4.1 Direct synthesis of MCM-41-like materials from optoelectronic waste powder
4.1.1 Motivation
Our prior study demonstrated the possibility of extracting silicate supernatant from optoelectronic waste powder which was obtained from the exhaust of CVD process in TFT-LCD plants. The alkaline fusion method was employed in our prior study and the obtained supernatant was further utilized for manufacturing silica materials. However, the disadvantages of alkaline fusion method for extracting the silicon from solid wastes are the long processing period (~24 h) and the low silicon recovery yield (20-33 %) [6,114]. Thus the industrial application of this fusion process would have certain limitations. An efficient and low energy method for the total recycling of optoelectronic waste powder into valuable mesoporous silica must be developed. In this part, a fast and low energy-consumed approach on the direct use of optoelectronic waste powder for complete recycling into silica resources is proposed. Detailed characterization and formation mechanism of silica materials via the direct usage of waste powder were investigated and discussed.
4.1.2 Synthesis
Mesoporous silica was synthesized via the direct utilization of waste powder as the silica source and cetyltrimethylammonium bromide (CTAB) was employed as the structure-directing template. The molar composition of the gel mixture was 1 SiO2 : 0.2 CTAB : xHF: 12 NH4OH: 120 H2O (x varies with molar ratio of HF/Si). In a typical procedure, 9.35 g of waste powder was firstly dissolved in 137 ml of DI water.
Subsequently, a given amount of hydrofluoric acid was added into the above solution.
Meanwhile, 7.28 g of CTAB was dissolved in 25 ml of DI water and then it was added dropwise into the above solution. Then, 32.06 g of ammonium hydroxide solution was slowly added to promote the hydrolytic condensation of silica-surfactant mixture. All the above procedures were performed under continuous stirring. The resulting gel
mixture was aged at room temperature for 8 h; the resultant solid was recovered by filtration, washed with DI water and dried in an oven at 110oC for 6 h. Finally, the organic template was removed by using a muffle furnace in air at 550oC for 6 h. The mesoporous MCM-41 synthesized by direct usage of waste powder as the silica source in the absence of hydrofluoric acid was denoted as MCM-41(DU), while the MCM-41 sample synthesized by direct usage of waste powder in the presence of hydrofluoric acid was denoted as MCM-41(DU)-F.
4.1.3 CO2 adsorption measurement
Tetraethylenepentamine (TEPA) was selected as the amine agent to enhance the adsorption performance of the mesoporous materials on capturing CO2 gas.
Mesoporous adsorbents were mixed with TEPA at a weight ratio of 1:1 in ethanol by the wet impregnation method. The mesoporous adsorbents were preheated at 120 oC for 1 h. And then TEPA was mixed with ethanol and the resulting solution was stirred for 30 min. After pretreatment, the mesoporous adsorbents were dispersed into the flask containing TEPA solution and the mixture was then refluxed at 80 oC for 2 h.
After cooling to room temperature, the obtained materials were dried at 60 oC.
To obtain the adsorption capacity and breakthrough curve of the adsorbents, CO2
adsorption experiment was carried out in a packed column with an internal diameter of 0.75 cm. The adsorption column was packed with 1.0 g of adsorbents (packing height = ~ 5 cm) and placed in a temperature-controlled oven. In a typical process, adsorbents were pretreated under a N2 flow of 0.1 L/min at 110 C for 1 h, and then cooled to 60 C. Subsequently, the gas flow was switched to 10% (v/v) CO2 gas stream (balanced with N2) under a flow rate of 0.1 L/min. The concentration of CO2
was continuously measured by a CO2 analyzer (Molecular Analytics AGM 4000 Gas Analyzer). The CO2 adsorption capacity (q, mg/g) at a certain time (t, time) was estimated as
are the influent and effluent CO2 concentrations (mg/L), respectively, which are expressed in terms of percent in volume (%). The adsorption capacity of zero gas (N
only) was deducted from the adsorption capacities of adsorbents.
Mixer
Adsorption column
Sampling Inlet Sampling Outlet
To CO2
Analyzer
MFC
Heating zone
N2
CO2
Figure 4. 1 CO2 adsorption via pack column system
4.1.4 Results and discussion
Figure 4.2(a) and Figure 4.2(b) respectively depicts the alkaline fusion process and the direct utilization process for the recovery of silica materials from the optoelectronic waste powder. For alkaline fusion process, the formation mechanism of silica materials from optoelectronic waste powder can be explained by considering the chemical reactions presented in Equations (1)-(3). In brief, optoelectronic waste powder consisting of (NH4)2SiF6 and SiO2 was decomposed into gaseous silicon tetrafluoride (SiF4) and ammonium fluoride (NH4F) solid residue when the temperature was above 220 oC [115]:
(NH4)2SiF6(s) 550oC
SiF4(g)↑ + 2NH4F(s) (1) Subsequently, the remaining NH4F which is thermally stable up to 850 oC would react with sodium hydroxide to produce gaseous ammonia and sodium fluoride solid residue:
NaOH(s) + NH4F(s) 550oC
NaF(s) + NH3(g)↑+ H2O(g)↑ (2) Meanwhile, the SiO2 presented in optoelectronic waste powder would also react with sodium hydroxide to produce sodium silicate:
2NaOH(s) + SiO2(s) 550oC
Na2SiO3(s) + H2O(g)↑ (3) The obtained sodium silicate was then utilized as the silica source to synthesize mesoporous MCM-41 in the presence of cationic surfactant of CTAB via hydrothermal treatment at 145 oC for 36 h as shown in Figure 4.2(a). Therefore, the gaseous pollutants of SiF4 and NH3 tend to be formed during the alkaline fusion process for waste recovery.
The optoelectronic waste powder is mainly composed of 85% of (NH4)2SiF6 and 15% of SiO2 and the total Si mass fraction (from (NH4)2SiF6 and SiO2) detected by ICP-MS analysis is 22.4% as obtained from our prior study. Silicon recovery yield was determined by the weight ratios of silicon contents in the raw waste powder and
Figure 4.2(a) that there is only 28% of silicon recovery yield for the MCM-41(AF) obtained via the alkaline fusion process. This is similar with the report by Chang and colleagues, who employed the alkaline fusion method for extracting silicon from coal fly ash and the silicon recovery yield was ca. 32% [108]. They stated that sufficient amount of sodium hydroxide and activation time are required for effective converting silica in the form of quartz and mullite phases into more soluble form of sodium silicate during the fusion process.
In the present study, the decomposition of (NH4)2SiF6 at 550 oC results in losing silicon species from the release of gaseous silicon tetrafluoride. It is reasonable that the silicon recovery yield would be significantly affected by the loss of silicon species from gaseous silicon tetrafluoride since the optoelectronic waste powder is mainly composed of 85% of (NH4)2SiF6 and 15% of SiO2. Therefore, even though the MCM-41(AF) with very high specific surface area can be manufactured, the alkaline fusion process might not be considered as an economic and energy effective approach due to its low silicon recovery yield, long processing time, gaseous pollution formation and high operational temperature.
Figure 4.2(b) reveals the direct utilization process proposed in this study. In a typical process, the optoelectronic waste powder was first treated with hydrofluoric acid. It is expected that the SiO2 presented in optoelectronic waste powder would react with hydrofluoric acid to produce hexafluorosilicic acid (H2SiF6) and the chemical reaction is presented as follows:
SiO2(s) + 6HF(aq) 2H+(aq)+ SiF6
2-(aq) + 2H2O(l) (4) Subsequently, the hydrolysis of (NH4)2SiF6 and SiF62- can be accelerated by the addition of ammonium hydroxide:
(NH4)2SiF6(aq) + 4NH4OH(aq) Si(OH)4(aq) + 6NH4F(aq) (5) 2H+(aq)+ SiF6
2-(aq) + 6NH4OH (aq) Si(OH)4(aq) + 6NH4F(aq) +2H2O(l) (6) Therefore, the overall reactions for the recovery of silica materials from optoelectronic waste powder in the presence of hydrofluoric acid and ammonium hydroxide can be summarized as follows:
(NH4)2SiF6 + SiO2 + 6HF + 10NH4OH 2Si(OH)4 + 12NH4F + 4H2O (7) Also it is well-known that the ionization of the Si(OH)4 produces anion silicate such as SiO(OH)3- and the condensation rate can be greatly enhanced using basic catalyst:
Si(OH)4(aq) + OH-(aq) SiO(OH)3
-(aq) + H2O(aq) (8) The ionization of the Si(OH)4 makes silanol more electrophilic and thus more susceptible to react with the cationic surfactants of CTAB by electrostatic force.
Finally, the cooperative assembly between the anionic silicate hydrolyzed from SiF6
2-and free charged micelles undergo extensive condensation 2-and polymerization, thus the mesoporous silica material can be rapidly regenerated at room temperature. The novel MCM-41(DU)-F material can be obtained from direct use and complete recycling of optoelectronic waste powder as the silica source. The process is simple, fast and low energy-consumed, which is more cost-effective when compared to the alkaline fusion process shown in Figure 4.2(a).
Figure 4.2 Schematic procedures for (a) conventional alkaline fusion process and (b) direct utilization process for the recovery of mesoporous silica from the optoelectronic waste powder.
Low angle powder X-ray diffraction patterns of MCM-41(DU), MCM-41(DU)-F and MCM-41(AF) are shown in Figure 4.3(a). The results reveal the presence of the hexagonal lattice of the MCM-41(AF) material prepared via alkaline fusion process, where two well-defined diffraction peaks of (100) and (110) located at 2θ of 2.5o and 4.1o were observed [19]. Similarly, MCM-41(DU) and MCM-41(DU)-F materials also exhibit two diffraction peaks of (100) and (110) located at 2θ of 1.7o and 3.5o, indicating that the silica materials with hexagonal mesostructure can be obtained by direct usage of optoelectronic waste powder as the silica source. It can be observed that MCM-41(DU)-F shows higher intensity of (100) reflection than that of MCM-41(DU), which suggests better pore arrangement of the MCM-41(DU)-F. On the other hand, there is a broad peak centering at 22o observed in Figure 4.3(b) for the MCM-41(DU) and MCM-41(DU)-F, respectively, which could be ascribed to the presence of amorphous SiO2 as extra-framework particles. Moreover, it is noticeable that the (100) and (110) reflection peaks of the MCM-41(DU) and MCM-41(DU)-F materials shifted to lower degrees as compared to those of MCM-41(AF). This implies that MCM-41(DU) and MCM-41(DU)-F materials exhibit larger d-spacing values than that of MCM-41(AF).
2 theta (degree)
Figure 4.3 (a) Low angle XRD patterns of MCM-41(DU), MCM-41(DU)-F as well as MCM-41(AF) and (b) wide angle XRD patterns of MCM-41(DU) and MCM-41(DU)-F samples.
The N2 adsorption-desorption isotherms of the waste powder, MCM-41(AF), MCM-41(DU) and MCM-41(DU)-F samples are plotted in Figure 4.4. Apparently, the optoelectronic waste powder revealed a typical type II isotherm of non-porous materials according to the IUPAC classification. On the other hand, MCM-41(AF), MCM-41(DU) and MCM-41(DU)-F samples clearly show capillary condensation steps with H1 hysteresis loops at a relative pressure of p/p0 =0.25~0.45, which is attributed to the textural mesoporosity and corresponded to capillary condensation of N2 molecules within the interparticle pores [34]. The less steep condensation for MCM-41(DU) sample indicates a less degree of uniform mesostructure as compared to that of MCM-41(AF) and MCM-41(DU)-F samples.
Furthermore, the capillary condensation steps of MCM-41(DU) and MCM-41(DU)-F samples tend to shift toward higher values of relative pressure, indicating that MCM-41(DU) and MCM-41(DU)-F samples possess larger pore diameter. This result is further confirmed by the BJH pore size distribution shown in Figure 4.5. It is clear to see that all samples show narrow pore size distribution, suggesting the uniform porosity of the obtained materials, while the sequence of pore diameter is in the order of MCM-41(DU)-F> MCM-41(DU)> MCM-41(AF). This is probably attributed to the formation of soluble NH4F in the process, which could be easily dissociated and the released fluorine ions are beneficial for the enlargement of the surfactant micelles.
The physical properties such as BET specific surface area, specific pore volume and average pore diameter derived from N2 adsorption-desorption measurements are summarized in Table 4.1. One can see that the specific surface area and the total pore volume of MCM-41(DU)-F are higher than those of MCM-41(DU), revealing that MCM-41(DU)-F possesses superior quality of the mesostructure and this result is in agreement with the XRD results shown in Figure 4.3. It is noticed that MCM-41(DU) and MCM-41(DU)-F exhibit similar wall thickness and they are thicker than that of MCM-41(AF), suggesting that MCM-41(DU) and MCM-41(DU)-F possess higher hydrothermal stability.
Relative pressure (P/Po)
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Amount of N2 adsorbed (cm3 g-1 , STP)
0
Figure 4.4 N2 adsorption-desorption isotherms of raw optoelectronic waste powder, MCM-41(DU), MCM-41(DU)-F and MCM-41(AF) samples.
Pore diameter (nm)
1 2 3 4 5 6 7 8
dV/dlogD, pore volume (cm3 g-1 , STP)
0
Figure 4.5 BJH pore size distribution of MCM-41(DU), MCM-41(DU)-F and MCM-41(AF) samples.
The chemical composition of the obtained MCM-41(DU)-F sample was investigated by EDS analysis and the result is presented in Figure 4.6(a). As seen from the spectrum that the obtained silica materials contain silicon and oxygen elements and there are no impurities observed at all, suggesting that highly purified siliceous materials can be obtained by the proposed method in this study. The textural structure was obtained from transmission electron microscopy (TEM) analysis as shown in Figure 4.6(b) and Figure 4.6(c). It can be observed that MCM-41(DU) material shows poor organized mesostructure whereas MCM-41(DU)-F exhibits well-organized hexagonal pore arrangement, implying that the addition of hydrofluoric acid is beneficial for the formation of well-organized mesoporous silica from waste powder.
This is consistent with the results of XRD and BET analyses. In addition, the uniform porosity of the obtained MCM-41(DU)-F revealed by TEM analysis is in agreement with the narrow pore size distribution (BJH) determined by N2 adsorption-desorption measurement. As a result, one can conclude that pure silica materials with ordered mesostructure can be directly recovered from the optoelectronic waste powder at room temperature with the assistance of CTAB, hydrofluoric acid and ammonium hydroxide.
Figure 4.6 (a) EDS spectrum of MCM-41(DU)-F; (b) TEM image of MCM-41(DU);
(c) TEM image of MCM-41(DU)-F.
b c
a
It could be speculated from the previous results of XRD, BET and TEM that silica materials with well-organized mesostructure can only be formed when all silicon species were firstly pre-activated by hydrofluoric acid. It was proposed by the reaction Equations (4)-(8) that the addition of hydrofluoric acid in the system can firstly induce the reactive fluorine ions with the non-reactive SiO2 in the waste powder and form more SiF62- species, which can be rapidly hydrolyzed to more reactive Si(OH)4 and/or SiO(OH)3- anions by adding ammonium hydroxide. Then, well-ordered MCM-41(DU)-F can be obtained via the strong electrostatic force between silicate species and CTAB molecules. On the other hand, if hydrofluoric acid was not added then the bulk SiO2 particles could not be dissolved and further hydrolyzed into reactive Si(OH)4 and/or SiO(OH)3- anions. Thus, fewer amounts of reactive silicate anions present in the mixture would result in lower specific surface area and pore volume of MCM-41(DU). Besides, the presence of undissolved SiO2 particles as extra-framework particles would also disrupt the uniformity of the mesostructure.
This is confirmed by the XRD analysis shown in Figure 4.3, where MCM-41(DU) shows stronger amorphous SiO2 crystallinity than that of MCM-41(DU)-F.
It is well-known that fluorine ions are beneficial for enhancing the silanol condensation of the silica materials. Therefore, it is expected that MCM-41(DU)-F should exhibit higher degree of silanol groups condensation. This hypothesis is further supported by the result of 29Si NMR analysis shown in Figure 4.7. The sharp peak centering at -112 ppm was observed for both as-synthesized materials of MCM-41(DU) and MCM-41(DU)-F, which can be assigned to the highly polymerized Q4 Si sites [Si(OSi)4] [116]. On the other hand, two distinguished peaks locating at -102 and -90 ppm, respectively, were observed as well. The former peak represents Q3 Si sites [Si(OSi)3OH] while the latter one represents Q2 Si sites [Si(OSi)2OH2] species [117]. It is noticeable that MCM-41(DU)-F showed higher Q4/ Q3 ratio, implying that MCM-41(DU)-F possessed higher degree of condensation of the silanol groups. This is probably due to the fact that the presence of fluorine ions which greatly enhanced the condensation reaction of the silanol groups. Generally, it is recognized that higher Q4/Q3 ratio of the silica materials represents higher hydrothermal stability [118].
Consequently, it can conclude that hydrofluoric acid in the system is not only favorable to the production of more reactive silicate species, but also beneficial for the enlargement of the pore diameter and it promotes higher degree of silanol
condensation.
The scanning electron microscopy (SEM) images of MCM-41(AF) and MCM-41(DU)-F samples are depicted in Figure 4.8. It can be clearly observed that the MCM-41(AF) prepared via alkaline fusion method was in tubular shape with length of ca. 1-2 µm and diameter of ~200 nm. On the contrary, MCM-41(DU)-F was in spherical shape with sizes of around 100-150 nm, which size was much smaller than that of MCM-41(AF). In direct utilization process, the pH value of synthetic medium is 6.9-7.0 and the silica condensation rate is a maximum at around pH value of 6-7 [119]. Therefore, the fast hydrolytic condensation of silica-surfactant mixture results in the formation of small-sized mesostructured particles.
Chemical shift (ppm)
-160 -150 -140 -130 -120 -110 -100 -90 -80 -70 -60
MCM-41(DU)-F MCM-41(DU) Q4
Q3
Q2
Figure 4.7 29Si NMR spectrum of MCM-41(DU) and MCM-41(DU)-F samples.
Figure 4.8 SEM images of (a) MCM-41(AF) and (b) MCM-41(DU)-F samples.
b a
The uniform mesostructure with large pore diameter and large pore volume of waste-derived mesoporous silica materials may imply them as potential adsorbents for CO2 capture. Figure 4.9 displays the breakthrough curves of 10% CO2 adsorption on three TEPA-functionalized waste-derived mesoporous silica adsorbents at 60 C via packed column reactor, including MCM-41(AF), MCM-41(DU) and MCM-41(DU)-F samples. Furthermore, MCM-41(NaSi) sample synthesized using pure chemical reagent of sodium silicate was also tested as support of adsorbent for its adsorption capacity. It is seen that initially the CO2 gas can be efficiently adsorbed on all adsorbents with capture efficiencies greater than 98%. The CO2 adsorption capacities of all adsorbents shown in Table 4.1 were in a range of 100-121 mg/g-adsorbent and follow the order of TEPA-MCM-41(DU) < TEPA-MCM-41(AF) ≈ TEPA-MCM-41(NaSi) < TEPA-MCM-41(DU)-F. This may be due to the fact that the textural properties of the supports play important roles in CO2 adsorption performance.
In a related work, Son et al. [66] employed a series of amine-functionalized mesoporous silica materials as support of adsorbent for CO2 adsorption. They found that the adsorption capacity was a function of the pore diameter of the support and followed the order of MCM-41 (2.8 nm, 1D) < MCM-48 (3.1 nm, 3D) < SBA-15 (5.5 nm, 1D) ≈ SBA-16 (4.1 nm, 3D) < KIT-6 (6.5 nm, 3D). However, in this study MCM-41(DU) with relatively larger pore diameter (4.4 nm) appears to have lower CO2 uptakes compared to those of MCM-41(AF) (3.0 nm), MCM-41(NaSi) (3.1 nm) and MCM-41(DU)-F (4.5 nm) (Figure 4.10(a)). On the other hand, it is noted from Figure 4.10(b) that a linear relationship with correlation coefficient R2 = 0.95 was observed between the total pore volume of the parent silica substrates and the CO2
adsorption capacity. This may suggest that the total pore volume of the parent support plays a predominant role instead of the pore diameter in CO2 capture.
The density of TEPA is 0.99 cm3 /g and the total pore volume of MCM-41(NaSi), MCM-41(AF), MCM-41(DU) and MCM-41(DU)-F is 1.00 cm3 /g, 0.99 cm3 /g, 0.52 cm3 /g as well as 1.10 cm3 /g, respectively. Thus the maximum TEPA amount theoretically loaded inside the pore channels was calculated to be 50%, 50%, 34% and 53%, for the MCM-41(NaSi), MCM-41(AF), MCM-41(DU) and MCM-41(DU)-F, respectively. It can be seen from Table 4.1 that TEPA-MCM-41(DU) exhibits tiny
TEPA. It is reported that more efficient contact between the CO2 gas and the impregnated amines could be achieved when a small space is still left insides the pores of the mesoporous silica after amine reagents loading [80]. As a result, the MCM-41(DU) had lower CO2 adsorption capacity due to the fact that it had smaller pore volumes, which may result in more constricted or blocked pores in the adsorbents.
In addition, the adsorption capacity appears to increase with an increase in pore diameter and follow the order of TEPA-MCM-41(AF) (3.0 nm) ≈ TEPA-MCM-41(NaSi) (3.1 nm) < TEPA-MCM-41(DU)-F (4.5 nm). This might be because larger pore diameter of the support could facilitate more TEPA into the pore channels more easily, leading to higher CO2 adsorption performance. As shown in Table 4.1, the pore volume of MCM-41(DU)-F and TEPA-MCM-41(DU)-F is 1.10 cm3 /g and 0.16 cm3 /g, respectively. Therefore, the amount of TEPA loaded into the pores of MCM-41(DU)-F was calculated to be 49%, while there was only 46% TEPA loaded into the pore channels of the MCM-41(NaSi) and MCM-41(AF), respectively.
In addition, the BJH pore size distribution shown in Figure 4.11 shows that the pore sizes of MCM-41(AF) and MCM-41(NaSi) are diminished to less than 2 nm after TEPA loading. In comparison, parts of pores still remain larger than 2 nm in TEPA-MCM-41(DU)-F. This could make the gas flow into TEPA-MCM-41(DU)-F to be easier, which is beneficial for CO2 adsorption. Consequently, it could be concluded that the large pore diameter and pore volume of MCM-41(DU)-F resulted in the highest CO2 capacity among all adsorbents.
Time (min)
Figure 4.9 CO2 breakthrough curves for TEPA-functionalized adsorbents of MCM-41(NaSi), MCM-41(AF), MCM-41(DU) and MCM-41(DU)-F samples.
Figure 4.10 (a) Relationship between the pore diameter of the mesoporous substrates and CO2 uptake and (b) correlation of total pore volume of the mesoporous substrates and CO uptake at 60 C.
Pore diameter (nm)
1 2 3 4 5 6 7 8 9 10 11 12 13
dV/dD, pore volume (cm3 g-1 , STP)
0.0002 0.0004 0.0006 0.0008 0.0010 0.0012
TEPA-MCM-41(NaSi) TEPA-MCM-41(AF) TEPA-MCM-41(DU)-F
Figure 4.11 BJH pore size distribution of TEPA-MCM-41(NaSi), TEPA-MCM-41(AF) and TEPA-MCM-41(DU)-F samples.