal transport properties of specimens was found to not change significantly under ambient humidity from 40 to 100 °C. The Hashin-Shtrikman model for conductivity in biphasic solids, with an extension developed by Schilling and Parstzsch, has been applied to determine the thermal conductivity of the intrinsic geopolymer binder and a first approximation of the interconnectivity of the gel. Introduction Geopolymeric binders1 can be described as a biphasic gel of water and aluminosilicate binder. The microstructure of geopolymeric materials is known to vary considerably with composition and processing conditions.2,3 While the effects of microstructure and composition on mechanical properties have been investigated,2,4 there has been little investigation into quantifying the microstructural connectivity of geopolymeric gels. While it is reasonable to expect that compositional factors will affect thermal properties of geopolymers, it should also be expected that the large range of microstructures observed for geopolymeric gels will also have a significant effect on the thermal properties. As such, the effects of both composition and microstructure are investigated in the current work. Geopolymer compositions in the current work can be nominally described by
MAlO2(SiO2)z‚wH2O
(1)
where M represents the alkali metal cation, z is the nominal Si/Al ratio, and w is the molar ratio of water and alkali metal. Alkali metal cations are required to balance the negative charge on AlO4- tetrahedral groups. Therefore, the amount of alkali metal and aluminum in each specimen is controlled to be equimolar. Water is required as a reaction intermediate, and is released during condensation to form pores and create the biphasic structure. The distribution of water in the biphasic gel composite (as porosity) is determined by solution chemistry during formation, which is primarily a function of the Si/Al ratio2,5 and alkali metal cation type, while the absolute pore volume is determined by the nominal water content. Therefore, both aspects of absolute water content and distribution of pores are expected to be critical in determining thermal properties. As the nominal water content affects many of the solution properties, the stoichiometric amount of water in eq 1 is held constant with respect to alkali metal, rather than constant with * To whom correspondence should be addressed. Fax: +61-3-83444153. E-mail:
[email protected].
respect to the weight fraction of solid as is common in ordinary portland cement (OPC) systems. Changes in the distribution of water in geopolymers are also observed to affect greatly the microstructure. Changes in microstructure are reflected in dramatic improvement in the mechanical properties of specimens when the water is distributed in small pores.2,5 The size of pores formed in these geopolymers has been observed to be so small as to be effectively part of the skeletal framework, which reduces the effective density of the gel and reduces the accessible pore volume. Though the effect of alkali metal on the mechanical properties of geopolymers is small, the molecular structure is directly affected by the presence of different alkali metal cations.6 Therefore, the distribution and interconnectivity of the pore structure, the shortrange ordering of the gel phase, and the nominal composition are all likely to play roles in determining the thermophysical transport properties of geopolymeric gels. Given that a large volume fraction of geopolymer composite structures are comprised of water, the thermophysical properties of geopolymeric materials are expected to vary with humidity at ambient conditions and during first heating. The current work will focus on the effect of humidity and specimen composition on the thermal conductivity, thermal diffusivity, and specific heat of a systematic series of geopolymers with nominal composition NayK1-yAlO(SiO)z‚5.5H2O, where 0 e y e 1 and 1.15 e z e 2.15. Thermophysical properties were measured simultaneously with the transient plane technique from 40 to 100 °C and different humidity environments. The Hashin-Shtrikman7 (HS) model for conductivity in biphasic solids is applied to determine the intrinsic conductivity of the gel, and to provide a first approximation of the interconnectivity of the gel phase from measured data. Theory The thermal conductivity of a homogeneous biphasic composite may be determined using the HS model.7 The thermal conductivity of a homogeneous matrix, in this case the geopolymer gel phase, containing a volume fraction of spherical
10.1021/ie060187o CCC: $33.50 © 2006 American Chemical Society Published on Web 10/13/2006
7782
Ind. Eng. Chem. Res., Vol. 45, No. 23, 2006
inclusions, assumed to be the pores, is described by the HS lower and upper bounds, after Weiderfeller et al.:8
λCompositeHS+ ) λG +
λCompositeHS- ) λW +
xW
; 1/(λW - λG) + (1 - xW)/3λG λG > λW (2) 1 - xW
; 1/(λG - λW) + (1 - xW)/3λW λG < λW (3)
where λComposite, λG, and λW are the thermal conductivities of the bulk composite specimen, the gel phase, and pore fluid, respectively. xW is the pore fraction, which is determined from the nominal water content. xW is calculated by assuming that all water in specimens is present as porosity. Like many properties of geopolymeric materials, the observed values are often of the material on a macro scale. Similarly, measurement of thermal conductivity involves the measurement of the composite biphasic structure. Therefore, it is possible to calculate the thermal conductivity of the gel phase in as-cured geopolymer, λG, as λW is known, xW is determined by the nominal water content, and λMeasured may be readily measured with appropriate instrumentation. The calculated thermal conductivity of the gel phase may then be utilized in further analysis of the thermophysical transport properties of the composite. However, eqs 2 and 3 are dependent only on the xW and do not account for the variation in microstructure that has been previously observed in geopolymers.2 The interconnectivity of the conductive phase (which describes the morphology of the microstructure) in the HS model has been described elsewhere according to Schilling and Partzsch:9
Xi )
λMeasured - λCompositeHSλCompositeHS+λCompositeHS-
(4)
where Xi theoretically varies between 0 and 1, for completely discontinuous distribution and infinite interconnectivity, respectively. However, the HS model is generally applied where the two phases in the composite material are well-known, and the HS upper and lower bounds have more physical significance. Unfortunately, due to the aqueous and heterogeneous polycondensation of geopolymeric gel during synthesis, it is impossible to synthesize fully dense geopolymer gel. Therefore, this adds significant uncertainty to the values of gel interconnectivity that can be obtained, which is discussed in more detail in the text of the article. Experimental Procedure Materials. Metakaolin was purchased from Imerys (U.K.) under the brand name of Metastar 402. The molar composition of metakaolin determined by X-ray fluorescence (XRF) was (2.3:1)SiO2‚Al2O3. More detailed analysis of this material is described elsewhere.2,3 The Brunauer-Emmett-Teller (BET) surface area10 of the metakaolin, as determined by nitrogen adsorption on a Micromeritics ASAP2000 instrument, is 12.7 m2/g, and the mean particle size (d50) is 1.58 µm. Alkaline silicate solutions based on three differing ratios of alkali metal Na/(Na + K) ) M (0.0, 0.5, and 1.0) with composition SiO2/M2O ) R (0.0, 0.5, 1.0, 1.5, and 2.0) and H2O/M2O ) 11 were prepared by dissolving amorphous silica in appropriate alkaline solutions until clear. Solutions were stored for a minimum of 24 h prior to use to allow equilibration.
Geopolymer Synthesis. Geopolymer samples were prepared by mechanically mixing stoichiometric amounts of metakaolin and alkaline silicate solution to allow Al2O3/M2O ) 1 to form a homogeneous slurry. After 15 min of mechanical mixing each slurry was vibrated for a further 15 min to remove entrained air before being transferred to two cylindrical Teflon molds (diameter of 50 mm and height of 30 mm) and sealed from the atmosphere. The specimens were cured in a laboratory oven at 40 °C and ambient pressure for 24 h before storage at ambient temperatures in sealed vessels for the prescribed period of time before being transferred into a controlled humidity chamber for measurement of thermal conductivity. The surfaces of the specimens were polished to ensure good thermal contact. For ease of discussion, specimens synthesized with different alkali metal cation types will be referred to as NaR, NaKR, and KR, where R is the Si/Al ratio. Thermophysical Property Measurement. Thermal conductivity, thermal diffusivity, and specific heat were measured with a Hot-Disk thermal constants analyzer (Hot-Disk AB Uppsala, Sweden). The Hot-Disk utilizes the transient place source (TPS) method for simultaneous transient measurement of thermal constants.11 Detailed description of TPS theory is provided elsewhere.11 The sensor was placed between the polished surfaces of each pair of specimens. Measurements at varied humidity were made at 40 °C, with specimens allowed to equilibrate inside the controlled humidity chamber. Due to the size of specimens in the current work, equilibration times were on the order of weeks. Measurements taken at higher temperatures were made by placing specimens in a standard laboratory oven at the prescribed temperature (40-100 °C). The probe diameter and heat power were respectively equal to 3.556 mm and 0.05-0.1 W, while the heating time was 80 s. The values of the thermal conductivity, thermal diffusivity, and specific heat were determined by fitting the thermogram with a unidimensional heat diffusion model.11 Each measurement quoted in the current work is the mean of 20 individual measurements, with variation of (0.003 W m-1 K-1 for thermal conductivity, (0.004 W mm-2 for thermal diffusivity, and (0.035 MJ m-3 K-1 for specific heat. Due to the small error associated with these measurements, error bars for thermal conductivity are accounted for in the size of data points in the figures. However, the mathematical derivation of specific heat values from the thermophysical transport values induces a significant increase in uncertainty about thermal diffusivity and specific heat measurements, which is reflected in the greater degree of scatter and uncertainty of the thermal diffusivity and specific heat measurements. The uncertainty associated with these measurements is NaK > K, which correlates with the trend observed in Figure 1a. However, the magnitude of difference in the thermal conductivities of the specimens based on alkali metal cation type is significantly larger for λG, on the order of 0.1 W m-1 K-1, as opposed to 0.02 W m-1 K-1 for λComposite. Unlike the similar values of λComposite observed in Figure 1a, λG can be generally observed to decrease slightly with increasing Si/Al ratio for all alkali metal compositions. This and the larger magnitude of difference between the different alkali metal series indicate that λG is affected by chemical
7784
Ind. Eng. Chem. Res., Vol. 45, No. 23, 2006
Figure 2. Calculated thermal conductivity of ([) Na, (2) NaK, and (9) K geopolymer gel phase, λG, with Si/Al ratio between 1.15 and 2.15.
composition (contrary to Figure 1a), but that this trend is obscured in Figure 1a by the variation in nominal water content of geopolymers with Si/Al ratio. Also, changes in short-range ordering6,17 and microstructure2 with alkali metal cation and Si/ Al ratio make it difficult to interpret the trend observed in Figure 2. This is especially so since λG is calculated assuming that the binder thermal conductivity is represented by the HS upper bound. In physical terms, this relates to the microstructure containing a perfectly homogeneous interconnected porosity. The validity of the values of λG in Figure 2 can be further elucidated by measurement of λComposite under different conditions, where the pore conductivity should vary without significant change of λG. The thermal conductivities and specific heats of geopolymers measured at 50% RH at 40 °C and after reequilibration at 100% RH are shown in Figure 3. There is a clear reduction in the thermal conductivity values for all specimens (Figure 3) compared to the as-cured state (Figure 1). The thermal conductivity of specimens generally increases monotonically. A 50% increase is observed between the Na1.15 and Na2.15 specimens. This trend is contrary to expectations, based on the reduction in water content with increasing Si/Al ratio, and must be explained by microstructural aspects. The thermal conductivity of Na geopolymer is greater than NaK and K geopolymer specimens for each Si/Al ratio measured at 50% RH (Figure 3). At 50% RH the difference between the observed thermal conductivities of specimens with different alkali metal ratios increases with increasing Si/Al ratio. This may be related to the different microstructures in these specimens and characteristics associated with the different alkali metal cation types. Despite the general trend of increasing thermal conductivity with increasing Si/Al ratio, the thermal conductivities of K and NaK specimens with Si/Al ratio of 1.40 measured at 50% RH exhibit
reduced values compared to specimens with Si/Al ratio of 1.15, while the Na specimens with Si/Al ratios of 1.15 and 1.40 exhibit similar values. This may be a result of the competing effects of reduction in the short-range ordering18,19 and the aluminum/alkali metal content of these specimens with increasing Si/Al ratio as discussed above. The thermal conductivity of specimens recorded after reequilibration at 100% RH are essentially constant with respect to changes in Si/Al ratio, though a slight decrease in conductivity with increase in Si/Al ratio can be observed in all alkali metal cation series. The trend is similar to that observed of the ascured specimens (Figure 1a). However, the values of thermal conductivity in Figure 3 are lower than those in Figure 1a. This suggests that the specimens are subject to environmental hysteresis. Physical and structural changes may occur in the gel as a result of dehydration, similar to those that are observable upon exposure to high temperatures,18,20,21 though no significant indication of this, such as shrinkage, is observed. The specific heats of geopolymers measured at 50% RH, presented in Figure 3b, increase with increasing Si/Al ratio. The difference between the values of the specific heat of Na, NaK, and K geopolymers increases significantly with increasing Si/Al ratio, implying that the alkali metal cation type plays a significant role in determining the amount of water retained in specimens at reduced humidity. The specific heats of the specimens reequilibrated at 100% RH exhibit a decrease in values with increasing Si/Al ratio, consistent with the small reduction in conductivity observed in Figure 3a. The densities of geopolymers as-cured (100% RH), at 50% RH, and after reequilibration at 100% RH are presented in Figure 4. A theoretical dry density of specimens, based on the as-cured densities with nominal water content subtracted, is also presented in Figure 4. The dry density provides a comparison of the extent of dehydration resulting from humidity. The ascured density of specimens increases slightly with increasing Si/Al ratio for each of the Na, NaK, and K geopolymers, which has been observed elsewhere.2 Furthermore, the density at each Si/Al ratio increases marginally in the order Na < NaK < K, as would be expected based on the increase in mass of potassium compared to sodium. The density of all specimens is observed to be significantly lower when measured at 50% RH, implying the loss of significant water from pores as predicted from specific heat data in Figure 3b. Evaporation of water from geopolymeric materials in unsaturated conditions is expected, given that water plays only an intermediary role in the synthesis and does not contribute to the physical structure of the material as intracrystalline water
Figure 3. (a) Thermal conductivities and (b) specific heats of ([) Na, (2) NaK, and (9) K geopolymers with Si/Al ratio between 1.15 and 2.15 measured as cured in 50% humid environment at 40 °C. Hollow data points represent measurements taken after reequilibration of the specimens in 100% RH environment at 40 °C.
Ind. Eng. Chem. Res., Vol. 45, No. 23, 2006 7785
Figure 4. Bulk density of (a) Na, (b) NaK, and (c) K geopolymer specimens measured in different environments.
Figure 5. Water content (weight of water/weight of solid) of ([) Na, (2) NaK, and (9) K geopolymers determined from mass of specimens: (s) as-cured, (- -) 50% RH, and (---) 100% RH.
(though water in pores is known to aid in compressive strength). Therefore, the decrease in the specific heats of all geopolymers in Figure 3b can be linked with the reduction in the density of the specimens by loss of freely evaporable water at 50% RH (Figure 4). The densities of the specimens reequilibrated at 100% RH are generally lower than the as-cured values, with more reduction in density observed at higher Si/Al ratios. Despite this, the NaK and K specimens with low Si/Al ratios exhibit largely identical densities after rehydration compared to the ascured specimens. As such, specimens with low Si/Al ratios containing potassium are less subject to environmental hysteresis than those of high Si/Al ratio or those containing only sodium. The change behavior of these specimens may be related to either microstructural aspects or changes in the gel structure as a result of dehydration. However, no significant change in gel structure has been observed with exposure to 50% RH, making it more likely that microstructure plays a large role. Figure 5 shows the weight fraction of water present in geopolymer specimens at various conditions. For as-cured specimens, the water content decreases with an increase in Si/ Al ratio; i.e., the solid component of the specimen increases for the same stoichiometric content of water. Despite this, the weight fractions of NaK and K geopolymers measured at 50% RH are remarkably similar at all Si/Al ratios. Also, the weight fractions of water in Na specimens are nominally 0.03-0.05 unit greater than for NaK and K geopolymers, which exhibit similar values. This implies that the water content is closely related to the mass of binder to adsorb to, rather than microstructure or composition in terms of Si/Al ratio. Therefore, the water content of specimens at reduced humidity may be driven by the amount of water associated at the surface of the gel. The greater fraction of water present in Na geopolymers
compared to NaK and K geopolymers at reduced humidity may be a result of the higher charge density of sodium, increasing its affinity for water.22,23 The similarity in the weight fractions of the NaK and K systems implies that these specimens exhibit similar affinity for water at the surface of the gel. After reequilibration at 100% RH, the water content of geopolymers increases to values close to the as-cured values. However, the water contents of specimens with low Si/Al ratios are closer to the as-cured values, because the specimens with high Si/Al ratios are unable to rehydrate to the same extent as the low Si/Al ratio specimens. Figure 6 shows the microstructures of geopolymers with different alkali metal cation compositions with Si/Al ratios of 1.15 and 2.15. At low Si/Al ratio, the microstructures of Na, NaK, and K geopolymers exhibit markedly different microstructures. NaK and K specimens exhibit larger pores compared to the finer microstructure of the Na specimen. At higher Si/Al ratios, all geopolymers exhibit a largely homogeneous microstructure. The different microstructures of geopolymers with low Si/Al ratios may be related to how well these specimens rehydrate at 100% RH. The similar water content of NaK1.15 and K1.15 geopolymers as-cured and after rehydration signifies that they rehydrate more completely than the Na1.15 specimen (Figures 4 and 5). Therefore, the microstructure of geopolymer gels does affect the extent of rehydration. A lesser degree of rehydration is observed in homogeneous gels (i.e., Na1.15). This correlates with the low water content in geopolymers after reequilibration at 100% RH at higher Si/Al ratio. The physical interpretation of this observation may be presented in terms of the presence of small pores in these gels not refilling upon rehydration. A higher than ambient pressure is required for water to intrude pores smaller than 10 nm.24 Also, zeolites become more hydrophobic as the Si/Al ratio increases.25 Therefore, the affinity of the gel to water will decrease with increasing Si/Al ratio, and the ability of water to intrude into small pores will be reduced. As the microstructure of geopolymers affects the thermal transport properties, it is important to quantify the interconnectivity of the gel phase. For this to be achieved using the HS method, the conductivity of the specimens must be determined with the pores filled with another fluid. This is achieved here by replacing the water with air at elevated temperatures, where the water content of the specimens is greatly reduced. Figure 7 displays the thermal conductivity of specimens in the current work measured between 40 and 100 °C at ambient humidity. The thermal conductivity of geopolymers increases uniformly with increasing Si/Al ratio in all series, and increases in the order K < NaK < Na. The thermal conductivity of individual geopolymers does not vary significantly up to a temperature of 100 °C from the value recorded at 40 °C.
7786
Ind. Eng. Chem. Res., Vol. 45, No. 23, 2006
Figure 6. SEM micrographs of (a) Na1.15, (b) Na2.15, (c) NaK1.15, (d) NaK2.15, (e) K1.15 and (f) K2.15 specimens.
Figure 7. Thermal conductivity of (a) Na, (b) NaK, and (c) K geopolymers with Si/Al ratios of (b) 1.15, (2) 1.40, (0) 1.65, ([) 1.90, and (9) 2.15, measured between 40 and 100 °C at ambient humidity (nominally 45%).
Figure 8 shows the proportion of water remaining in Na specimens after exposure to elevated temperatures and ambient humidity. Approximately only 20% of all water present in the sample is lost between 40 and 100 °C. The loss of a similar fraction of water is observed between the as-cured state and after reequilibration at 100% RH (Figures 1 and 3) results in reduction of thermal conductivity on the order of 0.1 W m-1 K-1. It is clear that the water lost in heating between 40 and 100 °C has a much reduced effect on thermal properties compared to a similar amount of water lost during changes in RH. Therefore, it is proposed that the water lost from variation
in humidity is from bulk pores, and has a substantial effect on thermal conductivity by greatly changing the conductivity of the bulk of the pores from that of water (0.6294 W m-1 K-1) to that closer to air (0.025 W m-1 K-1). The water lost during heating would be liberated from water absorbed on the gel and hydrating alkali metal cations. The effect of adsorbed and hydrating water molecules on the conductivity of the pores as a whole is small, due to the comparatively low thermal conductivity of air that fills the pore. This explains the insignificant influence of heating from 40 to 100 °C on the thermophysical properties. Any changes in conductivity in this
Ind. Eng. Chem. Res., Vol. 45, No. 23, 2006 7787
Figure 8. Fraction of water present in Na geopolymers after exposure to different environments.
temperature region may be due to a small level of structural and physical evolution of the specimens in this region that is unlikely to greatly affect the gel properties. The increase in thermal conductivity with increasing Si/Al ratio in Figure 7 indicates that the effect of total pore volume is very important in determining the thermal conductivity of geopolymer. However, a reduction of conductivity on the order of 50% resulting from a 8-10% reduction in porosity is too large to be a result of only changes in traditional series and parallel conduction,26 implying that other factors such as the pore size distribution and gel interconnectivity contribute significantly to determining the thermal conductivity. The microstructures of geopolymers with low Si/Al ratios contain larger pores (on the order of microns), while those of higher Si/Al ratio (i.e., Si/Al g 1.65) are highly homogeneous and contain nanometer-sized pores in Figure 6, which has been confirmed by nitrogen porosimetry.2 The change in the pore sizes by 3 orders of magnitude is not a result of the water content, but rather is a result of differences in the gel chemistry during synthesis.2 Conductivity is dominated by the smallest junctions of the path of high conductivity (the geopolymer phase in this case). Since all specimens contain similar water contents at 100 °C, the small pores in specimens with high Si/Al ratios result in a more interconnected structure, and may lead to a more efficient transfer of energy through the specimens. The microstructures of geopolymers are not significantly altered by heating to 100 °C.18,27 Hence the interconnectivity of the gel remains similar in all measurements over this temperature range. Therefore, the Schilling and Partzsch9 interconnectivity parameter (eq 4) can be evaluated by solving the upper and lower HS bounds for the thermal conductivity measurements of specimens as-cured and at 100 °C. These are only a function of the conductivity of the gel and the pores. Since the specimens measured at 100 °C have lower water content, the pores in these specimens can be assumed to contain only air, the thermal conductivity of which is known to be 0.025 W m-1 K-1. Figure 9 shows the values of Xi determined from the above description. The values of λG determined from the procedure described above are within (0.5% of those presented in Figure 2, and thus are not presented for brevity. Xi increases with increasing Si/Al ratio, which correlates with the increase in microstructural homogeneity observed in Figure 6. Furthermore, Na specimens can be observed to exhibit greater interconnectivity than NaK and K specimens, respectively, which also correlates with microstructural observations. Indeed, the interconnectivity of the NaK specimens at low Si/Al ratios (i.e.,
Figure 9. Gel interconnectivity, Xi, of ([) Na, (2) NaK, and (9) K geopolymers.
Si/Al e 1.40) is more closely related to K geopolymers than to Na geopolymers, which correlates with not only the micrographs in Figure 6, but also with the observation in the literature that low Si/Al ratio NaK specimens behave similarly to K specimens.27 Furthermore, high Si/Al ratio NaK specimens tend to have interconnectivity more closely likened to Na geopolymers, which also correlates with the literature.27 The consistency of the values of Xi with microstructural observations provides powerful corroboration of the validity of the Schilling and Parstzsch interconnectivity extenstion of the HS model. However, from the definition of Xi values in excess of 1 are invalid. Therefore, the generally high values of Xi in Figure 9 are conspicuous, and indicate that assumptions made in the evaluation of the parameter are subject to error, namely the calculation of total pore volume and the assumed stability of gel conductivity. The HS model was developed for spherical porosity,7 which is clearly not fulfilled for the description of geopolymeric composites (Figure 6). However, the isotropic approach of the HS model gives a first approximation for describing the interconnectivity of geopolymeric materials, which appear to be valid in comparison to microscopic evidence. The high values of Xi may reflect not only the nonideal geometry of geopolymeric materials, but also that the assumption of geopolymeric binder being unchanged by heating to 100 °C requires further analysis. In the current work, the pore volume of geopolymers is determined from the nominal water content. However, it is known that the pore volume and gel volume (i.e., skeletal density) of geopolymers are largely determined by the solution chemistry during synthesis.2 Therefore, the total pore volumes in Table 1 are also likely subject to significant uncertainty, with porosity subject to variation due to changes in skeletal density.2 Furthermore, the pore volume and gel density are known to vary to some degree in heating to 100 °C.20 Revision of total pore volume (to vary with temperature and skeletal density) and λG (to include possible densification effects) would have the effect of reducing Xi to more physically sensible values. However, inclusion of these small but significant structural changes in the specimens is not justified based on the accuracy with which these factors can be determined and the assumptions of the HS model when applied to the geometry of geopolymers. Nonetheless, this work presents the first systematic evaluation of the effects of nominal Si/Al ratio and alkali metal cation on the thermophysical properties of metakaolin geopolymers. Furthermore, this work is the first application of a mathematical model to provide a good first approximation at quantifying geopolymer gel interconnectivity.
7788
Ind. Eng. Chem. Res., Vol. 45, No. 23, 2006
Conclusions The thermal conductivity, diffusivity, and specific heat of geopolymers derived from metakaolin are largely insensitive to Si/Al ratio in the as-cured state. Small differences in thermal conductivity in the order Na > NaK > K result from the lower specific heat of potassium compared to sodium. A theoretical measure of the conductivity of the binder has been determined by utilizing the Hashin-Shtrikman model.7 From this, the thermal conductivity of the binder is observed to decrease with increasing Si/Al ratio, and is affected by both alkali metal cation density (associated with aluminum) and a reduction in structural ordering with increase in Si/Al ratio. Once exposed to low RH environments the thermal conductivities of geopolymers are largely governed by large reductions in the specific heat resulting from dehydration of water from pores. When measured at 50% RH, the thermal conductivity of Na geopolymers increases by over 50% with the increase of Si/Al ratio from 1.15 to 2.15, with similar trends observed for NaK and K systems but to a lesser extent. When rehydrated (100% RH), the thermal conductivity of specimens does not recover to the as-cured values. The interconnectivity of the gel phase was determined by applying the Schilling and Partzsch interconnectivity extension of the HS model. The interconnectivity of the gel increases with Si/Al ratio, and with increase in sodium content. Although the predictions of the gel interconnectivity are subject to significant error and uncertainty, the trends observed match closely microscopic evidence provided in SEM images. The HashinShtrikman model for biphasic thermal conductivity has been shown as an effective method for evaluating a first approximation of the interconnectivity of geopolymeric gels. Acknowledgment The authors gratefully acknowledge the financial support of the Particulate Fluids Processing Centre (PFPC), a Special Research Centre of the Australian Research Council (ARC). Literature Cited (1) Davidovits, J. GeopolymerssInorganic polymeric new materials. J. Therm. Anal. 1991, 37, 1633. (2) Duxson, P.; Provis, J. L.; Lukey, G. C.; Mallicoat, S. W.; Kriven, W. M.; van Deventer, J. S. J. Understanding the relationship between geopolymer composition, microstructure and mechanical properties. Colloids Surf., A 2005, 269, 47. (3) Lee, W. K. W.; van Deventer, J. S. J. Structural reorganisation of class F fly ash in alkaline silicate solutions. Colloids Surf., A 2002, 211, 49. (4) Rowles, M.; O’Connor, B. Chemical optimisation of the compressive strength of aluminosilicate geopolymers synthesised by sodium silicate activation of metakaolinite. J. Mater. Chem. 2003, 13, 1161. (5) Duxson, P.; Lukey, G. C.; Separovic, F.; van Deventer, J. S. J. The effect of alkali cations on the incorporation of aluminum in geopolymeric gels. Ind. Eng. Chem. Res. 2005, 44, 832. (6) Duxson, P.; Provis, J. L.; Lukey, G. C.; van Deventer, J. S. J.; Separovic, F. 29Si MAS NMR Investigation of molecular structuring in aluminosilicate geopolymer gels. Langmuir 2005, 21, 3028. (7) Hashin, Z.; Shtrikman, S. A variational approach to the theory of the effective magnetic permeability of multiphase materials. J. Appl. Phys. 1962, 33, 3125.
(8) Weidenfeller, B.; Ho¨fer, M.; Schilling, F. R. Thermal conductivity, thermal diffusivity, and specific heat capacity of particle filled polypropylene. Composites, Part A 2004, 34, 423. (9) Schilling, F. R.; Partzsch, G. M. Quantifying partial melt portion in the crust beneath the Central Andes and the Tibetan Plateau. Phys. Chem. Earth 2001, 26, 239. (10) Brunauer, S.; Emmett, P. H.; Teller, E. Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 1938, 60, 309. (11) Gustafsson, S. E. Transient plane source techniques for thermal conductivity and thermal diffusivity measurements of solid materials. ReV. Sci. Instrum. 1991, 62, 797. (12) Bouledroua, M.; Dalgarno, A.; Cote, R. Viscosity and thermal conductivity of Li, Na, and K gases. Phys. Scr. 2005, 71, 519. (13) Shaw, G. H.; Caldwell, D. A. Sound-wave velocities in liquid alkali metals studied at temperatures up to 150 °C and pressures up to 0.7 GPa. Phys. ReV. B 1985, 32, 7937. (14) Sachdev, K.; Sharma, K. B.; Saxena, N. S.; Sinha, A. K. Temperature dependence of thermal conductivity and thermal diffusivity of rocks and minerals. High Temp. High Press. 1995/1996, 27/28, 47. (15) Ramires, M. L. V.; Nieto de Castro, C. A.; Nagasaka, Y.; Assael, M. J.; Wakeham, W. A. Standard reference data for the thermal conductivity of water. J. Phys. Chem. Ref. Data 1995, 24, 1377. (16) Bigg, P. H. Density of water in SI units over the range 0-40 °C. Br. J. Appl. Phys. 1967, 18, 521. (17) Provis, J. L.; Duxson, P.; Lukey, G. C.; van Deventer, J. S. J. A statistical thermodynamic model for Al/Si ordering in aluminosilicates. Chem. Mater. 2005, 17, 2976. (18) Duxson, P.; Lukey, G. C.; Van Deventer, J. S. J. Structural evolution and phase stability of Na-geopolymers derived from metakaolin up to 1000 °C. Submitted for publication in Langmuir 2006, 22, 8750-8757. (19) Duxson, P.; Lukey, G. C.; Mallicoat, S.; Kriven, W. M.; van Deventer, J. S. J. The effect of alkali and Si/Al ratio on the mechanical properties of metakaolin-based geopolymers. Colloids Surf., A, in press. DOI: 10.1016/j.colsurfa.2006.05.044. (20) Duxson, P.; Lukey, G. C.; van Deventer, J. S. J. Physical evolution of Na-geopolymer derived from metakaolin up to 1000 °C. J. Mater. Sci. 2005, in press. (21) Rahier, H.; Van Mele, B.; Wastiels, J. Low-temperature synthesized aluminosilicate glasses. 2. Rheological transformations during low-temperature cure and high-temperature properties of a model compound. J. Mater. Sci. 1996, 31, 80. (22) Babu, C. S.; Lim, C. Theory of ionic hydration: Insights from molecular dynamics simulations and experiment. J. Phys. Chem. B 1999, 103, 7958. (23) Kollman, P. A.; Kuntz, I. D. Studies of cation hydration. J. Am. Chem. Soc. 1972, 94, 9236. (24) Lefevre, B.; Saugey, A.; Barrat, J. L.; Bocquet, L.; Charlaix, E.; Gobin, P. F.; Vigier, G. Intrusion and extrusion of water in hydrophobic mesopores. J. Chem. Phys. 2004, 120, 4927. (25) Weitkamp, J.; Ernst, S.; Gunzel, B.; Deckwer, W. D. Separation of gaseous water ethanol mixtures by adsorption on hydrophobic zeolites. Zeolites 1991, 11, 314. (26) Maqsood, A.; Kamran, K.; Gul, I. H. Prediction of thermal conductivity of granite rocks from porosity and density data at normal temperature and pressure: in situ thermal conductivity measurements. J. Phys. D 2004, 37, 3396. (27) Duxson, P.; Lukey, G. C.; Van Deventer, J. S. J. The thermal evolution of metakaolin geopolymers: Part 1sPhysical evolution. J. NonCryst. Solids 2006, in press.
ReceiVed for reView February 15, 2006 ReVised manuscript receiVed May 21, 2006 Accepted September 7, 2006 IE060187O