Optimization of nanostructured/nano sized rice husk ash preparation

The objective of the study is developing a procedure for production and characterization of rice husk ash (RHA). The effects of rice husk (RH) amount, burning/cooling conditions combined with stirring on producing of RHA with amorphous silica, highest SiO2, lowest loss on ignition (LOI), uniform particle shape distribution and nano structured size have been studied. It is concluded that the best amount is 20 g RH in 125 ml evaporating dish Porcelain with burning for 2 h at temperature 700 °C combined with cooling three times during burning to produce RHA with amorphous silica, SiO2 90.78% and LOI 1.73%. On the other hand, cooling and stirring times affect the variation of nano structured size and particle shape distribution. However, no crystalline phases were found in RHA in all cases. Results proved that the Attritor ball mill was more suitable than vibration disk mill for pulverizing nano structured RHA with 50% of particle size (D50) lower than 45 m and 99 % of particle size (D99) lower than 144 m to nanosized RHA with D50 lower than 36 nm and D99 lower than 57 nm by grinding time 8.16 min to every 1 g RHA without changes in morphousity of silica.


Introduction
According to food and agriculture organization of the United Nations (FAO), more than 769 Mt of rice paddy are annually produced in the world (around 510 Mt, milled basis) in 2018 (1). Rice is considered as one of the main sources of food for people. Disposal of agricultural residue RH by useful way is aimed as it accounts around 20 % by weight of rice (2). Variation in chemical composition of RH is found from one sample to another due to the differences in the climate and geographical conditions (3). But, generally, RH is composed of both organic and inorganic matter. The organic element analysis shows that the average organic composition of rice husk was 39.8-41.1 wt.% carbon, 0.5-0.6 wt.% oxygen, 5.7-6.1 wt.% hydrogen and 37. 4-36.6 wt.% nitrogen while silica is the major component of inorganic minerals (4). However, uncontrolled combustion of RH produces RHA with low reactivity (5). The preparation of RHA should be optimized to be used as a partial replacement of Portland cement (6). The partial replacements of Portland cement depend on pozzolanic activity which depends on parameters such as content of amorphous silica, specific surface area of particles and particle size distribution (7). According to (8)(9)(10)(11), production of amorphous silica with some carbon and metallic impurities obtained from burning of RH depends on rate of heating, ultimate temperature, retention time, atmosphere cooling rate, and grinding time of the ash. Temperature ranging below 800 °C and higher than 500 °C was highly recommended. In this regard, incineration of rice husk at temperatures below 500 °C is not desirable and produces an amount of unburned carbon which can result in adverse effects on ash pozzolanic activities. On the other hand, burning at temperature higher than 800 °C produces crystalline silica instead of amorphous one, which is not desirable due to the decrease in the pozzolanic reactivity. However, in all cases quickly cooling of RHA by directly removing it from the oven is preferable to avoid crystallization. Also, grinding is recommended to increasing the pozzolanic reactivity by increasing the specific surface area and decreasing the particle size.
In this study, we focus on studying the effect of RH amount, cooling times with stirring and grinding process. Finally, the chemical and physical properties of nano structured RHA, before and after grinding, were determined, to obtain high reactive and optimal RHA suitable for replacements of Portland cement according to (ASTM C618) (12). The effect of both nanostructured and nanosized produced RHA was studied in terms of physical and mechanical properties of cement and mortar (13).

Materials and methods
The sample of RH used in the study was obtained from Kafr El Sheikh, Delta, Egypt. Its chemical composition is shown in Table 1. Selection of the best method to determine RHA particle size depends on using of Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK) then particle size distribution D50 and D99 at different grinding time was measured by using Laser diffraction apparatus combined with laser diffraction-SSA, laser particle size analyzer (Model BT-9300S by Bettersize) according to the International Standard ISO 13320 (17). The particle size of RHA before grinding was tested by using sieving test according to (ASTM C618) (12) to avoid irregular shapes and to confirm the fineness requirements for fly ash and natural pozzolana to be less than 34 % of the material retained on the 45 m sieve. Also, the particle size of final grinding time sample was confirmed by using TEM, (Model JEOL JEM-2100).

Chemical analysis
The chemical analysis in Table 2 shows that the SiO2 content of 100 g RH was affected slightly (only 3 %) by increasing burning time from 1 to 2 h. On the other hand, it increased 25.5 % by decreasing RH amount from 100 g to 30 g and increased 29.4 % in case of decreasing RH amount to 20 g.  Table 2 indicate that LOI of RHA is more affected by RH amount than burning time,

Effect of grinding
After grinding, LOI increased from 1.73 % to 6.73 %, as shown in Table 4, when particle size decreased from D50 45 m to D50 36 nm. Grinding for a long time will decrease SiO2 by 9.6 % due to the increase of LOI by 5 % after grinding which may be attributed to the extraction of internal carbon content remained in internal structure of RHA . The increase in Al2O3 may be attributed to contamination coming from the grinding mill components.

Particle size, distribution and shape of silica
The RHA which cooled three times was the best due to achieve the higher content of amorphous SiO2 by 1% increase and the lower amount of LOI than RHA cooled one time and approximately had the same analysis of RHA cooled six times. However, it has lower variation in SiO2 size and the particle shape distribution shows more uniformity than RHA cooled one and six times as shown in Fig. 2, so this sample has been selected to apply the grinding process.

Figure 2.Effect of cooling times on TEM micrograph
Zeta-sizer was used to measure the particle size in the case of grinding by using vibration disk mill. Fig. 3 shows the dramatic decreasing of RHA particle size to 433.9 nm measured by using 0.5 % concentration of RHA with 0.5 % dispersing agent isopropanol and sonication time for 30 min for the de-agglomeration and increasing the dispersion efficiency .
The coarse RHA particles with irregular shapes and lightweight before grinding were founded. So, particle size was difficult to track by the stable way using Zeta-sizer. RHA without grinding was 3667 nm which less than the value after grinding one hour, 4000 nm. During the sieve test, RHA percentage retained on mesh (90μ) was 4.8 % and 33.2 % retained on mesh (45μ) which confirmed that it is not suitable for RHA particle size to be tracked by Zeta-sizer apparatus due to its irregular shapes and light weight. The grinding of RHA for 6 and 8 h resulted around 200 nm as shown in the TEM micrograph Fig.4. While Zeta sizer resulted in around 450 nm for the same samples. For tracking the particle size, BT-9300S Laser Particle Size Analyzer was used. It is featured by auto water level measurement and auto dispersing that ensure the accuracy. However, the resulted particle size range agrees with TEM results.

Figure 3.Effect of grinding on average particle size
Black spots appeared in TEM micrograph (Fig. 4). This is attributed to using vibration disk mill. When the vessel temperature increased to 90 °C, the RHA particles adhered to each other and form layered sheets. In order to avoid the appearance of that black spot and prevent adhesion as well as increasing productivity, Attritor ball mill combined with cooling system have been applied. . Table 5 shows the particle size distribution D50 and D99 using BT-9300S Laser Particle Size Analyzer, it was being analyzed every 6 hours grinding time by using Attritor ball mill.
Particle size distribution D50 and D99 of RHA decreased dramatically to D50 7 μm and D99 19 μm with increasing grinding time to 30 h (Fig.5). After 30 h of grinding, the particle size started to increase until 42 h to D50 21 μm and D99 36 μm then decreased dramatically to D50 36 nm and D99 57 nm after 102 h grinding. Decreasing rate of particle size was high during the first period of grinding within 6 h also after 48 h and slows with time above 6 h until 30 h. This may be attributed to the elongated grinding time which gains the particles extra charges that reinforced the agglomeration

Surface analysis
The grinding of RHA from micro to nano scale using Blaine method caused increasing of surface area from 0.2997 m 2 /g to 2.0849 m 2 /g (Table 6). However, there are some limitations for using the Blaine measurement. It does not reflect the accurate RHA specific surface area (SSA) specially when it is larger than 0.5 m 2 /g, so it becomes unreliable (21). Another method was tested, laser diffraction-SSA. It was found that, as grinding time increase, the SSA of RH increased slightly started from 0.0864 m²/g, after 72 h, a sudden increase happened till surface area increased to 29.4811 m²/g with increasing the grinding time to 102 h, it appears clearly in Fig.5. These results agree with BET surface area (29.794 m²/g) as stated in Table 6 which indicates that the internal surface area disappeared due to grinding effect. For the RHA before grinding, the BET surface area is 87.517 m 2 /g. After grinding, it reaches 29.794 m 2 /g. The significant drop in surface area occurring after grinding due to particles agglomeration occurs during the grinding process. The enormous decrease in internal surface area was confirmed by cumulative volume of pores results. The pore volume decreased from 0.127 cm 3 /g to 0.042 cm 3 /g (Table 6). This decreasing in RHA internal surface area was due to the destroying the cellular structure existing in the RHA during grinding. This conversion of RHA into tiny particles filled the spaces and decreased the volume of pores.

Figure5.Effect of grinding time on specific surface area of RH
Most of the pore size lies in the mesopores region (2-50 Å) (Fig. 6). Using the N2 gas adsorption/desorption isotherms (Fig.7), the hysteresis can be classified as typical mesoporous of type IV isotherm according to IUPAC classification (22).

Mineralogy
No obvious change in mineralogy is noticed between XRD diffractograms of RHA before and after grinding (Fig. 8), both of them have a strong broad peak of RHA with high content of silica centered on range ≈ 22-23° (2θ) which is characteristics of amorphous SiO2 (23). Also, XRD results confirmed that avoiding long period of burning prevents crystallization, which is observed for temperatures exceeding 700°C. Increasing the retention time produces amorphous RHA. Also, the grinding process does not affect the mineralogy of RHA.
Figure8. Mineralogy state of RHA before and after grinding.

Conclusion
The optimum conditions to produce nano structured RHA with highest amorphous SiO2 and lowest LOI was burning for 2 h at temperature 700 °C combined with cooling 3 times and mass of rice husk (g)/volume of the evaporating dish (ml) is 0.16. These optimum conditions resulted 90.78 % of amorphous SiO2 and 1.73 % of LOI . Using Attritor ball mill is the best to convert micro RHA to nano RHA combined with uniform particle shape distribution. Tracking RHA particle size requires laser particle size analyzer featured by auto dispersing which ensures the accuracy and overcomes the RHA irregular shapes and light weight. Tracking RHA specific surface is more accurate by using BET method due to RHA cellular structure. Grinding RHA doesn't affect its mineralogy and adsorption/ desorption isothermals type, but it decreases the BET surface area.