Arsenic-contaminated sludge remediation induced generation of coloured glass using conventional and microwave heating

The present work incorporates toxic arsenic-contaminated sludge (AsCS) obtained from groundwater �ltration into a chemically durable borosilicate glass matrix to overcome the environmental exposure's arsenic (As) toxicity. Upto 25 wt% of AsCS loaded borosilicate system found X-ray amorphous that ensures the successful glass formation. The optical absorption spectra reveal the utilization of AsCS to develop heat absorption property (NIR absorption ~ 1000 nm) into the borosilicate matrix. 3–5 wt% of AsCS was found suitable for developing a green colour glass for various purposes. Furthermore, green colour in glass can be turned yellow with the adoption of novel microwave heating (MW), which reduces energy demand. From a structural viewpoint, the presence of tetrahedral boron (B 4 ) and tetrahedral Fe 3+ within the silicate glassy network collectively help to maintain a low thermal expansion coe�cient (CTE) and higher characteristic thermal properties (T g, T d, etc.)Further, the structural integration of iron (major component in AsCS) as bridging network helps to maintain the thermal and chemical durability of the system. Signi�cantly lower leaching rate (7.47 *10 − 7 g.cm − 2 .h − 1 in 42 days) followed by a successful immobilization of toxic As, the glass becomes an alternative and most attractive material for AsCS immobilizing material.


Introduction
Arsenic contamination in groundwater has been proven to be deadly for human health (Berg et Yamauchi et al. 2004).Researchers across the globe are still working to establish a route to make arsenic below the permissible limit in drinking water (Mohora et al. 2018; Najib and Christodoulatos 2019).Several technologies (Berg et al. 2006; Jain and Singh 2012; Litter et al. 2019) are available now to make the water arsenic-free.Among them, the adsorption of arsenic from contaminated groundwater using iron oxide nanoparticles or iron oxide-based compounds via a membrane-based ultra ltration method was found effective.This method provides an effective solution for removing arsenic contamination; simultaneously it generates an arsenic-rich sludge that can potentially cause arsenic contamination.Therefore, plenty of research works are currently found in this area (Clancy et al. 2013).Dumping of sludge is a common practice in the land eld.Thus, uncontrolled environmental exposure of the separated sludge (AsCS) again enhances the possibility of iron, and arsenic contamination in nearby fresh water.Therefore, safe disposal of AsCS is a critical issue.However, some of the possible solutions regarding the safe disposal of AsCS is found in tiles and bricks preparation (Hassan et al. 2014; Rathore and Mondal 2017; Weng et al. 2003) like various ceramic products ( Roy et al. 2019; Kizinievic et al. 2013).Nevertheless, the higher amount of sludge loading is accompanied by higher leaching of arsenic (Hassan et al. 2014).Recently, Hong Quan and coworkers suggested the use of arsenic-contaminated sludge in cementitious material for long-term stabilization (Quan et al. 2022 ).Therefore, the concern remains due to the very low permissible limit of As (≤ 10 ppb) in drinking water recommended by WHO.Arsenic oxide is water soluble; hence a considerable amount of arsenic leaching from the disposed media can trigger the problem in the near future.
To address this, glass can be a material and borosilicate matrix will be suitable for incorporating AsCS (Lima and Monteiro, 2001).Borosilicate glass is known for its high chemical durability and corrosionresistive property both in acidic and basic medium that resists the leaching of ions from the glass surface.Borosilicate glass is commonly used to immobilize nuclear waste (Rautiyal et al. 2021).
Borosilicate matrix seems superior to the commercial soda-lime silicate glass, where the melting temperature reduces with a substantial amount of alkali content.This results in poor mechanical properties with low chemical durability.The AsCS enriches with iron, the major toxic element arsenic, and different amounts of alkaline earth oxides.Interestingly, these sludge constituents are commonly used raw materials in glass melting.In addition, different transition metal elements (Iron, etc.) are also used to prepare different colour glasses.Furthermore, arsenic remediation is possible by properly utilizing it as a ning agent to obtain bubble/seed-free glass.AsCS should be categorized as low-volume toxic waste produced in limited amounts with respect to the volume of water puri ed ( The present study focuses on 1) the safe immobilization of AsCS in a highly durable borosilicate system;2) optimization of sludge loading to evaluate the safe limit of AsCS incorporation in the studied matrix; and3) Finally, the in uence of MW heating in the change of optical properties.

Materials and Method
A series of borosilicate glass (50-300g) with the composition listed in

Characterizations
The X-ray diffraction pattern of AsCS and all the AsCS incorporated glasses were recorded in a powder Xray Diffractometer (Ultima IV, Rigaku) using Ni-ltered Cu-K radiation (λ = 1.5418Å) within the range (2θ) 10 -90 following a step size of 0.02 with a scan rate of 6 steps/second.The UV-Vis-NIR optical ℃ absorption spectra of the AsCS-loaded glasses were recorded in a UV-Vis-NIR spectrophotometer (PerkinElmer) in the working range of 200-2500 nm.The Fourier-transform infrared (FTIR) re ectance spectra of the AsCS content glasses (G1, G2, G3&M1, M2, M3) were recorded in an FTIR spectrophotometer (Frontier IRL 1280119, PerkinElmer) in the wavenumber range of 400-1500 cm − 1 .
The FTIR transmission spectrum of the AsCS powder was collected in the 400-4000 cm − 1 range using a KBr pellet.The eld emission scanning electron microscopy-energy dispersive X-ray spectroscopy (FESEM-EDX) technique was used to identify the major unknown elemental components of the sludge sample before incorporation in glass.The dried AsCS power was analyzed as a pellet with carbon coating using eld emission scanning electron microscopy (FESEM, Zeiss, SupraTM 35VP, Oberkochen, Germany).The characteristic Raman spectra of glasses(G1, G2) and M1, M2) were recorded in a Raman spectrometer : HORIBA Jobin Yvon, France (Model: Lab Ram HR 800 EV) using the excitation source of 488 nm Argon ion laser.The Raman curve tting for the different silicate units (Q n Si , where n is the number of bridging oxygen per silicon tetrahedra, n = 0-4) was performed using the Gaussian function in PeakFit software.The tting parameter (r 2 ) was considered up to 0.9997.The thermal properties of the AsCS-containing glasses (G1, G2, G3 & M1, M2, M3) were carried out using a push rod (alumina) horizontal Dilatometer (DIL 402C, Netzsch) with a heating rate of 5K/min.Cylindrical glass samples having dimensions 25 mm length and 6 mm diameter were used to collect the data of coe cient of linear thermal expansion (CTE, ) in the range 50-400 C temperature, glass transition temperature(T g ), dilatometric softening point (T d ).

Leaching study with G1 sample:
A polished glass of G1 (17.70*17.25*2mm) was immersed in a distilled water bath for 42 days with a thermal cycle of 75 C for 8h/day.A magnetic stirrer was used to create temporal turbulence in the solution.Leaching analysis was carried out after 14 and 28 days with the leachate solution for Fe and As.Further, it was kept for 42 days, and the leaching rate and UV-Vis spectrum were recorded.

Leaching test for G3 samples
Four sets of glasses (~ 2g each) G3 were immersed into 100 mL deionized water in a 100 mL Te on beaker for 14 days, 1 month 14 days, 3 months 17 days, and 6 months 19 days respectively, i.e., with 1, 2 and 3 months of time intervals in room temperature.600 µm-sized particles of G3 were used to expose the large surface area.Using a Te on beaker ensures no leaching of ions from the interface of the beaker.
The total exposed surface area (SA) of the glass particles is calculated using the formula (1) below, --------(1) The terms w, ρ, and r are the sample weight, density, and the glass particle's radius (assuming the glass particles' spherical geometry), respectively.Particle radius data was obtained from standard British mesh SA = 3w ρr size (B.S.S.).The density of the glass for the leaching study was measured utilizing Archimedes' buoyancy principle on a Mettler Toledo digital balance tted with a density measurement kit using solid glass samples.

Result and Discussion
The AsCS powder was rst analyzed using FESEM-EDX to determine elemental composition.The inset of Fig. 1(a) depicts the microscopic view of the sludge surface analyzed in one of the three positions for component veri cation.The composition was nearly identical irrespective of the position scanned, presented in Fig. 1(b).Figure 1(c) exhibits the XRD pro le of the dried AsCS, indicating a substantial amount of amorphous nature with some crystalline phase in it.Sludge usually contains iron in the form of Fe(OH) 3 precipitate; the major amorphous phase of the sludge is hydrated iron oxide (Sarkar et al. 2008).The adsorbed arsenic by the iron oxide particle is believed to exist in that amorphous phase.Calcium is a major element present in the AsCS (Fig. 1(b)), and eventually, the crystalline phases identi ed match well with calcium oxide (ICDD-00-028-0775) as well as calcium carbonate (ICDD-01-085-0849).FTIR spectrum in Fig. 1(d),the asymmetric stretching vibration of carbonate was recorded at 1468 and 1422 cm − 1 followed by the characteristic bending vibration at 876 cm − 1 con rming the presence of carbonate functionality.Literature reveals that a vibrational band at 712 and 876 cm − 1 is characteristic of calcite (Xyla et al. 1989).A broad absorption band at 3420 cm − 1 can be assigned for O-H stretching.Apart from molecular water, this band can also contribute to the iron hydroxide component in AsCS.
The AsCS-loaded glasses, as per the studied composition (Table 1), were prepared under the meltquenching route.Figure 3(a)1st row shows the glasses (G1, G2, G3, G5) prepared with 5-25% of AsCS in a resistive heating furnace, and the 2nd row indicates the 5-20% AsCS incorporated glasses prepared under MW heating.
(b) XRD pattern of prepared glass samples after incorporation of AsCS (5-25 wt%) Figure 2(a) depicts the visible appearance of the glasses with subsequent addition of sludge up to 25%.This preliminary observation indicates the darkening of glass with the incorporation of 10 wt% AsCS or more.A distinct change can be noticed between the glass G1 and M1 in Fig. 2(a).This signi es a strong role of MW heating in altering the optical properties of glass containing transition metals like Fe.The multicomponent sludge (AsCS), mainly iron oxide, imparts colour to the glass matrix.
Figure 2(b) illustrates the XRD Pro le of different amounts of AsCS-containing glasses with waste loading varying within 5-25%.The amorphous pro le depicts the sludge embedded within the glassy matrix without any crystalline phase separation.To further con rm the limit of AsCS loading in this borosilicate matrix, the waste loading was gradually enhanced, keeping the glass former components (SiO 2 and B 2 O 3 ) xed (G1, G2, G3).The maxima of the broad humps shift towards the lower 2θ side in the higher AsCS-containing glasses (G3, G5).This might indicate that the higher amounts of constituting ions in the sludge highly modify the primary glass network.
The conventional glasses up to G3 were x-ray amorphous, as observed in Fig. 2 A similar non-crystalline nature was evident in the XRD analysis of MW-melted glasses M2 and M3 samples (Fig. S2).The energy-saving and shorter processing time aspect is also clearly re ected in the Tt-P pro le of MW melting (Fig. S1).The MW melted glass M1 appears more transparent than G1 with a distinct colour variation.The UV-Vis-NIR absorption spectra were recorded to elucidate the observational changes.Figure 3(a) shows the characteristic optical absorption behavior of the sample G1 and G2 having 5% and 10% AsCS content, respectively.
The absorption coe cient increases gradually with subsequent incorporation of sludge (AsCS) from 5 to 10 wt%.The absorption maxima at 1µm due to Fe(II) in octahedral coordination is almost double.This suggests that the AsCS is enriched with iron with an almost homogeneous distribution throughout the sludge.The UV-Vis-NIR spectra depict the presence of iron in two different oxidation states (Fe 2+ , Fe 3+ ).In addition to that, the octahedral and tetrahedral coordination of Fe 2+ is resolved in the NIR region of ~ 1µm and ~ 2µm, respectively.However, except six-fold and four-fold coordination, evidence of ve-fold coordination of Fe 2+ was found in other spectroscopic studies (Rossano et al. 2008).Further, it is also found that the gradual increase of AsCS (5 to 10 wt%) results in a redshift near the cutoff region.This aligns with the previous observations of the modi er role of Fe 2+ regardless of its coordination states.
Conversely, the transmission spectra presented in Fig. 3(b) highlight the potential of MW heating over altering the optical property of glass with identical sludge content.The 5% AsCS-loaded glass G1 appears green with the corresponding visible transmission (~ 60%) maxima at 536 nm, whereas it turns yellow with an enhanced optical transmission (80% in NIR and > 70% in visible) for M1.Enhanced optical transparency and a moderate heat absorbing property in MW melting represent its uniqueness over conventional heating.Fe(II)/∑Fe ratio plays an important role in determining the visible color difference of the 5 wt.%AsCS containing sample.Therefore, it is quite obvious that the modulation of colour from green to yellow depends on the Fe 2+ content.The colour of the glass becomes yellow with increasing Fe 3+ content in the glass with its predominant blue absorption (436 nm absorption band near cutoff).
The AsCS is enriched with iron due to the iron oxide components (Berg et al. 2006;Hao et al. 2018) for effective As(III) and As(V) adsorption from contaminated water.As evident from UV-Vis-NIR absorption spectra, this iron in higher melting temperature gets equilibrated from Fe(III) to Fe(II) state.Further, it is also evident that the distribution of Fe(II) in the octahedral environment is more than the tetrahedral environment.This can be interpreted based on higher crystal eld stabilization of Fe(II) in the octahedral eld relative to the tetrahedral eld.
Figure 4(a) and 4(b)depict the variation of linear thermal expansion of the different AsCS-loaded samples derived from resistive and microwave heating.The respective data regarding the thermal properties of the samples are represented in Fig. 4(c) and 4(d) to show their nature of variation.In each case, the dilatometric softening temperature (T d ), which indicates a structural deformation point, was found reasonably high even after 20 wt% of AsCS loading.The dilatometric softening point produces a gradually increasing trend for conventional glasses (G1, G2, G3) while a sharp decrease of T d in the case of M3 is due to the higher boron content in the sample at the expense of SiO 2 .
Usually, the glass transition temperature (T g ) decreases with increasing non-bridging oxygens(NBO) in the structure, leading to structural depolymerization.However, T g shifted towards the higher temperature side with xed glass former content(SiO 2 , B 2 O 3 ), enhancing AsCS from 5-10%.Again, a decreasing trend of T g is found for the MW-prepared glasses (M1, M2, M3), which is similar to the decreasing SiO 2 content (Table 1).Large incorporation of AsCS consisting of signi cant modi er decreases T g of glasses (G3, M3).Hence, the property modi er role of iron oxide is clear with this data on the borosilicate system.However, the structural in uence of iron redox states in silicate networks is quite different and complex.
Researchers have found that the ferrous and ferric states show different bonding natures in silicate glasses.While Fe 2+ mostly plays the modi er role, Fe 3+ is found to act as a weakly former associated with the Si-O-Fe bridging network.It is worth mentioning that the relatively higher polarising power of the Fe 3+ ion imparts covalency to the Fe-O bond.The network bridging capability of Fe 3+ is re ected in the T g pro le from 5 to 10% AsCS-containing glasses.
In both cases, the coe cient of thermal expansion (CTE) decreases with the incorporation of AsCS.This result shows that the well-known cross-linking of the borate group with the silicate network (Hubert and Faber 2014) results in a more polymerized structure showing a lowering CTE trend.Notably, the extent of decrease in CTE value in G1 to G2 is found more than in the case of G2 to G3 (Fig. 5c).G1 to G2, a lowering of CTE value is associated with a slight increment of Tg.Hence, it could be the change in boron coordination from 3-fold to 4-fold with charge compensation from modi ers in uenced by AsCS constituent ions.But,10% AsCS (G2 to G3) as in G3 involves BO 3 to BO 4 − conversion, and the rest of it is responsible for NBO generation re ected in lower Tg value.This observation is quite similar in MWprepared glasses.Gradually, the increasing Fe 3+ content in the network and its subsequent participation in the former with BO enforces behind the lower CTE still in G3.
The property of the glass depends on its structure; changing its bonding pattern formation of NBOs with the incorporation of modi er ions signi cantly in uences its physical and chemical durability.However, a systematic structure-property correlation is very di cult to establish here.First, the predominant amount of modi er ions (from AsCS) enriched with iron is an external and non-stoichiometric component.Thus, the subsequent replacement of the stoichiometric amount of any sludge component is a signi cant limitation for the structural study.Still, with the help of the corresponding FTIR spectra shown in Fig. 5(a), (b) its bonding variation can be analyzed to establish a stable network as evidenced by thermal analysis.
The most intense and broadband around 1038 cm − 1 (Fig. 5a) and 1039 cm − 1 (Fig. 5b) can be assigned for Si-O b -Si asymmetric stretching vibration between two SiO 4 tetrahedra.Silica is the major glass former, and B 2 O 3 is known to polymerize within the glass network, creating Si-O-B linkage at the expense of some Si-O-Si network (Kamitsos et al. 1990).In the presence of modi er ions from AsCS, the charge compensates BO 4 units can cross-link into the silicate network.This structural argument is suitable for practically observing shifting the band at a higher energy side.A decrease in thermal expansion and a slight increase in T g from G1 to G2 support the argument(Fig.4c).Further, a decrease in T g with the additional modi er cations can be expected from the formation of NBOs from Si-O-Si bridging linkage of silicate unit without much degrading the Si-O-B linkage supported by previous ndings.The lower bond strength of Si-O in comparison to B-O might favour the fact.In the case of MW melted glasses, the gradual decrease in silica content in the batch material (Table 1) re ects the gradual decrease in Si-O-Si stretching band intensity.The shifting of the band exhibits a similar nature.Two low intense bands in the higher energy region 1300-1450 cm − 1 are expected for B-O-B stretching in three-coordinated boron (Kamitsos et al. 1990).
The structural unit of the glass matrix was examined with Raman spectral analysis, shown in Fig. 5(c),(d) for conventional and MW melted glasses with 5 and 10% AsCS loading.Few critical observations have been revealed through this study.First, the structural depolymerization effect is clear enough in both systems from 5 to 10% AsCS loading.The intensi cation of Q 2 Si band at the expense of the Q 3 Si band establishes the structural reorganization of the glass matrix to host the AsCS constituent modi er ions.However, this structural depolymerization effect is not re ected in the T g pro le of both systems.Usually, the glass network depolymerization effect is associated with the thermal property of the glass sample (lowering of T g and a gradual increment of CTE value).Here, both properties follow the opposite trend.
Thus, the major guiding factor preventing the decrease of T g is the formation of the Fe 3+ -O bond, which has a weakly bridging nature under predominant tetrahedral coordination interconnecting the silicate network (Cochain et al. 2012).Second, A peak at 632 cm − 1 is denoted for borosilicate ring vibration (Manara et al. 2009) where essentially boron is four-coordinated (B 4 ).Hence, the incorporation of B 4 within the silicate network is justi ed.
The deconvoluted Raman bands presented in Fig. 5[(e),(f),(g),(h)] depict the changes in silicate network connectivity upon adding AsCS from 5 to 10 wt%.For G1 and G2 (with xed SiO 2 and B 2 O 3 content) the incorporation of AsCS is an increment of the Q 2 and Q 4 bands at the expense of Q 3 .The corresponding values of the percent area occupied by the bands have been listed in Table 2. Further, it should be noted that the Q 2 band around 980 cm − 1 is also due to the contribution of Fe 3+ in FeO 4 tetrahedra (Cochain et al. 2012;Nayak and Erwin Desa 2018).Thus, the increment of the Q 2 si band is not simply the depolymerization effect of Q 3  Si ; the involvement of FeO 4 is also in uencing.This phenomenon to some extent supports the thermal property observations.Interestingly, the relative population of Q 2  Si and Q 4 Si are found more in M1 having less Fe 2+ (Fig. 3b) than in G1 with identical batch composition.This is also re ected in relatively higher T g for M1 (than G1).Therefore, the incorporation of Fe 3+ in the borosilicate structure as a bridging unitis rmly supported in these AsCS immobilized glasses.

Leaching analysis
The leaching analysis was initially carried out with a G1 sample slit at an elevated temperature due to its better visible transmission as colour glass.Figure 6(a) depicts the difference in optical absorption property of the sample (G1) before and after 42 days of immersion in deionized water with a subsequent thermal cycle as stated in section 2 (Materials and methods).The leaching rate was determined from the sample and found to be 7.47*10 − 7 g.cm − 2 .h− 1 .In the period of continuous thermal cycle, the leachate analysis after 14 and 28 days revealed that a trace amount of 0.016 and 0.019 ppm of Fe leached out (estimated by colorimetric method) from the G1 glass slit.In contrast, arsenic remained undetected in commercial Merck test kit (detection limit 10 ppb).In this context, a similar study focusing on the detailed analysis of the chemical durability of iron-containing sodium borosilicate glass can be found elsewhere (Konon et al. 2022).
The UV-Vis-NIR result is consistent with the leaching analysis data, showing a slight NIR absorption decrease(highlighted in Fig. 6(a)).This might result from a lower amount of iron detected in the leaching analysis, but the visible transmission remains unaltered.Further, G3, having 20 wt% AsCS content, was chosen for the leaching analysis test, and the leaching behavior is depicted in Fig. 6(b).The respective leaching rate and the total exposed surface area of the glass particles (600 µm) of G3 were calculated and found to be 4.80 (± 0.75) *10 − 7 g/cm 2 .dayand 79.04 ± 0.06 cm 2 , respectively.All the samples of G3 do not produce any detectable amount of leaching of arsenic in the leached solution examined by a commercial mark testing kit for arsenic, except G3(199), where 5 ppb of As has been recorded.This ensures that the targeted batch composition, more signi cantly the choice of borosilicate matrix, is one of the best ways to incorporate arsenic safely.Finally, the G3 and higher arsenic content G5 samples were subsequently analyzed by ICP-MS to verify that the toxic iron and arsenic were trapped in the glass matrix.The arsenic content in G3 and G5 was found to be 349.935and 639.945 ppb, whereas iron content was 227.6 and 302.6 ppm, respectively analyzed from 0.2 g of glass powder each.These data further con rm that arsenic and iron in the toxic AsCS are successfully immobilized into the borosilicate glass matrix.

Conclusion
A borosilicate glass composition has been identi ed to safely incorporate As-containing sludge from contaminated groundwater.Upon systematic compositional modi cation, up to 25 wt% of sludge could be incorporated within the borosilicate glassy phase.However, colour glass articles could be fabricated incorporating < 5 wt%, maintaining visible transparency.After incorporating sludge, the characteristic thermal properties (T g , T d ) were not degraded to a greater extent.Additionally, the lowering CTE trend in a broad temperature range of 50-400°C with a gradual increase of AsCS was found suitable for the practical use of these colour glasses.5% sludge-loaded glass at elevated temperature with a thermal cycle (75°C: 8h/day) produces minimum leaching of iron.Furthermore, the 20% sludge-loaded glass particles at room temperature up to 199 days of immersion in distilled water show no leaching of toxic arsenic.Only 5 ppb of As detection in the 20% sludge sample for 199 days is negligible, indicating the effective immobilization methodology for arsenic-contaminated sludge.
MW heating is suitable for immobilizing toxic AsCS in borosilicate glass in an energy-e cient way with improved visible transmission.It is a novel addition to the MW heating technique for sludge management.Further, it is also established that the colour of the glass matrix can be tuned from green to yellow using an alternative novel microwave heating.The correlating factor behind this observation is the Fe 2+ absorption (maxima at 1000 nm) in the octahedral eld, as revealed in the UV-Vis-NIR absorption spectra.

Declarations
Figures al. 2001; Rahman et al. 2009).The people face the severity of this problem, particularly near the Ganga Delta region of West Bengal and Bangladesh (Chakraborty et al. 2022; Chowdhury 2022; Ivy et al. 2022) and many people are suffering from arsenic poisoning (Gorby 1988; Sen Gupta et al. 2017; Yamauchi and Takata 2021).Numerous studies establish the adverse effect of chronic arsenic toxicity (Ahmed et al. 2021; Bhadauria and Flora 2007; Hall 2002; Prasad et al. 2020; (b).Henceforth, further addition of 25 wt% AsCS in the glass matrix results in phase separation of magnetite (Fe 3 O 4 ) and hematite (Fe 2 O 3 ) as identi ed with the ICDD database [see in Fig. 2(b) inset].The glass composition was modi ed to maximize sludge loading with subsequent replacement of SiO 2 with B 2 O 3 .Enhancing B 2 O 3 content at the expense of SiO 2 produces a similar x-ray amorphous nature of the sample (G5).