Ceramics International xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Ceramics International
journal homepage: www.elsevier.com/locate/ceramint

Synthesis of colloidal nanosilica from waste glass powder as a low cost

precursor
⁎
Zohreh Asadi, Reza Norouzbeigi
School of Chemical, Petroleum and Gas Engineering, Iran University of Science and Technology, Narmak, Tehran, Iran

A R T I C LE I N FO A B S T R A C T

Keywords: Waste glass powder was used as a low cost precursor for production of colloidal nanosilica for the first time. The
Colloidal nanosilica process includes production of wet silica gel and thermal peptization of the wet gel. Purification of the glass
Peptization powder and wet gel production were initiated by acid washing. The obtained powder was reacted with sodium
Waste glass powder hydroxide to produce wet silica gel. Type of the applied acid was examined in one factor at a time route.
Taguchi design
Temperature of the alkaline step and concentrations of the applied acid and base were investigated using
Taguchi design of experiments. After finding the best combination of the investigated factor levels in production
of the wet gel, time of the stabilization in thermal peptization was studied. Characterizations of the wet gel and
colloidal silica were performed by XRF, DLS, FESEM, TEM, FTIR and N2 sorption evaluation. Accordingly pure
and stable colloidal nanosilica (98.50%) with average particle size of 21.9 nm was produced from the glass
powder successfully. Specific surface area of the dried porous optimum sample was 83.63 m2/g.

1. Introduction was proposed by Ahlberg and Simpson in order to obtain a more stable
colloid [15]. However, Rhee et al. [16] studied effects of some im-
Colloidal silica is dispersion of silica particles which are between portant experimental parameters such as concentration, pH of the gel-
1 nm and 1 µm in size showing specific deviations from true solutions ling and washing solutions, aging and duration of the peptization step.
[1]. Colloidal silica has growing and versatile applications. It can act as It has been reported that lower temperature differences between two
a mineral binder in cement and concrete [2], polishing slurry in pro- steps of the thermal peptization may lead to fabrication of colloids with
duction of mirrors and lenses [3], coating in textile and paints [4–6], lower average particle sizes [17].
carrier in drug delivery systems and precursor in production of catalysts Some minerals such as olivine, vermiculite and chrysotile have been
[7]. applied for production of nanosilica as inexpensive available Si sources
There are several processes for production of colloidal silica such as [18–21]. However, almost all of the mineral based synthetized nano-
ion exchange, acid-neutralization, peptization [8], direct oxidation, silica products reported in the literature are powder. In our previous
milling and dispersion [9], dialysis and electrodialysis [10]. Although work, perlite was used to obtain colloidal nanosilica [17]. Particle size
ion exchange is the most common process, but it has some dis- of the produced nanosilica was mostly about 8–74 nm. In this study,
advantages such as recycling of ion exchange resins and residual Na waste glass powder has been selected to produce colloidal nanosilica
which causes aggregation and lower stability of the colloid [9]. with smaller particle size for the first time. Glass powder is a chemically
Peptization is a simple, practical and environmental friendly neutral and abundant low cost source of silica. Reusing of waste glass
method for production of colloidal silica. Two distinct methods have has various advantages such as saving energy and materials, reduction
been proposed for peptization of a gel: Thermal and ultrasonic-assisted of the production costs, decrease of landfill space demands and so on
routs [11,12]. Thermal peptization has some benefits. Trail [13] con- [22,23]. Effective factors have been optimized using Taguchi design of
cluded that if there is no live steam, a tight 90% filled autoclave is more experiments. The utilized process had two steps: wet gel production and
effective for peptization. White and coworkers [14] investigated the thermal peptization. Studied factors of the thermal peptization were
impacts of temperature and pressure. According to their report, suitable type and concentration of the acid solution, concentration and tem-
temperature and pressure ranges for thermal silica gel peptization are perature of the alkaline solution and finally stabilization time which
80–200 °C and 7–225 psi, respectively. A two- step thermal peptization was considered as the second step of the thermal peptization. Wet pure

⁎
Corresponding author.
E-mail address: [email protected] (R. Norouzbeigi).

https://doi.org/10.1016/j.ceramint.2018.09.050
Received 5 August 2018; Received in revised form 5 September 2018; Accepted 6 September 2018
0272-8842/ © 2018 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Please cite this article as: Asadi, Z., Ceramics International, https://doi.org/10.1016/j.ceramint.2018.09.050
Z. Asadi, R. Norouzbeigi Ceramics International xxx (xxxx) xxx–xxx

Table 1 silica gel and colloidal nanosilica with smaller particle sizes were se-
Wet gel production parameters and their levels. lected as targets of the experiments.
Symbol Wet gel production parameter Unit Level 1 Level 2
2. Materials and method
A Hydrochloric acid concentration M 2 4
B Alkaline step temperature °C 70 90
C Sodium hydroxide concentration M 2.5 4.5
2.1. Materials

A piece of a flat waste domestic glass was selected as the Si source

Table 2 and crushed. It was washed by distillated water and ethanol (Merck Co.,
Taguchi experimental design (factors-levels combination). 99%) and put in a ball mill. The mass proportion of ceramic ball to
crushed glass was 3:2 and the time of milling was 15 h. Other materials
Run Factor Response
including sodium hydroxide (97%), sulfuric acid (Analytical grade,
A B C Yi SiO2 content (%) 95–97%), hydrochloric acid (Analytical grade, 37%) and nitric acid
(Analytical grade, 65%) were provided from Merck Company
1 1 1 1 Y1 89.72
(Germany) and were used without any purification.
2 1 2 2 Y2 88.81
3 2 1 2 Y3 93.69
4 2 2 1 Y4 98.50 2.2. Preparation of colloidal nanosilica

2.2.1. Wet gel production

Table 3 The procedure of the wet gel production was similar to the previous
Chemical composition of the glass powder.
work [17]. Wet silica gel fabrication process includes following steps:
component wt% reactions in the acidic and alkaline media and gel preparation. First, 5 g
of the obtained glass powder were mixed with 100 ml of acid solution at
SiO2 72.80
Na2O 12.86
80 °C for 5 h in a reflux system. Settled materials were separated by
CaO 7.62 filtration using a Buchner funnel and then dried at 110 °C for 12 h.
Al2O3 2.57 Dried residues of the acidic step were reacted with an aqueous solution
MgO 3.23 of sodium hydroxide for 90 min. The impurities were settled in the al-
K2O 0.68
kaline step and sodium silicate was separated by filtration again. Fi-
Fe2O3 0.11
TiO2 0.079 nally, wet silica gel was obtained by addition of sulfuric acid (2.5 M) to
P2O3 0.006 the sodium silicate solution. Resulted silica gel was washed by sulfuric
MnO 0.005 acid solution and distilled water (150 ml) to remove sodium sulfate.
S2O3 0.005 In order to study importance of the acid washing, three experiments
were conducted using sulfuric, hydrochloric and nitric acid solutions
(4 M). Other investigated gel production influential factors were the
alkaline step temperature and concentrations of the sodium hydroxide
and acid solutions. Taguchi experimental design was utilized in order to
decrease the number of the experiments and evaluate the effects of the
factors. Experiments of the wet gel production were arranged by
Taguchi L4 orthogonal array. Table 1 indicates the controlling para-
meters and their levels.
Eqs. 1, 2 and 3 present related reactions of acidic, alkaline and gel
production steps, respectively. In the first equation, M can be con-
sidered as metallic impurities presenting in the initial waste glass
powder (Al, Mg, Na, and Fe) [17].

Glass powder + HCl → M(Cl)n + SiO2↓ (1)

SiO2 + 2NaOH → Na2SiO3 + H2O (2)

Na2SiO3 + H2SO4 → SiO2 + Na2SO4 + H2O (3)

Fig. 1. FESEM picture of glass powder particles (5KX magnification). Designed experimental runs were conducted according to Table 2.
Statistical analyses were performed to optimize the synthesis condition.
Consequently, the experiments were continued with thermal peptiza-
Table 4
tion of the obtained wet gel.
XRF elemental analysis for experiments with various acids.
Run SiO2 (%) Al2O3 (%) Na2O (%) SO3 (%) MgO (%) 2.2.2. Wet gel peptization
1 98.175 0.598 0.145 0.029 0.053
The wet gel was dispersed in a sodium hydroxide solution (0.1 M)
2 98.475 1.078 0.410 0.037 0 under vigorous stirring. Two distinct steps were performed. Duration of
3 97.773 0.722 1.463 0.042 0 the first thermal peptization step was 10 h and the temperature was
remained at 120 °C. The temperature was fixed at 160 °C and the
duration was set at four levels (1–4 h) in the second step.

2.3. Sample characterization

X-ray fluorescence (XRF) analysis was applied to determine the

chemical composition of the initial (raw) glass powder and dried

2
Z. Asadi, R. Norouzbeigi Ceramics International xxx (xxxx) xxx–xxx

Table 5
XRF elements analysis.
Experiment No. Al2O3(%) Na2O (%) SO3(%) MgO (%) K2O (%) Fe2O3(%)

1 1.7 7.54 0.74 0.16 0.08 0.06

2 1.51 8.7 0.74 0.12 0.06 0.06
3 1.53 4.44 0.19 0.08 0.07 0
4 0.63 0.79 0 0.04 0.04 0

Fig. 2. Wet gel production factors main effect plot.

Table 6 less than −30 mV [24]. Dynamic light scattering (DLS) and zeta po-
ANOVA table for Taguchi L4 designed experiments. tential analyses were performed using a Malvern light scattering unit
Source of Sum of DOF Mean of Contribution (ZEN 3600) with a 633 nm He–Ne laser.
variance square square percentage (%) For the other characterizations such as field emission scanning
electron microscopy (FESEM), Fourier transform infrared spectroscopy
A 46.649 1 46.649 79.56
(FTIR) and gas physisorption the samples were dried at 80 °C for 10 h.
B 3.803 1 3.803 6.49
C 8.180 1 8.180 13.95
FESEM (Zeiss, sigma) studies were operated at a 30 kV used for mor-
Error 0.000 0 – – phology assessment. Fourier Transform Infrared spectroscopy evalua-
Total 58.632 3 – – tion was carried out in the wavenumbers ranged from 450 to
4000 cm−1 (Perkin Elmer Spectrum, RX1) applied to analyze the sur-
face functional groups. Besides, the specific surface area of the opti-
mized sample was measured by N2 sorption method and the
Brunauer–Emmett–Teller (BET) model (NOVA® Station B). Finally, for
TEM analysis (Zeiss, EM10C, 100 kV), a drop of the optimized sample
was dried on a substrate.

3. Results and discussion

3.1. SiO2 preparation (wet gel)

XRF result and FESEM image of the initial glass powder are pre-
sented in Table 3 and Fig. 1, respectively. Elemental compositions of
Fig. 3. Wet gel production yield. the samples prepared under various acidic media are listed in Table 4.
Samples 1, 2 and 3 were obtained by hydrochloric, sulfuric and nitric
acids. Results show that using hydrochloric acid leads to a higher purity
resulted silica gels (Philips, PW 1800). Particle size distribution and (SiO2 content (%)) of the product.
zeta potential can be considered as two crucial features of the final H+ cations must diffuse through the solid phase to solve it in an
product. Average particle size and the polydisparity index (PDI) re- acidic solution [25]. As a result, more powerful acids like HCl can re-
present the quality and stability of colloids. PDI indicates the wideness lease more hydronium ions and subsequently remove the impurities. To
of the particle size distribution. As a result, lower PDI value shows improve the SiO2 content of the product, besides the temperature and
uniform particulate dispersion and is desired for high quality products. concentration of the alkaline step, the acidic concentration were in-
Another crucial property of practical colloids is stability. The measured vestigated according to the Taguchi experimental design. The Optimum
zeta potential values of the stable colloids are more than + 30 mV or experimental conditions and impact of the studied factors can be

3
Z. Asadi, R. Norouzbeigi Ceramics International xxx (xxxx) xxx–xxx

Y3 + Y4 93.69% + 98.5%
A2̅ = = = 96.095%
2 2
EfA = A2̅ − A1̅ = 96.095% − 89.265% = 6.83% (4)

The effect of parameters B and C were obtained 1.95% and

−2.86%, respectively. Results show that the acidic solution con-
centration is the most important factor hence the impurities removal is
the main result of the initial reactions done in the acidic medium.
Similar outcome is reported in the literature [27]. Temperature of the
alkaline step and concentration of the sodium hydroxide solution have
fewer impacts.
Fig. 2 shows factors main effect plot. The optimum experimental
conditions can be inferred from the levels with higher mean values.
According to Fig. 2, the optimum condition can be introduces as com-
bination of A2, B2 and C1. Accordingly, the optimized levels are acid
Fig. 4. Particle size distribution of colloidal silica samples in term of number.
solution concentration of 4 M, alkaline reaction temperature at 90 °C
and sodium hydroxide solution concentration of 2.5 M. As it is indicated
in Table 2, the most purified sample was produced in the fourth ex-
periment (98.50%).
Analysis of variance (ANOVA) reveals the contribution percentage
of each factor (Table 6). Data given in Table 6 authenticate that the acid
solution concentration can be considered as the most important factor.
ANOVA results confirm the factor effect analysis.
Optimized combination of factors levels is in agreement with our
previous research in which perlite was selected as Si source.
Accordingly, higher acid solution concentration and alkaline tempera-
ture cause a higher rate of separation. Reports show that silica and
alumina have tendency to form polysialate or zeolite (aluminosilicates)
in the concentrated alkaline media. High sodium hydroxide con-
centration leads to sedimentation of silica species including impurities
[17].
Process yields related to silica purification for experiments were
computed by Eq. (5).
Fig. 5. Polydispersity index of colloidal samples. mi
PYi (%) = ×100
m0 (5)

where PYi and mi are process yield and dried gel mass related to the ith
experiment, respectively. Indices i and 0 refer to the run numbers
(Table 2) and the raw waste glass. The process yield is defined by di-
vision of the dried silica gel and the initial raw glass waste mass (5 g).
According to Fig. 3, the production yield (%) of the first experiment
is more than 50%. For synthesis of this sample, temperature and con-
centration of the alkaline step were in the least level, so the sedi-
mentation of silica along with impurities was less than the other ex-
periments [28].

3.2. Particle size distribution, PDI and FTIR

For investigation of the average particle size and particle size dis-
tribution, DLS analysis was carried out. The obtained results are illu-
strated in Fig. 4. The average particle sizes of the four prepared samples
Fig. 6. FTIR spectrum of dried obtained colloidal nano-silica.
were 23.3 nm, 21.9 nm, 35.9 nm and 37.9 nm respectively. Hence,
sample 2 made under 2 h stabilization duration had the smallest par-
estimated by factor main effect measurements (Ef) and analysis of ticle size.
variance (ANOVA) [26]. The weight percentage of silica in the dried gel PDI values of the obtained colloids are shown in Fig. 5. PDI of the
was considered as response of the experiments. XRF results and re- second experiment is less than the others. Particle size distribution
sponses are presented in Tables 2, 5, respectively. would be optimized when thermal stabilization is less than 2 h [12].
For calculation of Ef, the mean values of responses should be de- Smaller particles will sediment on the larger ones when higher silica sol
termined for each level. Accordingly, the factor effect can be computed temperature and alkaline conditions are applied (Ostwald ripening)
by subtracting the lowest mean value from the highest one [27]. Effect [29].
of the factor A (EfA) was calculated according to Eq. (4). In the second step of thermal peptization, functional groups existing
on the surface of nanosilica particles were eliminated, so the produced
Y + Y2 89.72% + 88.81%
A1̅ = 1 = = 89.265% colloid was more stable. On the other hand, Ostwald ripening was
2 2
happened in this step. As a result, the optimized time which balanced
two proposed phenomena was 2 h. Zeta potential of colloidal particles
can be considered a reliable criterion for stability assessment. The zeta

4
Z. Asadi, R. Norouzbeigi Ceramics International xxx (xxxx) xxx–xxx

The intensity of peaks at 3518 and 1128 cm−1 decreased by in-

creasing the duration of second step of thermal peptization. As a result,
by increasing the duration of the stabilization step, amounts of the si-
lanol groups and adsorbed water molecules are decreased.

3.3. Morphology and surface area assessments

Sample 2 was synthesized under the optimum conditions. FESEM

and N2 sorption investigations were performed for sample 2. Figs. 7, 8,
show FESEM micrographs of the initial glass powder and the dried
colloidal nanosilica. According to the figures, particles are spherical
with a good uniformity in shape and size. FESEM image of sample 2 is
in a good agreement with the average particle size obtained by DLS
analysis. Fig. 8 also shows the TEM micrographs of sample 2. According
to the figure, nano SiO2 particles are successfully achieved.
Fig. 9 shows the results of nitrogen sorption isotherm. This graph
Fig. 7. FESEM micrograph of dried colloidal nano-silica (50KX magnification) exhibits an IV type isotherm with a H1 type loop. The isotherm in-
from sample 2. dicates that dried silica particles are porous in texture and capillary
condensation has been occurred during the sorption [29]. The specific
surface area and mean pore diameter of sample 2 calculated from the
BET and BJH models, were 83.63 m2/g and 0.557 nm, respectively.
Similar N2 sorption isotherms are reported for nanosilica particles
prepared under similar routes or conditions [29,32].

4. Conclusion

Pure colloidal nanosilica was produced from waste glass powder

successfully. The optimum levels of acid concentration, alkaline solu-
tion concentration and temperature for production of wet silica gels,
were 4 M, 2.5 M and 90 °C, respectively. The optimized silica gel sample
had SiO2 content of ~98.50% and the average particle size was
21.9 nm. The optimum stabilization duration was obtained 2 h.
Statistical analysis showed that acid solution concentration was the
most effective factor to prepare a pure wet gel. Under the optimized
conditions, the obtained yield percentage was more than 50%.

Fig. 8. TEM micrograph of dried colloidal nanosilica (100KXmagnification) Acknowledgement

from sample 2.
The authors gratefully acknowledge financial support from the Iran
National Science Foundation, Iran (INSF, grant number: 91049554) and
research council of Iran University of Science and Technology.

References

[1] K.S. Birdi, Handbook of Surface and Colloid Chemistry, 2nd edition, CRC Press LLC,
Boca Raton, 2003.
[2] M. Aly, M.S.J. Hashmi, A.G. Olabi, M. Messeiry, E.F. Abadir, a.I. Hussain, Effect of
colloidal nano-silica on the mechanical and physical behavior of waste-glass cement
mortar, Mater. Des. 33 (2012) 127–135.
[3] M. Sivanandini, S. Dhami, B. Pabl, Chemical mechanical polishing by Colloidal si-
lica slurry, Int. J. Eng. Res. Appl. 3 (2013) 1337–1345.
[4] B. Mahltig, H. Böttcher, Modified silica sol coatings for water-repellent textiles, J.
Sol-Gel Sci. Technol. 27 (2003) 43–52.
[5] J.J. Yuan, S.X. Zhou, B. You, L.M. Wu, Organic pigment particles coated with col-
Fig. 9. N2 adsorption-desorption isotherm for dried colloidal nanosilica of loidal nano-silica particles via layer-by-layer assembly, Chem. Mater. 17 (2005)
sample 2 at 80 °C. 3587–3594.
[6] T. Mizutani, K. Arai, M. Miyamoto, Y. Kimura, Application of silica-containing
nano-composite emulsion to wall paint: a new environmentally safe paint of high
potential of sample 2 prepared under the optimum condition was performance, Prog. Org. Coat. 55 (2006) 276–283.
−43.6 mv. The measured zeta potential less than −30 mV confirms the [7] T.K. Barik, B. Sahu, V. Swain, Nanosilica-from medicine to pest control, Parasitol.
Res. 103 (2008) 253–258.
stability of the obtained colloid [12]. [8] H.E. Bergna, W.O. Roberts, Colloidal Silica: Fundamentals and Applications,
FTIR spectroscopy analysis was performed to evaluate the surface Surfactant Science Series, CRC Press, New York, 2005.
functional groups of the product. Fig. 6 illustrates the related spectrum. [9] M.J. Comstock, The Colloid Chemistry of Silica, Maple Press, Washington, 1994.
[10] H.M. Lim, J. Lee, J.H. Jeong, S.G. Oh, S.H. Lee, Comparative study of various
The peaks observed at wavenumber of 490, 820 and 1128 cm−1 can be preparation methods of colloidal silica, Engineering 2 (2010) 998–1005.
assigned to asymmetric Si-O bond stretching, symmetric Si-OH bond [11] R.K. Iler, The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface
stretching and Si-O-Si bond bending vibration, respectively [30]. Also, Properties, and Biochemistry, John Wiley & Sons Inc, New York, 1979.
[12] V. Jafari, A. Allahverdi, M. Vafaei, Ultrasound-assisted synthesis of colloidal na-
the peaks located at 1650 and 3518 cm−1 can be related to H-O-H nosilica from silica fume: effect of sonication time on the properties of product, Adv.
stretching and bending modes of adsorbed water [31]. The peak pre- Powder Technol. 25 (2014) 1571–1577.
sented at 3650 cm−1 is related to Mg-O impurities [20]. [13] H.S. Trail, US Pat.2572578, 1948.

5
Z. Asadi, R. Norouzbeigi Ceramics International xxx (xxxx) xxx–xxx

[14] J.F. White, US Pat.2 375 738, 1945. [24] J. Jiang, G. Oberdorster, P. Biswas, Characterization of size, surface charge, and
[15] J.E. Ahlberg, E.A. Simpson, US Pat.2900348, 1959. agglomeration state of nanoparticle dispersions for toxicological studies, J. Nano
[16] Y. Lee, Preparation of colloidal silica using peptization method, Colloids Surf. 173 Part. Res. 11 (2009) 77–89.
(2000) 109–116. [25] S. Arabia, Extraction of alumina from local clays by hydrochloric acid process, Eng.
[17] Z. Asadi, R. Norouzbeigi, Optimization of colloidal nanosilica production from ex- Sci. 20 (2009) 29–41.
panded perlite using Taguchi design of experiments, Ceram. Int. 43 (2017) [26] D.C. Montgomery, Design and Analysis of Experiment, 5th ed., John Wily and sons,
11318–11323. New York, 2004.
[18] A. Lazaro, H.J.H. Brouwers, G. Quercia, J.W. Geus, The properties of amorphous [27] R.K. Roy, Design of Experiments Using the Taguchi Approach: 16 Steps to Product
nano-silica synthesized by the dissolution of olivine, Chem. Eng. J. 211–212 (2012) and Process Improvement, John Wiley & Sons Inc, New York, 2001.
112–121. [28] P.L. Antonnucci, M.L. Crisafulli, N. Giordano, N. Burriesci, Zeolitization of perlite,
[19] M. Zhao, Z. Tang, P. Liu, Removal of methylene blue from aqueous solution with Mater. Let. 3 (1985) 302–307.
silica nano-sheets derived from vermiculite, J. Hazard. Mater. 158 (2008) 43–51. [29] A. Lazaro, M.C. van de Griend, H.J.H. Brouwers, J.W. Geus, The influence of process
[20] K. Liu, Q. Feng, Y. Yang, G. Zhang, L. Ou, Y. Lu, Preparation and characterization of conditions and Ostwald ripening on the specific surface area of olivine nano-silica,
amorphous silica nanowires from natural chrysotile, J. Non Cryst. Solids 353 (2007) Microporous Mesoporous Mater. 181 (2013) 254–261.
1534–1539. [30] E. Rafiee, S. Shahebrahimi, Nano silica with high surface area from rice husk as a
[21] L. Wang, A. Lu, C. Wang, X. Zheng, D. Zhao, R. Liu, Nano-fibriform production of support for 12-tungstophosphoric acid: n efficient nano catalyst in some organic
silica from natural chrysotile, J. Colloid Interface Sci. 295 (2006) 436–439. reactions, Chin. J. Catal. 33 (2012) 1326–1333.
[22] C. Shi, K. Zheng, A review on the use of waste glasses in the production of cement [31] M. Zhao, Z. Tang, P. Liu, Removal of methylene blue from aqueous solution with
and concrete, Resour. Conserv. Recycl. 52 (2007) 234–247. silica nano-sheets derived from vermiculite, J. Hazard. Mater. 158 (2008) 43–51.
[23] A. Shayan, A. Xu, Value-added utilization of waste glass in concrete, Cem. Concr. [32] P. Lu, Y. Lo Hsieh, Highly pure amorphous silica nano-disks from rice straw, Powder
Res. 34 (2004) 81–89. Technol. 225 (2012) 149–155.