Low energy pre-blended mortars: Part 1 – Control of the sand drying process using a lime drying technique

Construction and Building Materials 101 (2015) 466–473

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Low energy pre-blended mortars: Part 1 – Control of the sand drying process using a lime drying technique D.C. Hughes ⇑, J.M. Illingworth Bradford Centre for Sustainable Environments, University of Bradford, Bradford, West Yorkshire BD7 1DP, UK

h i g h l i g h t s
A technique is described for low-energy drying of wet sand to produce pre-blended mortars.
Quicklime is used as the drying medium.
The dominant processes are chemical combination and evaporation of free water.
Four principal process control factors have been identified.
The drying process produces slaked lime to form a component of the binder phase.

a r t i c l e

i n f o

Article history: Received 13 May 2015 Received in revised form 19 August 2015 Accepted 14 October 2015

Keywords: Mortar Sand drying Quicklime Slaking Process control

a b s t r a c t Production control methods allow factory produced mortars to be supplied to a more consistent formulation than site produced mortars. However, there is scope to enhance their ‘‘sustainability” credentials by addressing the methods of drying the wet sand and the use of lower energy hydraulic components. This paper describes the development of a technique in which quicklime is added in controlled quantities to remove free water by both chemical combination and evaporation. The slaked lime so generated is porous and a third mechanism of absorption is suggested which, however, might have adverse effects during storage of the pre-blended mortar. The principal process-control factors are lime addition based upon a ratio of the stoichiometric requirements for complete slaking of the quicklime, free moisture content of the sand, mixing time of the combined sand and quicklime, and storage of the mixed material. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction The use of pre-blended dry silo mortars is a relatively new concept in UK construction, but is standard practice throughout mainland Europe. Dry silo (and bagged) mortars are produced in computer controlled facilities which allow accurate gravimetric batching of fine aggregate, binders and appropriate admixtures. Hence, a wide range of site specific mix designs may be created and many of the potential problems arising from ‘on-site’ volumetric batching are eliminated. Further advantages of the pre-mix system include reduced wastage, lower labour costs and cleaner, quieter construction practice [1]. Unfortunately, these advantages are somewhat offset by an increase in the embodied energy of pre-blended mortars. Quarried sands typically contain significant amounts of free water, typically ⇑ Corresponding author at: 10 High Fold Lane, Utley, West Yorkshire BD20 6ES, UK. E-mail addresses: [email protected] (D.C. Hughes), jamesillingworth@ hotmail.co.uk (J.M. Illingworth). http://dx.doi.org/10.1016/j.conbuildmat.2015.10.043 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.

within the range of 5–10% and this must be removed prior to mixing with the hydraulic binders in order to prevent deleterious hydration reactions occurring during storage. In this paper we take free water to be that in excess of the water required to saturate the pore space within the sand particles and therefore exists both on the surface and between the particles. The drying of the wet aggregate usually takes place in large diesel-fuelled kilns, an energy intensive process which imposes an additional environmental cost upon the material. This is of particular concern in the context of the current drive for greater sustainability in construction materials. Lime mortars have been used for millennia. One traditional technique for their production is known as ‘hot lime slaking’ in which freshly calcined quicklime is mixed with the wet sand and worked during the slaking period or whilst the material is hot [2]. In this approach the aim is to slake the lime prior to use after a period of several days have elapsed and thus the slaked lime is simply viewed as the binder. However, at the extreme where the contents of water and quicklime are ‘‘in balance” the outcome is a fully slaked lime and dry sand. Should it be possible to sensibly achieve this balance under operational conditions, the process

467

D.C. Hughes, J.M. Illingworth / Construction and Building Materials 101 (2015) 466–473

may yield a material which forms the basis for a factory produced mortar to which an additional hydraulic binder component can be added. In this case the lime is both a drying agent and a component of the binder phase. In the current work the production of ‘low-energy’ pre-blended mortars using a controlled quicklime drying technique for the aggregate is described. The ‘as-received’ wet aggregate is mixed with controlled amounts of quicklime and water is removed from the aggregate via the hydration of the quicklime. Further water is removed by evaporation as a consequence of the heat generated by the exothermic hydration reaction; thus, there is no need to utilise the full stoichiometric amount of quicklime to fully combine with the moisture within the sand (i.e. 3.113 g of CaO being required to combine with 1 g water). The ‘green’ credentials of these mortars may be further enhanced by careful selection of the hydraulic component. Recycled waste materials such as ground granulated blast furnace slag (ggbs) and pulverised fuel ash (PFA) possess far lower embodied energy than the more commonly used Portland cement (PC) or natural hydraulic lime (NHL) [3]. In order to produce a true ‘low-energy’ pre-blended mortar, this research concentrates primarily on the addition of ggbs as the hydraulic phase and demonstrates that a range of mortars with differing physical properties may be engineered. The work is reported in two parts. The first describes the control methodology of the drying process itself whilst the second [4] records the properties of ggbs/lime mortars and Roman cement mortars which have been retarded by a pre-hydration technique [5]; both of which utilise the lime-drying process. It is apparent that variations in the as-delivered moisture content of the sand will be accompanied by variations in the amount of quicklime required and, hence, the amount of slaked lime produced. This variation is accounted for in the mix design procedure described in Part 2 of the paper. Two research projects have been progressed alongside each other but not necessarily contemporaneously. One has been a laboratory investigation whilst the other was based in the commercial premises of Lime Technology Ltd in Didcot, UK where the principal focus has been the application of principles arising from the first project (reported here-in) as they apply to locally sourced materials. Such a factor was deemed important to minimise the carbon footprint of the product. This paper is a combination of data obtained in both projects. 2. Experimental procedures Nitrogen adsorption–desorption isotherms were determined at 77 K using a Micromeritics ASAP 2000 automated gas adsorption apparatus. The samples were outgassed at 100 °C under vacuum for 18 h prior to analysis. The specific surface area (ABET) of the samples was calculated by the Brunauer, Emmett and Teller (BET) method [6] using adsorption data in the relative pressure (P/Po) range 0.05–0.25. The total pore volume (VP) was determined from the volume of adsorbed N2 at P/Po = 0.99, assuming a liquid density for N2 of 0.8081 g/cm3. Thermogravimetric analysis (TGA) was used to determine the proportion of Ca (OH)2, CaCO3 and absorbed water in both the quicklimes and derived slaked limes. The TGA profiles were produced using a Stanton Redcroft TG761 with a heating rate of 20 °C/min under a dry nitrogen flow. The weight loss events at 400–550 and 600– 850 °C were used to determine the quantity of Ca(OH)2 and CaCO3 respectively whereas the weight loss up to 250 °C was used to determine the moisture content of the sample. The free lime measurements were obtained by titration using the method described in BS EN 459-2 for the analysis of air limes. The same Standard was used to evaluate the soundness of slaked limes produced during the drying process. The bulk density of sands, quicklime and slaked lime powders was determined using a Hosokawa Powder Densometer. The values cited in this paper are the average values of the aerated and tap densities. The former is obtained by vibrating the powder through a 0.710 mm sieve into a cylinder of known volume whilst the latter is determined by compacting the powder for 3 min at the standard tap rate. The absorption of the sands was obtained by first soaking the sand in excess water for 24 h. Then it was gently dried in a frying pan to SSD and weighed. The determination of the SSD condition is made when the finest particles adhere to a metal spatula when rotated vertically but are detached by a gentle tap. The sample was then oven dried and the absorption expressed as a percentage of the oven dry weigh.

An Environmental Scanning Electron Microscope (FEI Quanta 400 E-SEM) was used in back-scattered mode to illustrate the morphology of slaked lime particles. Unless otherwise stated all oven drying of materials was conducted in a closed system in air at 110 °C circulated through silica gel and soda lime.

3. Sand characterisation Three silica sands have been used in this work and the results of dry sieve analysis are shown in Table 1. Sand 1 has been previously used by Lime Technology Ltd as a component of their conventional dry silo mortar which uses natural hydraulic lime as the binder. Sand 2 was identified as a more sustainable source being closer to their Didcot factory and was identified following the commencement of the investigation. Sand 3 is an atypical sand with a grading with only minimal overlap in grading with the lime which permitted the physical extraction of lime for analysis with only a small contamination of sand resulting from surface degradation of the sand during the mixing phase. The water absorption values were determined to be 1.5%, 3.1% and 1.0% for sands 1–3 respectively; their densities are 1636, 1671 and 1614 kg/m3 respectively. 4. Lime characterisation Six powdered quicklimes were sourced for initial evaluation and finally two were selected for mortar production as being indicative of highly reactive and less reactive quicklimes and of commercial interest. In the interest of conciseness only details of these two limes are included here and are detailed in the first column of Table 2. Given the range of suppliers there was no consistency in the measure of reactivity used to classify them. In order to assess their performance as agents for drying sand, mixes have been made using sands 1 and 3 and in all cases only a single mix has been made at each condition; hence, only broad conclusions will be drawn. In the first series a 1 kg mix of sand 1, pre-soaked to yield free moisture contents of 5%, 7.5%, 10% and 12.5% after allowing for the absorption of the sand, and the quicklimes dosed at 50% of the stoichiometric ratio required for full conversion (i.e. 1.556 g of CaO for 1 g of water) were mixed in a domestic Kenwood Chef (speed 1 – approximately 60 rpm) for a period of 1 h. No account of the purity of the quicklime was taken in determining the weight of material to be added to each mix. A type K thermocouple was secured within the charge and the temperature logged every second. At intervals of 5 min the mixing was momentarily halted and the system weighed so that the weight loss due to evaporation could be determined. In the second series quicklime C1 was used at 40%, 50% and 60% stoichiometry together with sand 1 at free moisture contents of 5% and 10% and similarly evaluated. In the third series, sand 3 was used to facilitate the separation of the lime from the aggregate following the mixing period, thus allowing a detailed analysis of the resulting slaked limes. Mixes of 1 kg of sand at 5%, 7.5% and 10% free moisture were prepared at 50% stoichiometric ratio and mixed for 20 min. Subsequently, the mixture was dried at 110 °C until constant weight achieved and then placed over a 0.15 mm sieve to separate the sand and

Table 1 Sand gradings. Mesh size (mm)

1

2

3

4 2 1 0.5 0.25 0.125 0.063

99.3 95.8 91.7 74.8 31.3 7.6 1.9

95.0 83.8 67.4 45.6 25.2 13.6 7.1

100 100 98.4 3.8 0.2 0 0

468

D.C. Hughes, J.M. Illingworth / Construction and Building Materials 101 (2015) 466–473

Table 2 Analyses of quicklime and slaked lime following drying of sand 3 at three free moisture contents. Source

Quicklime

C1

Free lime (%) Insolubles (wt%) CaO (wt%) Ca(OH)2 (wt%) CaCO3 (wt%) ABET (m2/g) VTOT (cm3/g) Bulk density (kg/m3)

93.5 6.0 92.4 1.4 1.3 1.2449 0.004 1250

E

Free lime (%) Insolubles (wt%) CaO (wt%) Ca(OH)2 (wt%) CaCO3 (wt%) ABET (m2/g) VTOT (cm3/g) Bulk density (kg/m3)

92.2 6.1 90.9 1.7 3.8 2.2192 0.007 942

Slaked lime 5% m/c

7.5% m/c

10% m/c

70.8 5.7 0.5 92.9 3.1 14.46 0.088 580

71.0 5.9 0.9 92.6 2.9 16.23 0.094 574

70.9 5.8 0.3 93.2 2.9 15.19 0.098 585 68.7 4.8 0 90.8 6.9 18.16 0.109 559

lime fractions. Both quicklime and slaked lime were analysed for chemical species, surface area, porosity and density. The temperature profiles and weight loss due to evaporation during the first series of tests are shown in Fig. 1. It is apparent that, for a given lime, the peak temperature increases with moisture content as does the initial rate at which the water is evaporated when expressed as a percentage of free water evaporated after 5 min (Table 3). It was noted that the more reactive limes showed similar performance and both 10% and 12.5% free moisture contents. It can be seen in Table 2 that the reactive component of the quicklimes is only 90.9% for quicklime E and 92.4% for quicklime C1. Had this purity of the quicklime been accounted for an evaporation of approximately 54% would be required to remove all water assuming full slaking of the quicklime occurs. As such, it can be seen that a 5 min drying time yields a wet sand. However, a drying time of 1 h yields theoretically dry sands, although those produced using the less reactive lime C1 is marginal whilst the reactive lime would appear to have caused the evaporation of such quantities of the free water so as not to provide sufficient for the complete slaking of the quicklime, particularly at 5% free moisture content. It is apparent that after 1 h mixing the most reactive limes indicate the greatest evaporation potential being with the lowest sand moisture content – the reverse of early age behaviour. However, such long mixing times are unlikely to be commercially attractive even if sand was delivered at this relatively low moisture content. The commercial considerations include excessive energy consumption and the impact on other plant operations if the mixer was monopolised for so long on a single operation. Table 4 shows the results from the second series of tests. The amount of free water remaining has been estimated after accounting for that absorbed in the sand particles, assuming that they remain saturated. The degree of lime conversion has been estimated by first determining the amount of water which has reacted by calculating the difference between the total amount of water added to each mix and the sum of the water evaporated during mixing and the residual water determined as remaining within the mix which could be removed by heating at 105 °C until constant weight was achieved. Subsequently, the amount of CaO present as Ca(OH)2 was calculated assuming stoichiometric combination of CaO and water. As before, it is apparent that the maximum temperature increases with the moisture content of the sand as well as with increases in the stoichiometric ratio of quicklime utilised. It can be seen that low additions of quicklime do not fully remove all the free water, despite the visual appearance of the mixture being of a free-flowing dry material, whilst

high quicklime additions do not necessarily fully slake all of the quicklime (these data have been emboldened in Table 4). The former case could lead to degradation of the combined mixture of lime-dried sand and hydraulic binder during the storage period in a silo before use on site whilst the latter may yield an unsound product should the mix water of the final mortar not rapidly slake the remaining quicklime during the mixing process. The use of quicklime C1 at 60% stoichiometric ratio and 10% moisture content generates such reactivity that the measured weight loss is greater than that attributable to the original free water content and it is surmised that water is removed from the pores of the sand particles and is reported in Table 4 as 0% water remaining. It is emphasised that only broad trends should be extracted from this data since subsequent experimentation using thermogravimetry has shown that slaking of CaO proceeds during the oven drying process until all of the water is removed from the system. Columns 2–4 of Table 2 show the analyses of the original quicklimes together with those of the slaked limes produced at 3 free moisture contents of sand 3 (series 3); comparison with a commercial CL90 slaked lime may be found in Table 5. The purity of the limes is reduced by the presence of insoluble residues, probably clays and sand, and small amounts of slaked and carbonated lime. Following the drying process it is apparent that the insoluble residue is commonly greater than that present in the original quicklime whereas it might be expected to reduce. This is attributed to a small fraction of sand which has been ground to <0.15 mm during the drying process. Consequently, the surface areas of the slaked product cited in Table 2 are likely to be a slight underestimate of the key hydrated material essential to the hydration of the ggbs, i.e. Ca(OH)2. Nevertheless, it can be seen that the surface area of the slaked lime is at least 1 order of magnitude greater than that of the original quicklime and, further, increases with increasing moisture content of the sand reflecting the increasing reactivity with increases in moisture content as shown by the increased maximum temperatures (Fig. 1). The surface areas obtained, for the quicklimes and slaked limes, fall within the typical range for such materials [7]. Similarly, the porosity of the slaked lime is greater than that of the original quicklime and increases with moisture content of the sand (see Table 2). Fig. 2 shows significant differences in the N2 adsorption characteristics of the quicklime and its associated slaked lime. The as-received quicklime displays a typical type II isotherm indicative of a non-porous solid [8]. The isotherm is completely reversible and hysteresis is absent. In contrast, the slaked lime displays a type IIb isotherm consisting of a normal type II adsorption branch with type H3 hysteresis on desorption [9]. Similar adsorption isotherms (for N2 on Ca(OH)2) have also been observed by other workers [10]. This form of isotherm indicates the presence of aggregates of non-porous plate-like particles, the individual particles consisting of the crystallites of calcium hydroxide. According to Pashalidis and Theocharis [11], the observed hysteresis loop may be assigned to intercrystalline adsorption of nitrogen within the aggregates in ill-defined pores at the point where crystallites come into contact. The associated SEM image (Fig. 3) appears to confirm the suggested aggregate morphology. Although the pore network proposed from the gas adsorption data is below the resolution of the SEM, larger pores between the crystallites are clearly visible. The presence of such porosity raises the possibility of a third ‘‘water removal” process, that being absorption of water within this space. Of course, this water remains within the system as liquid water and if it were to be ‘‘available” to the ggbs there may be a problem of degradation during storage of the pre-blended mortar. The density of the slaked lime is some 46–59% of that of the source material (see Table 2). However, whilst the variation in density of all of the quicklimes assessed is some 330 kg/m3 that of the

469

D.C. Hughes, J.M. Illingworth / Construction and Building Materials 101 (2015) 466–473

50 60 50

40

40

30

30 20 20

7.5% 10

10 5%

0 0

0

10

20

30

40

50

60

70

80

Temperature (OC)

Temperature (OC)

60 10%

70

(b)

90

70

Free H2O evaporated (%)

12.5% 80

60

10%

70

50

7.5% 60 50

40

40

30

30 20 20 5%

10

10

0

Free H2O evaporated (%)

(a)

90

0 0

10

20

30

40

50

60

Time (min)

Time (min)

Fig. 1. Temperature (°C) and weight loss (%) profiles for limes C1 (a) and E (b) using sand at various values of free moisture content.

Table 3 Percentage evaporation from 50% stoichiometric mixes following 5 min and 1 h of mixing. Lime

C1 E

5 min

1h

5.0%

7.5%

10.0%

12.5%

5.0%

7.5%

10.0%

12.5%

16.4 23.8

21.3 30.1

23.8 36.5

27.0 –

54.2 59.8

54.1 56.7

52.5 56.2

52.9 –

slaked limes is only 43 kg/m3; an important parameter in the volumetric proportioning of the final mortar mixes. Comparison of the slaked product with a commercial CL90 slaked lime (from the supplier of lime C1) shows that, whilst the density of the materials are similar, the sand drying process appears to yield materials with higher surface area and porosity. In summary, the reactivity of the mix is increased as the moisture content of the sand to be dried is increased and as the stoichiometric ratio of quicklime is also increased. To ensure a ‘sound’ hydrated lime, the mixing process must be calibrated in order to balance the evaporation of water with that combined chemically as calcium hydroxide. If the quantity of quicklime added is too high, the exothermic reaction leads to rapid evaporation and insufficient water remains in the aggregate to allow complete conversion of the quicklime. Conversely, too little quicklime will produce a ‘sound’ hydrated product, but the evaporation rate will be too low to facilitate the complete removal of free water leading to potential hydration during storage following the addition of the hydraulic phase. During this stage of work several mortar mixes were produced following lime drying in which the drying phase was halted when the mixture was visually dry and free flowing prior to the addition of ggbs. These were complemented with similar mixes produced with sands pre-conditioned to saturated surface dry (SSD). Both sets of ‘‘dry” mortars were stored in sealed boxes for approximately 3 months prior to being mixed with sufficient water to yield fresh mortar. Whilst the SSD materials could be stored without degradation, those produced using visually assessed lime-dried

Table 4 Performance of mixes with 40%, 50% and 60% stoichiometric additions of quicklime C1. Stoich ratio (%)

TMAX (°C)

Free H2O remaining (%)

CaO converted (%)

5% m/c

10% m/c

5% m/c

10% m/c

5% m/c

10% m/c

40 50 60

34.6 37.1 46.0

49.1 60.2 70.4

16 0 0

21 1 0

99 99 98

99 100 90

sands exhibited degradation; the mortars had the consistency of wet sand and exhibited excessive bleeding in the plastic state, whilst subsequently yielding low strengths in the hardened state. Consequently, it is necessary to refine the degree of control of the drying process in order to achieve an optimum quicklime addition which satisfies both criteria of achieving SSD sand and full conversion of quicklime to calcium hydroxide. 5. Control of the lime drying process The first step is to determine the amount of water to be evaporated such that the remainder is sufficient to satisfy the requirements of the lime to be fully slaked. Thus, the quantity of water removed via chemical combination (CaO to Ca(OH)2) can be calculated according to Eq. (1):

W R ¼ ðC S  C P Þ=100

ð1Þ

where WR = free water removed (% of total free water content), Cs = stoichiometric ratio (%), CP = CaO content of as-received quicklime available for slaking (%). Hence, to ensure an adequately dry material, the remaining water must be removed by evaporation during the mixing process (Eq. (2)):

W E ¼ 100  W R

ð2Þ

where WE = required evaporation (% of total free water content) and is shown as the linear line in Fig. 4. All sands were oven dried at 105 °C for 24 h in an oven without a circulated atmosphere and allowed to cool before the addition of sufficient water to yield the required free moisture content. The wet sands were allowed to stand overnight in an air-tight box to ensure saturation before the addition of the quicklime. Values of free moisture content and lime addition for four combinations of lime and sand are shown in Table 6. Mixing was conducted in a Kenwood Chef (speed 1 – approximately 60 rpm) and the evaporation of free water from the wet-sand/lime mixtures was assessed

Table 5 Analysis of commercial CL90. Source CL90

Free lime (%) Insolubles (wt%) CaO (wt%) Ca(OH)2 (wt%) CaCO3 (wt%) ABET (m2/g) VTOT (cm3/g) Bulk density (kg/m3)

71.3 3.8 0.9 93.0 2.3 11.26 0.072 575

470

D.C. Hughes, J.M. Illingworth / Construction and Building Materials 101 (2015) 466–473 70

VADS (cm3/g) STP

60 50 40 30 Slaked lime 20 10

Quicklime

0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

P/Po Fig. 2. Nitrogen adsorption of quicklime before and after drying process.

gravimetrically following mixing times of 5, 10 and 15 min; a maximum mixing time of 15 min having been agreed based upon commercial considerations. The initial phase of work focused on sand 1. Fig. 4 shows the results for the E/1 combinations of quicklime and sand as specified in Table 6. Optimal drying is considered to occur when the measured evaporation equals that determined by Eq. (2) (see Fig. 4a), whilst sub-optimal drying does not remove sufficient water and super-optimal drying removes too much water leaving a residual amount of free lime. The optimum lime stoichiometric ratio has been determined for each value of free moisture content and mixing time (from Figs. 4a–d) and shown on Fig. 5 (solid symbols). High correlation coefficients were obtained for linear relationships for each mixing time (r2 = 0.954, 0.960 and 0.962 for mixing times of 5, 10 and 15 min respectively). It is apparent that as the moisture content increases a small reduction in the stoichiometric ratio of lime required for optimal drying is observed; this was anticipated following the earlier observations that increased moisture contents of the sand increased the reactivity of the lime and increased the initial rate of evaporation (Table 3). The data also show that as the mixing time is increased, the optimum stoichiometric ratio decreases. Again, this is to be expected as a greater percentage of the free water is removed by evaporation leaving less water to

Fig. 3. Micrograph of lime C after drying process showing aggregate morphology (3000, BSE).

be removed by chemical reaction. The same determinations of optimum lime addition for the C1/1 combinations have been superimposed on Fig. 5 as open symbols for each mixing time. It can be seen that the lower reactivity of lime C1 results in the need for higher lime additions at a given moisture content although the difference is reduced as the mixing time is increased. Additionally, the difference in efficacy of two limes is greater at the higher moisture content. At a later stage sand 2 was identified as a possible commercial source for use with lime C1 and a fresh supply of lime was obtained (C2) from the same supplier. In the intervening period the supplier had altered the production process resulting in an increase in the surface area to 2.0685 m2/g, compared to 1.2449 m2/g for C1. A comparison of the efficacy of limes C1, C2 and E in drying sand 1 of 5% free moisture can be seen in Table 7; it is apparent that the reactivity of the second batch of lime C is much higher than that of the original supply and approaches that of lime E. A comparison of optimum drying conditions for the 4 combinations of lime and sand specified in Table 6 and mixed for 15 min is shown in Fig. 6. The data show that, for a given sand the amount of lime required is reduced as the reactivity of the lime is increased. The same ranking of limes shown in Table 7 is apparent in Fig. 6 with less lime being required as their reactivity increases. The lower reactivity leads to lower evaporation and a greater quantity of the water which must be removed by chemical combination; hence a higher stoichiometric ratio is required. There is the suggestion that limes C2 and E might yield very similar optimum conditions at high values of sand moisture contents reflecting the higher reactivity in the presence of more water. The similar reactivity of limes C2 and E (Table 7) provide similar evaporation and, consequently, the optimum stoichiometric ratios are also quite similar. In contrast, the less reactive lime (C1) requires higher lime additions for optimum drying performance. The type of aggregate also plays a minor role with the more well-graded sand (sand 2) yielding the larger evaporation and requiring a lower lime addition for a given sand moisture content. Thus, it is possible to specify a combination of lime stoichiometry and mixing time to achieve optimum drying for a sand of given free moisture content.

6. Influence of storage In Sections 4 and 5 it was assumed that sub-optimal drying would yield a sound but wet product which could degrade during storage whilst super-optimal drying would minimise storage problems but be responsible for unsoundness. In order to test these assumptions four mixes have been produced with sand 3 and combinations of moisture content and stoichiometric ratios of lime addition determined from the use of Lime E with sand 3 to yield sub-optimal, optimal and super-optimal mixes of processed material; optimal drying conditions were specified for both 5 and 15 min mixing periods. Sub-optimal and super-optimal mixing was for 15 min mixing and at 50% stoichiometric ratio, 5.69% free moisture and at 67% stoichiometric ratio, 5.69% free moisture respectively. The processed material was stored in air-tight containers and analysed at various time intervals following mixing up to a maximum of 12 weeks storage. The resulting data are presented in Tables 8–11. The first analysis was undertaken immediately after the prescribed mixing time in order to provide base-line data. The sand was removed by sieving over a 0.15 mm sieve to leave the lime fraction. This was analysed for total free lime by the titration method and CaCO3 and Ca(OH)2 using TGA. Following dissolution of the lime fraction in 1 M HCl it is apparent that a small amount of ground sand was generated during the drying process which

471

D.C. Hughes, J.M. Illingworth / Construction and Building Materials 101 (2015) 466–473

b - 5.5% m/c

a - 5% m/c 75

75

70

70

Super-optimal drying

65

indicates optimum drying

Evaporation (%)

Evaporation (%)

65

60 55 Sub-optimal drying

50 45

15 min mixing 40 10 min mixing

35

60 55 50 45 40 35 30

30 5 min mixing

25

25 20

20 35

40

45

50

55

60

65

35

70

40

45

50

CaO Stoichiometric ratio (%)

60

65

70

65

70

d - 7.5% m/c

c - 6.5% m/c 75

75

70

70

65

65

Evaporation (%)

Evaporation (%)

55

CaO Stoichiometric ratio (%)

60 55 50 45 40 35

60 55 50 45 40 35

30

30

25

25 20

20 35

40

45

50

55

60

65

35

70

40

45

50

55

60

CaO Stoichiometric ratio (%)

CaO Stoichiometric ratio (%)

Fig. 4. Determination of optimum drying conditions using different mixing times for sands of different free moisture contents; (a) 5%, (b) 5.5%, (c) 6.5%, (d) 7.5%. Key shown on (a) applies to all figures.

Table 6 Specification for drying regimes for 4 combinations of lime and sand. Lime/sand

Free m/c of sand (%)

Stoichiometric ratio (%)

E/1 C1/1 C2/1 C2/2

5, 5.5, 6.5 & 7.5 5 & 7.5 3, 5, 7 3, 4, 5, 6 & 7

55, 50, 55, 50,

60, 55, 60, 55,

Table 7 Evaporation (%) from 3 mixes using sand 1 of free m/c 5% and lime content of 55% stoichiometric ratio.

65 & 70 60, 65 & 70 65 60, 65 & 70

5 min 10 min 15 min

Optimum stoichiometric ratio (%)

75

70

5 min mixing

65

10 min mixing

60

15 min mixing

55

50 4

4.5

5

5.5

6

6.5

7

7.5

8

Sand moisture content (%)

Fig. 5. Optimum lime-drying parameters using quicklimes C1 & E and sand 1.

could be quantified. This allowed a correction to be applied to the amount of CaO present as CaCO3, Ca(OH)2 and CaO in the lime frac-

C1

C2

E

17.0 29.2 37.4

24.5 34.8 41.1

26.2 37.2 42.8

tion alone. Soundness evaluations were made in accordance with BS EN 459-2 after various periods of storage; a maximum expansion of 2 mm is permitted for a sound lime. It is apparent that the assumption of different moisture contents with different drying regimes has been confirmed with super-optimal drying yielding the driest material and suboptimal drying the wettest. The assumption concerning soundness was broadly correct, yet was not immediately recognised. It can be seen that immediately following mixing the CaO conversion had not progressed to completion resulting in an unsound lime. However, a 24 h period of storage increased the quantity of Ca(OH)2 with a corresponding decrease in CaO, the CaO effectively acting as an in situ desiccant. For optimal and sub-optimal conditions, the storage period also facilitated the production of a sound hydrated lime whereas the super-optimal material remained unsound for the first 4 weeks of storage. Following these data a period of storage of the dried sand for 24 h was added to the specification of stoichiometric ratio, mixing time and free moisture content to yield optimum drying conditions. It is only after this

472

D.C. Hughes, J.M. Illingworth / Construction and Building Materials 101 (2015) 466–473 Table 10 Sub-optimal drying following 15 min mixing for E/3.

Optimum stoichiometric ratio (%)

65

C2/1

C1/1

60 C2/2

E/1

55

50 3

4

5

6

7

8

Sand moisture content (%) Fig. 6. Optimum lime-drying parameters using quicklimes C1, C2 & E and sands 1 & 2.

Table 8 Optimal drying following 5 min mixing for E/3. Sample

H2O (wt%)

ABET (m2/g)

% Ca(OH)2 (as CaO)

% CaCO3 (as CaO)

% CaO

Soundness exp (mm)

CaO After mix 24 h 1 week 2 weeks 4 weeks 6 weeks 8 weeks 12 weeks

0 4.3 2.5 2 1.9 2 1.7 1.6 1.3

2.2 19.9 23.5 21.7 21.6 20.6 19.9 18.1 14.3

1.3 80.4 90.8 91.0 90.6 90.7 91.0 90.8 90.4

2.2 3.9 4.0 3.8 4.1 4.1 4.0 4.2 4.4

92.8 12.0 1.5 1.5 1.6 1.6 1.4 1.4 1.5

>4 >2 0.16 – – – – – –

Table 9 Optimal drying following 15 min mixing for E/3. Sample

H2O (wt%)

ABET (m2/g)

% Ca(OH)2 (as CaO)

% CaCO3 (as CaO)

% CaO

Soundness exp (mm)

CaO After mix 24 h 72 h 96 h 1 week 2 weeks 4 weeks 6 weeks 8 weeks 12 weeks

0 5.4 5.1 5.1 4.8 5 4.7 4.4 3.2 1.7 1.2

2.2 20.6 23.7 – – 23.2 22.9 22.9 21.1 20.3 16.0

1.3 74.7 89.5 89.6 89.5 89.3 89.0 89.3 89.1 89.0 88.4

2.2 6.0 5.8 5.8 5.9 5.8 6.1 5.9 6.4 6.5 6.9

92.8 15.6 1.1 1.0 1.0 1.2 1.2 1.1 0.8 0.9 1.0

>4 >2 0.09 – – – – – – – –

storage that the ggbs is blended into the mixture and the dry mortar stored until required for use. The 24 h storage period may ultimately be shown to be conservative; however, the potential for a combination of further slaking and alkali-activated hydration of the ggbs resulting from any residual water in the system must be acknowledged. The initial increase in Ca(OH)2 observed during the first 24 h is accompanied by an increase in the BET surface area, in accordance with previous data. The long-term development of surface area is a function of the drying conditions. The storage of optimally dried material produces a reduction in surface area following the increase observed in the first 24 h. The rate of decrease increases with period of storage. Fig. 7 shows adsorption isotherms and

Sample

H2O (wt%)

ABET (m2/g)

% Ca(OH)2 (as CaO)

% CaCO3 (as CaO)

% CaO

Soundness exp (mm)

CaO After mix 24 h 1 week 2 weeks 4 weeks 8 weeks 12 weeks

0 10 12.5 12.4 12.8 10.6 9.6 7.0

2.2 20.9 23.3 24.2 25.9 24.9 24.6 23.1

1.3 77.5 89.3 90.6 90.0 89.2 89.1 88.6

2.2 5.9 5.4 4.7 5.3 5.8 6.3 6.9

92.8 12.9 1.6 1.0 1.0 1.4 1.0 0.9

>4 >2 0.00 – – – – –

corresponding BJH plots for a fresh sand slaked lime and the same material following a 12 weeks storage period. The data show a reduction in the adsorption capacity of the lime following storage and a large reduction in the volume of smaller pores (ca. <100 Å). It is possible that these effects could result from carbonation of the Ca(OH)2 plates at their points of contact, where the water within the stored mixtures would most likely condense. This may lead to the creation of closed porosity and a corresponding decrease in the surface area accessible to the nitrogen adsorbate. In contrast, the sub-optimal drying conditions produce a more prolonged period of increase in surface area, up to an age of 2 weeks before a more gradual rate of decrease. The data show that the extended period of increase is not related to formation of Ca(OH)2 from CaO and is likely related to the high water content of the material. It is proposed that the high water content facilitates dissolution and subsequent re-crystallisation of the Ca(OH)2, similar to that observed during the storage of lime putties [12]. These workers observed significant increases in surface area during storage of lime putties and proposed the generation of ‘platelike, submicrometer portlandite crystals from preexisting large prismatic crystals due to preferential dissolution of prism faces and heterogeneous secondary crystallisation of submicrometer, platelike portlandite crystals on preexisting large portlandite crystals’. The reduction in surface area of the slaked lime may have an impact on the water demand of stored mortars to achieve a given level of workability. In particular, mortars with a high proportion of lime might be expected to require less water following prolonged storage with associated impacts on strength performance. As expected, the data in Table 11 show a larger proportion of unreacted CaO within the super-optimal lime when compared to the optimal and sub-optimal materials, thus resulting in a lower surface area after the initial 24 h storage period. The surface area then remains relatively stable throughout the 12 weeks storage period. It is considered likely that the main reason for this contrasting behaviour is the disparity in carbonation between the different materials. When compared to the other limes, the CaCO3 content of the super-optimal material does not show an increase during storage, presumably as a consequence of the drier conditions in the sand-lime mixture. Hence, the carbonation mechanism proposed in the previous paragraph for the optimal lime does not take place. The lack of residual water also prevents an increase in the surface area by the dissolution mechanism proposed for the sub-optimal lime. It should be noted that the CaO content of the super-optimal lime gradually decreases during storage to eventually produce a sound material after 4 weeks. This in situ slaking occurs via reaction with atmospheric water vapour and/or absorbed water present within the pore structure of the aggregate. It would normally be expected that this additional slaking would cause a significant increase in the surface area of the lime. However, previous workers have shown that air-slaking results in hydrated limes with very low surface area [13,14].

D.C. Hughes, J.M. Illingworth / Construction and Building Materials 101 (2015) 466–473

sand increases. Thus, the stoichiometric amount of quicklime required to achieve optimum drying decreases as the free moisture content of the sand increases.
The variability in density of the slaked lime is less than that observed between the original quicklimes.
Although the required amount of water has evaporated following the specified mixing time to yield optimum drying a further storage period is advised to complete the slaking of the quicklime.

Table 11 Super-optimal drying following 15 min mixing for E/3. Sample

H2O (wt%)

ABET (m2/g)

% Ca(OH)2 (as CaO)

% CaCO3 (as CaO)

% CaO

Soundness exp (mm)

CaO After mix 24 h 1 week 2 weeks 4 weeks 8 weeks 12 weeks

0 1.2 1.2 0.9 1 1 1.1 1.0

2.2 18.7 20.5 20.2 20.8 20.3 19.5 18.9

1.3 81.0 85.9 89.4 89.4 89.6 90.5 90.4

2.2 4.3 4.4 4.3 4.4 4.3 4.5 4.5

92.8 11.0 6.1 2.7 2.6 2.4 1.3 1.4

>4 >2 >2 2.29 2.19 1.86 0.23 0.22

473

Acknowledgements The authors wish to acknowledge the financial support of the Engineering and Physical Sciences Research Council (EP/ D025036/1), KTP funding from Technology Strategy Board (Prog. No.: 876) and Lime Technology Ltd, and Castle Cement Ltd, Civil and Marine Ltd and Lhoist UK Ltd for the supply of materials.

70

VAds (cm3/g) STP

60 50 40

References

Fresh lime

30

[1] 20

[2]

10 Stored lime

[3]

0 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

P/Po Fig. 7. Nitrogen adsorption for optimally dried sand – lime sampled immediately after mixing (fresh) and the same material following a 12 weeks storage period (stored); solid line – adsorption, dotted line – desorption.

[4]

[5]

[6]

7. Conclusions It has been shown that the addition of quicklime to wet sand can result in a material suitable for producing dry silo mortars. However, the use of visual appearance is inadequate to ensure optimal drying. The main conclusions drawn from this study are that;

[7] [8]

[9] [10]

[11]


Drying of the wet sand is achieved by both hydration of the quicklime and evaporation as a result of the exothermic reaction. It is suggested that a further mechanism may account for absorption of water into the porous calcium hydroxide produced during the process.
For a given quicklime the peak temperature and initial rate of evaporation are increased as the free moisture content of the

[12] [13] [14]

Mortar Industry Association, Factory Produced Ready to Use Mortar for Masonry, Data Sheet No. 2, Issue 3, Mortar Industry Association Publications, London, 2005. A. Forster, Hot-lime mortars: a current perspective, J. Archit. Conserv. 3 (2004) 7–27. G. Hammond, C. Jones, Inventory of Carbon and Energy (ICE) Summary, Version 2.0, 2011. Available at: (accessed 20.07.15). D.C. Hughes, J.M. Illingworth, V. Starinieri, Low energy pre-blended mortars. Part 2. Production and characterisation of mortars using a novel lime drying technique, Constr. Build. Mater. (2015) (accepted for publication). V. Starinieri, D.C. Hughes, C. Gosselin, D. Wilk, K. Bayer, Pre-hydration as a technique for the retardation of Roman cement mortars, Cem. Concr. Res. 43 (2013) 1–13. S. Brunauer, P.H. Emmett, J. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60 (1938) 309–319. J.A.G. Oates, Lime and Limestone, Wiley-VCH, New York, 1998. K.S.W. Singh, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Reporting physisorption data for gas solid systems with special reference to the determination of surface area and porosity, Pure Appl. Chem. 57 (1985) 603–619. F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids: Principals, Methodology and Applications, Academic Press, London, 1999. C.R. Theocharis, D. Yeates, Changes in the surface properties of calcium hydroxide upon ageing: a spectroscopic and gas sorption study, Colloids Surf. 58 (1991) 353–361. I. Pashalidis, C.R. Theocharis, The effect of sorbed toluene on the surface properties of calcium hydroxide, J. Chem. Technol. Biotechnol. 71 (1998) 223– 226. C. Rodriguez-Navarro, E. Hansen, W.S. Ginell, Calcium hydroxide crystal evolution upon aging of lime putty, J. Am. Ceram. Soc. 81 (1998) 3032–3034. T.C. Miller, A Study of the Reaction between Calcium Oxide and Water, Report to the National Lime Association, Washington, USA, 1961. J.A. Murray, Summary of Fundamental Research on Lime and Application of Results to Commercial Problems, Report to the National Lime Association, Washington, USA, 1956.