Effects of sticky rice addition on the properties of lime-tile dust mortars

The present study focused on the characteristics of lime-tile dust mortars with sticky rice addition, which were prepared according to historic Chinese accounts. Results showed that addition of sticky rice into mortars could improve water retention capacity, prolong setting time, and reduce bulk density values. Sticky rice addition also had long-term detrimental impacts on mortar strength development, concurrently increasing their water absorption values. FTIR, XRD and TG-DSC analyses revealed that the addition of sticky rice into mortars could slow down the carbonation as well as hydration process, at the same time forming both calcite and aragonite as carbonation products.


Introduction
Sticky rice, also known as glutinous rice or waxy rice, is a type of rice mainly grown in East and Southeast Asia, characterized by its opaque appearance, high amylopectin content and sticky texture. Apart from being a staple food in China, sticky rice was widely applied in various other fields including architectural construction. For instance, many historic Chinese literatures accounted the use of lime-sticky rice mortars to build river dams [1], barns [2], flood gates [3] and tombs [4]. Sticky rice porridge was also commonly added in lime-clay-sand mortars to build forts [5] and Fujian Tulou (Hakka earth buildings) [6,7].
Relevant studies have reported the effects and mechanism of sticky rice addition in air lime mortars. Yang et al. [8] found that sticky rice could significantly improve mechanical strengths and compatibility of air lime mortars and control the growth of CaCO 3 crystals. Zhao et al. [9] reported that the addition of sticky rice porridge at 5% -w/w in air lime mortars could accelerate mortar setting and hardening, increase compressive and bonding strength, at the same time slightly reducing mortar density and water resistance. Zhang et al. [10] reported that the mechanical strengths of lime-sticky rice mortars can be significantly decreased by cyclic wetting-drying and dilute sulfate acid actions, causing more porous matrices and coarser particles. Yang et al. [11] concluded that sticky rice addition could induce pronounced heterogeneity in mortars, leading to inhomogeneous strain field and produce micro-cracks during wetting-drying cycles. The results of Zeng et al. [12] showed that biomineralization may occur in lime-sticky rice mortars, where sticky rice functioned as template to control the growth of CaCO 3 crystals.
To date, however, research on sticky rice addition in mortars with potential hydraulicity has been relatively scarce [13,14]. Therefore, the present study aims to understand some preliminary effects caused by sticky rice addition in hydraulic mortars. To achieve this objective, it was proposed by the current study to first search in historic Chinese accounts to come up with a mortar mix to be executed with a certain grade of plausibility; then prepare lab-specimens which are based on the mortar recipe; finally after experimental study, try to understand the effects caused by the sticky rice addition in the studied mortar. Results of the present study could contribute to the understanding of the material, subsequently shedding light on its innovative use in conservation and repair.

Open Access
*Correspondence: yangfwnwu@163.com 1 School of Cultural Heritage, Northwest University, Xuefu Avenue 1, Xi'an 710000, China Full list of author information is available at the end of the article

Mortar recipe for specimen preparation
Some historic sources [1,[15][16][17] from Ming (1368-1635 A.D.) and Qing dynasty (1636-1912 A.D.) accounted that adding tile dust to lime would make mortars impervious and strong. Some other literatures [15,18] even suggested adding organic additives such as sticky rice into lime-tile dust mortars. It is generally accepted that fired clay materials such as brick/ tile dust, when formulated with air lime, can react with lime and form calcium silicate hydrates and calcium aluminate hydrates, which could increase mortars' setting speed [19], mechanical strengths [20] and durability [19].
In the current study, the choice of prepared mortar mix compositions and proportions are based on historic accounts, and the reasons for making such choices are listed as follows.
• Inorganic components. A 17th-century Chinese literature on military defense (武备志 [16]) reported that sand can be excluded from the lime-tile dustsand recipe; for mortars with organic addition, fine tile dust is commonly used. Thus, in the current study, the inorganic components in the mortar specimens were lime and fine tile dust, and sand was excluded to simplify the mortar mix. • Proportion between lime and tile dust. A historic literature [16] revealed that the hydraulic lime-tile dust recipe was in fact derived from the traditional Chinese tabia (lime-clay) recipe. Clay is of course different from brick/tile dust which are fired materials possessing pozzolanic reactivity. However, the relevant historic accounts on lime-brick/tile dust mortars are scarce, and they do not contain some necessary information for mortar specimen preparation in the present study. Fortunately, the historic Chinese lime-brick/tile dust mortars evolved from lime-clay mortars (tabia), and the accounts on the latter are abundant, which could in turn serve as references for the former. Various proportions of lime: clay can be found in literatures (e.g. 5:3-v/v [5,21], 14:3-v/v [22], 3:1-v/v [4]), yet one of the most famous and commonly adopted recipes was accounted in 天工开物 [23], a renowned 17th-century Chinese encyclopedic work on agriculture, manufacture and industry: "To construct tombs and reservoirs, take 1 part of lime and 2 parts of clay and sand, and knead them with sticky rice porridge or kiwi vine juice. " Therefore, in the present study, the chosen ratio between lime and tile dust for lab-prepared mortar specimens was 1:2v/v. • Proportion between sticky rice and inorganic components. In historic records, the most frequently accounted sticky rice proportion in masonry mortars was raw sticky rice: lime = 1:20-v/v [3,[24][25][26], which was adopted in the present study for specimen preparation. • Proportion between sticky rice and water in sticky rice porridge. Unfortunately, the proportion between sticky rice and water wasn't found in historic literature. However, with the knowledge that the proportion between raw sticky rice grains (V RSR ) and lime powder (V L ) is 1: 20-v/v (previous paragraph), such information can be deduced. Considering sticky rice porridge as mixing liquid, the proportion in weight between raw sticky rice (m RSR ) and water (m W ) in porridge can be roughly estimated as where D RSR and D L are respectively the loose bulk density of raw sticky rice grains and lime powder, R W/L is the water/lime ratio needed to achieve a lime paste with appropriate consistency for building operations. • Temperature of sticky rice porridge. In some literatures [25][26][27], barrels are included in the construction preparation list for storage of sticky rice porridge, indicating that the porridge could be used at room temperature; while one other literature [21] described that the porridge should be cooked while mixing inorganic mortar components, suggesting that the porridge should be used while still hot. In the current study, mortars with both hot and cold sticky rice porridge addition shall be tested and analyzed to see the possible effect of mixing liquid temperature on such mortars.
To sum up, for the present study, the mortar mix for specimen preparation is composed of hydrated lime, fine tile dust and sticky rice porridge/ water, with ratio of hydrated lime to tile dust of 1:2-v/v, and both hot and cold sticky rice porridge will be added.

Raw materials
Analytical-grade calcium hydroxide (Ca(OH) 2 , 95.0%, Sinopharm Chemical Reagent), fine tile dust, sticky rice and distilled water were used for specimen preparation. Fine tile dust was purchased from a Chinese lacquerware workshop in Xi'an (China). Sticky rice (Jiulihu Food Co. Ltd.) was purchased from a local supermarket in Xi'an (China). The loose bulk density values (determined according to EN 1097-3: 1998 [28]) of calcium hydroxide powder, fine tile dust powder and sticky rice grains were respectively 500, 700 and 880 kg m −3 at ambient temperature and moisture (~ 25 °C, ~ 50% RH), which were calculated as the quotient of the mass of raw materials filling a graduated cylinder without compaction divided by the capacity of the graduated cylinder.
Particle size distribution of fine tile dust ( Fig. 1) was measured by laser diffractometry using Malvern Mastersizer 2000 (Malvern Instruments, U.K.). Particle size of the tile dust was relatively small, with medium grain size (d (0.5)) of 8.6 μm and d (0.9) of 30.1 μm. Mineralogy of fine tile dust was analyzed by a PANalytical X'pert 3 Powder X-ray diffractometer (PANalytical, Netherlands), and Fig. 2 showed that the characteristic mineral compounds were quartz, feldspars (e.g. albite, microcline), mica, hematite and a small amount of calcite. Elemental composition analysis was carried out with an X-ray fluorescence spectrometer (XRF-1800, Shimadzu, Japan), and Table 1 indicated that the tile dust contains major quantities of SiO 2 and Al 2 O 3 , while other components are present in small amounts.

Specimen preparation
Compositions of lab-prepared mortar specimens were shown in Table 2. As mentioned in "Mortar recipe for specimen preparation" Section, the chosen ratio between lime and tile dust was 1:2 -v/v. Since the loose bulk density values of hydrated lime powder and fine tile dust were 500 and 700 kg m −3 ("Raw materials" Section), the lime to tile dust mass ratio was (500*1): (700*2) = 1: 2.8-w/w in the current study.
As mentioned in "Mortar recipe for specimen preparation" Section, the proportion between weight of sticky rice (m RSR ) and water (m W ) in porridge can be roughly estimated as m RSR /m W = D RSR / (20 × D L × R W/L ). After several try-outs, the water/ lime ratio was determined at ~ 0.8 -w/w, and knowing the loose bulk density of raw sticky rice grains (880 kg/m 3 ) and lime powder (500 kg m −3 ) from "Raw materials" Section, m RSR /m W = 880/ (20 × 500 × 0.8) = 11: 100-w/w. Thus, the sticky rice content in porridge would be m RSR /(m W + m RSR ) = 11/ (11 + 100) = 9.91% -w/w, which can be rounded to 10%-w/w. Sticky rice porridge was prepared by first grinding sticky rice grains into fine powder with an electrical pulverizer for 180 s, then mixing the powder with distilled water at 1:9w/w, and finally cooking the mixture for 4 h in an electrical cooker. Water was constantly replenished during cooking to keep the original mixture volume, so that the final sticky rice content in the porridge was 10% -w/w. After cooking,   sticky rice porridge was added in mortars at both "cold" and "hot" temperatures: half of the porridge was taken out and brought to room temperature (~ 25 °C), designated as "cold sticky rice (CSR)"; the other half was left in the cooker while keeping the temperature at ~ 80 °C, designated as "hot sticky rice (HSR)". Mortar mixes were prepared by first mixing the inorganic components, then continuing mixing while adding distilled water (at room temperature) or sticky rice porridge. Water or sticky rice porridge was added according to personal experience so that the mortar would fit the workability features generally required in building operations. Additionally, it was also intended that the same water amount is present in each mortar mix. To be more specific, in Table 2 the proportions among the components of L-TD-W are hydrated lime: tile dust: distilled water = 1: 2.8: 1.8-w/w, and those of L-TD-HSR/CSR are hydrated lime: tile dust: sticky rice porridge = 1: 2.8: 2-w/w. Since the water content in sticky rice porridge is 90% -w/w (as mentioned in previous paragraph), the theoretical water amount in L-TD-HSR/CSR is the same as L-TD-W. For FTIR, XRD and TG-DSC analyses, mortar mixes without tile dust were also prepared-while keeping the original proportions of hydrated lime to water (1: 1.8-w/w) or sticky rice (1: 2-w/w)-intending to further evidence the effects of sticky rice on hydrated lime.
Specimens were prepared in two dimensions: 5*5*5 cm 3 cubes for hardened mortar density, water absorption and compressive strength tests, and Φ 6*1.5 cm 3 discs for FTIR, XRD and TG-DSC analyses. Specimens were stored under both dry (~ 25 °C, ~ 50% RH) and humid (~ 25 °C, ~ 95% RH) conditions for 180 days. Table 3 shows the property tests conducted on fresh and hardened mortars, and the number of specimens tested.  Water retention test was conducted according to Chinese standard JGJ/T70-2009 [29]. Freshly-mixed mortar was poured into a metal ring (Φ 100*25 mm 3 ), then quickly came in contact with filter papers for 2 min, pressed upon by a lead cube weighing 2 kg to simulate the suction from porous substrates. Water retention capacity of the mortar mix was characterized by the mass of water it retained after filter paper absorption. Then, knowing the water content of the tested mortar (Table 2), water retention was calculated by

Specimen Composition Specimen dimension Tests/analyses performed
where m 2 is the quantity of water retained by the filter paper, m 1 is the quantity inside the studied mortar. For each mortar mix, two samples were used for water retention test, and the mean value and standard deviation of the two samples was reported.
Setting time was determined by Vicat apparatus according to BS EN 480-2: 2006 [30], which observes the rate of mortar hardening by dropping a needle from a fixed height and measuring its penetration into the mortar. Initial setting time was determined as the time measured from completion of mixing until the time at which the distance between the needle and the base-plate was (4 ± 1) mm. Final setting time was determined as the time measured from completion of mixing until the time after which the needle no longer penetrates 0.5 mm into the specimen. One sample from each mortar mix was used for setting time test.
Bulk density values of fresh and hardened mortar were respectively determined according to European standards BS EN 1015-6: 1999 [31] and BS EN 1015-10: 1999 [32], by the quotient of its mass and the volume which it occupies. For each mortar mix, two fresh mortar samples were used for fresh mortar bulk density test, and the mean value and standard deviation of the two samples was reported. For each mortar mix, three hardened mortar specimens were used for hardened mortar bulk density test, and the mean value and standard deviation of the three specimens was reported.
Compressive strength test was conducted on the 5 cm test cubes according to Chinese standard JGJ/T70-2009 [29], after 60 and 180 days of curing by a WDW-20 electronic universal testing machine (Zhongluchang Manufacturing Co., Ltd, China) at 0.2 kN/s loading rate. For each mortar mix, three mortar specimens were used for the test, and the mean value and standard deviation of the three specimens was reported.
Water absorption test was conducted on 5 cm test cubes after 60 and 180 days of curing according to Chinese standard JGJ/T70-2009 [29]. Before test, the specimen cubes were dried in oven at 80 °C to constant mass to fully evaporate the water present in mortars. When taken out and cooled to room temperature, with the top surface facing down, they were placed in a flat-bottom container, supported by two steel rods (Φ 10 mm), and partially immersed in 35 mm-deep distilled water under ~ 25 °C and ~ 95% RH. The weight of each specimen was measured by an electrical balance (accuracy 0.01 g) at 1, 3, 5, 10, 15, 20, 25, 30, 40, 50, 60 min, and 2, 3, 4, 5, 6, 24, 36 and 48 h after immersion. It was suggested by the standard that the water absorption (m%) should be calculated as W = (m 1 -m 0 )/m 0 , where m 0 and m 1 are masses before and after absorption, respectively. However, the authors decided to report water absorption values as water absorption (g m −2 ) vs time 1/2 (s 1/2 ) so that the data can better represent the permeability of mortars. Water absorption coefficient values were calculated according to standard BS EN 1925: 1999 [33]. The calculation principle was that the plotted specimen graphs can be approximated by two straight lines, and the coefficient of water absorption by capillarity can be represented by the slope of the regression line of the first part (the increasing part) of the graph. For each mortar mix, three mortar specimens were used for the test, and the mean value and standard deviation of water absorption and water absorption coefficient of the three specimens were reported.

Analyses
Fourier transform infrared spectroscopy (FTIR), X-ray diffraction analysis (XRD) and thermogravimetry and differential scanning calorimetry (TG-DSC) were performed on specimens after 60 and 180 days of curing. Before analysis, all specimens were vacuum-dried in oven for 48 h at 60 °C to evaporate excess water in mortar pores and capillaries.
TG-DSC was performed by means of a Mettler Toledo TGA/DSC1 instrument (Mettler Toledo, Switzerland). Samples were placed in aluminum crucibles, with temperature ranging from room temperature to 1000 °C under pure N 2 atmosphere, heating rate at 10 °C min −1 . Table 4 showed that lime-tile dust mortars with sticky rice addition had higher water retention values than those without. Sticky rice porridge which is comprised of water and starch as gelling agent, could be regarded as an aquagel [34]. When homogeneously incorporated into mortar mixes, a large portion of the water in sticky rice porridge was physically bound and could not easily leakoff into a porous substrate [35].

Water retention, setting time and bulk density
Sticky rice porridge increased both initial and final setting time of the studied lime-tile dust mortar. The prolongation effect of saccharide additives on the setting time of cement mortars was well-known, which might be linked to adsorption of admixtures onto surfaces of the hydrated and/or anhydrous phases [36,37] through the complexation of saccharides and their alkaline degradation products [38] with metal ions such as Ca 2+ [39,40], forming semipermeable coating and delaying the formation of hydration products [41].
Addition of sticky rice porridge into lime-tile dust mortars decreased both fresh and hardened bulk density values under both dry and humid curing. As mentioned earlier, sticky rice porridge can be considered as an aquagel. When homogeneously incorporated into fresh mortar mixes, it became encased within the mortar. As mortar dried, the aquagel shrank, and void spaces were developed. Moisture from within the aquagel particles slowly diffused to the surface of the mortar and evaporated, further reducing mortar density. [34,42]

Water absorption
After curing for 60 days (Fig. 3a), at 48 h after immersion, specimens with sticky rice cured under both dry and humid conditions achieved similarly high absorption values (averaged at 24,080-24,840 g m −2 , Table 5). Meanwhile, specimens without sticky rice had the lowest water absorption values (averaged 22,000 g m −2 for L-TD-W-50% and 20,400 g m −2 for L-TD-W-95%), demonstrating increased water absorption with sticky rice addition, which was also reported in Zhao et al. [9]. Such increase of water absorption should be due to more pore volume in mortars with sticky rice addition, which was resulted from the void spaces formed after sticky rice porridge dried and shrank (discussed in "Water absorption" Section).
After 180 days of curing, the absorption values of specimens cured under 50% RH basically didn't change (Table 5), yet the values of those cured under 95% RH declined ( Fig. 3b and Table 5 (Table 5). Reduced water absorption and water absorption coefficient values could indicate diminished pore size, volume and/or capillary interconnections, which should be due to the continuous formation of hydration products under humid curing condition, taking up more space in mortar pores, making mortar denser and more compact. Figure 4 showed that under 50% RH curing, mortars without sticky rice addition achieved the highest compressive strength values after 60 and 180 days of curing (0.97 ± 0.08 and 2.08 ± 0.18 MPa), while L-TD-HSR-50% (0.41 ± 0.03 and 0.79 ± 0.02 MPa) and L-TD-CSR-50% (0.44 ± 0.01 and 0.81 ± 0.08 MPa) had similarly lower values. Such results are in accord with the lower 28 d hardened mortar density values of L-TD-HSR-50% and L-TD-CSR-50% (Table 4), since lighter material would result in weaker structures, thereby decreasing mortar strengths. Additionally, Yang et al. [43] also reported decreasing compressive strength in hydraulic lime mortars (lime-slag, lime-fly ash, lime-silicon fume) with increasing sticky rice content in mortars, which could indicate an inhibiting effect of polysaccharides and proteins on the pozzolanic reaction of such specimens.

Compressive strength
The 60 and 180 d compressive strength values of L-TD-W-95% (4.27 ± 0.38 and 6.49 ± 0.41 MPa) were much higher than those of L-TD-W-50% (0.97 ± 0.08 and 2.08 ± 0.18 MPa), due to the enhanced pozzolanic reactions under higher relative humidity [44,45]. The strength values of L-TD-HSR-95% (0.80 ± 0.06 and 1.85 ± 0.15 MPa) and L-TD-CSR-95% (0.62 ± 0.03 and   1.87 ± 0.07 MPa) were significantly lower than those of L-TD-W-95%, demonstrating that the strong retarding effect of sticky rice addition on mortar hydration could last for months, as also shown in [46]. Meanwhile, all specimens cured under 95% RH for 180 days had higher compressive strength values than those cured for 60 days, corroborating the water absorption test results ("Water absorption" Section) that water absorption and water absorption coefficient values were reduced after 180 days of curing.

FTIR
According to Fig. 5a and Table 6, for all specimens cured under 50% RH, the relative intensity of the  [47,48,61,62]. In L-TD-W-95% and L-TD-HSR/CSR-95%, although the bands at ~ 1421 and ~ 875 cm −1 are present in their spectra, the intensity of ~ 710 cm −1 band remained relatively low throughout 180 days of curing, which should be due to less calcite formation under high relative humidity. For specimens with sticky rice cured under dry condition (L-HSR/ CSR-50%, L-TD-HSR/CSR-50%), the intensity of ~ 710 cm −1 band was also low (Fig. 5b), which indicates the absence of significant amounts of crystalline calcium carbonate [63]. Hydroxyl and carboxyl functional groups present in sticky rice and its alkaline degradation byproducts could adsorb onto positively charged Ca(OH) 2 faces [55], inhibiting nucleation and limiting the crystallization of calcium carbonate.
All specimens cured under 95% RH showed peaks at 2θ≈ 11.7° and 23.4°, indicating presence of tetracalcium aluminate hydrate (C 4 AH 13 [71]) and monocarboaluminate (C 3 ACcH 11 [72]. These products were formed as a result of reactions between reactive aluminates in metakaolin (or as in this case from pozzolanic tile dust) and CO 3 2− ions in mortar mix [73]), which were both reported to be capable of providing strength in concrete [74,75], and concords with the results of compressive strength test ("Compressive strength" Section) that specimens cured under 95% RH achieved higher compressive strength values than their counterparts cured under 50% RH.
Weight loss within 350-450 °C should be attributed to dehydration of Ca(OH) 2 [81]. For specimens containing sticky rice, a part of the weight loss could still be caused by the depolymerization of polysaccharides. According to Rodriguez-Navarro et al. [55], within 100-395 °C, an additional weight loss could be present, corresponding to thermal decomposition of organic additives adsorbed on portlandite crystals and/or distributed among such crystals. Additionally, DTG and DSC curves (Fig. 8) of specimens containing sticky rice (~ 400-422 °C) generally had lower decomposition temperatures than the specimens without (~ 411-430 °C), which might confirm the interaction/reaction between lime and sticky rice.
Within 550-750 °C, weight losses of the raw materials (hydrated lime (1.40%), tile dust (0.77%), and sticky rice powder (2.01%)) can be attributed to the carbonaceous matters in these materials (Table 7). Under 50% RH curing for 60 days, the weight loss values of L-HSR-50% (15.48%) and L-CSR-50% (22.38%) were less than L-W-50% (27.45%). After 180 days of curing, however, the weight loss differences among L-W-50%, L-HSR-50% and L-CSR-50% (31.53, 29.68, 30.62%, respectively) became little. It seemed that the addition of sticky rice porridge, especially at hot temperatures, could result in slower initial carbonation rate on lime mortars; then the specimen gradually carbonated throughout 6 months of curing, and finally achieved similar amounts of carbonates as in specimen L-W-50%. The slower initial carbonation rate could be related to the decreased ion diffusion activities due to the confinement of lime particles inside sticky rice aquagel network. Additionally, according to Table 7, lime-tile dust mortar specimens cured under 50% RH (L-TD-W-50%, L-TD-HSR/CSR-50%) had more weight loss within 550-750 °C than those cured under 95% RH (L-TD-W-95%, L-TD-HSR/CSR-95%), which shows that more carbonates were formed in the specimens cured under 50% RH. Air lime mortars gain strength through carbonation, as calcium carbonate forms and fills mortar pores and capillaries [82]. However, for the specimens cured under 95% RH, pozzolanic reactions between lime and tile dust were facilitated (which makes the mortar hydraulic), and such hydraulic lime mortars could usually result in higher mechanical strengths [20] than air lime mortars, as indicated by the compressive strength test results in the present study (Fig. 4).

Conclusions
The present study reported some preliminary effects of lime-tile dust mortars with sticky rice addition. Results showed that mortars with sticky rice porridge addition had higher water retention values, which could prevent excessive water loss and induce better mortar properties when applied on substrates. Addition of sticky rice porridge into mortars could decrease fresh and hardened mortar bulk density values under both dry and humid curing, thereby increasing their water absorption values. However, by curing mortars under high relative humidity (95% RH), their water absorption coefficient values could be reduced, especially over a long period of curing time (180 days). The addition of sticky rice into hydraulic lime mortars could slow down carbonation as well as hydration process (especially within the first 60 days of curing), which not only prolonged mortar setting time, but also impaired mortar strength development. Apart from calcite, aragonite was also formed as carbonation product.