
==== Front
iScience
iScience
iScience
2589-0042
Elsevier

S2589-0042(24)01872-8
10.1016/j.isci.2024.110647
110647
Article
Self-sensitized photodegradation and adsorption of aqueous malachite green dye using one-dimensional titanium oxide nanofilaments
Walter Adam D. 1
Benamor Hiba 2
Ferrer Lucas M. 1
Reji Treesa 1
Curran Tracey 3
Schwenk Gregory R. 1
Hadji Mohamed 2
Creighton Megan A. 4
Barsoum Michel W. barsoumw@drexel.edu
15∗
1 Department of Materials Science and Engineering, Drexel University, Philadelphia, PA, USA
2 Laboratoire d’Etudes et Recherche en Technologie Industrielle, Université of Saad Dahlab Blida 1, Blida, Algeria
3 Academy of Natural Sciences of Drexel University, Philadelphia, PA, USA
4 Department of Chemical and Biological Engineering, Drexel University, Philadelphia, PA, USA
∗ Corresponding author barsoumw@drexel.edu
5 Lead contact

03 8 2024
20 9 2024
03 8 2024
27 9 11064719 4 2024
5 6 2024
30 7 2024
© 2024 The Author(s)
2024
https://creativecommons.org/licenses/by-nc/4.0/ This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Summary

Truly one-dimensional titanium oxide nanofilaments with a lepidocrocite structure (1DLs) were explored in the adsorption and photocatalytic degradation of aqueous malachite green (MG), a toxic polluting dye. Decolorization is monitored by ultraviolet-visible spectroscopy, and mineralization is confirmed by total organic carbon analysis. The 1DL/MG flocs are characterized by scanning electron microscopy and X-ray diffraction. 1DLs, a colloidal nanomaterial, exhibit flocculating behavior while demonstrating high affinity for MG, with a maximum uptake of >680 mg/g rapidly via ion exchange. Additionally, 1DLs decolorize MG under visible light only, unlike most available titania products, via a self-sensitization effect. MG is decolorized by 1DLs by >70% in 30 min under 1 sun exposure of visible light. Counterintuitively, dye adsorption increases as the normalized concentration by mass of 1DL decreases. Demonstrating high adsorption capacity and dye mineralization supports the use of 1DLs in water treatment and self-sensitization for photoelectrochemical devices, like solar cells.

Graphical abstract

Highlights

• One-dimensional titanium oxide material with 5 × 7 Å2 cross section

• Rapid, high adsorption capacity of malachite green primarily due to ion exchange

• Decolorization of dye occurs under visible light irradiation only, unlike other nano TiO2

• Photocatalytic activity causes mineralization of dye

Physics; Nanotechnology; Materials science

Subject areas

Physics
Nanotechnology
Materials science
Published: August 3, 2024
==== Body
pmcIntroduction

Malachite green (MG) is an organic basic (cationic) dye whose name is derived from the similarity of its greenish-blue hue to the mineral malachite. It is extensively used in the textile industry and is commonly chosen as a model dye in effluent treatment studies.1,2,3,4 It has been used since the 1930s in fish culture as a fungicide and antiseptic.5 As a member of the triphenylmethane dye family, MG has been proven to be toxic for humans and animals.6,7,8,9,10 In light of this, in 2022, the US Food and Drug Administration (FDA) named MG as an unapproved aquaculture drug with “high enforcement priority,” making its removal from waste streams of paramount importance.

Adsorption is a popular method for removing MG from water because it is cost-effective, simple, and practical. The most common material utilized is activated carbon (AC),6,11 with many research efforts focusing on identifying natural precursors such as adsorbents derived from rice husks,12 nut shells,13 coffee beans,14 or tea leaves.15 Clays16,17,18 and layered titanates19 have also been tested due to their ability to uptake charged dye species via ion exchange mechanisms. Additional engineered adsorbents like graphene oxide,20,21 zeolitic imidazolate frameworks (ZIFs),22 and metal organic frameworks (MOFs)23,24 have also been investigated and demonstrated uptake improvements over more common adsorbents. Although adsorption has shown potential, it is only a separation technology; further post-processing is needed to dispose of the dye. The spent adsorbents are generally landfilled or regenerated by desorbing the dye into another waste stream.25,26 Said otherwise, the dye problem remains.

Another approach to dealing with organic dyes is to degrade them via advanced oxidation processes (AOPs). These have been investigated for MG degradation in various reaction schemes.27,28,29 AOPs rely mostly on the generation of highly reactive oxygen species (ROS), like hydroxyl radicals (•OH) and superoxide radicals (-•O2). Many technologically mature AOPs are based on the Fenton reaction,30 relying on a homogeneous dispersion of ferrous ions to generate the very strong oxidizing agent •OH.31 Fenton reactions generally require high pH and strong oxidizers like hydrogen peroxide (H2O2) and result in a toxic sludge that requires further treatment. A more recent development is the heterogeneous Fenton process. This system utilizes solid catalysts containing iron that can be ionized to form Fe2+—required to oxidize the peroxide and form •OH.32 These have shown promising results, but their adoption is hindered due to the ongoing required presence of the peroxide species and the tendency for the system to leach ionized Fe3+ that ultimately forms sludge. Other AOPs include the use of UV/H2O233,34 or ozone,35 both of which present processing concerns and hazards.

AOPs that use photocatalysis have emerged recently due to their ability to produce ROS upon illumination under mild aqueous conditions, without the need for other oxidizing species. In many cases photocatalysts are water-insoluble metal oxide semiconductors like titanium dioxide (TiO2) or zinc oxide (ZnO). When photoactivated, these materials produce electron-hole pairs that are ultimately responsible for generating ROS by using either the hole to oxidize, or the free electron to reduce, water and/or dissolved oxygen. There is an array of detailed work on photocatalysis and its mechanisms available in previous studies.36,37,38,39,40,41

The majority of semiconductor photocatalysts with band-gap energies wider than 3 eV lack visible light activity. To solve this problem, dye sensitization is a common method meant to improve their overall spectral usage. Typical photocatalytic systems rely on the incident photon to directly excite an electron from the valence band of the semiconductor into its conduction band.42,43,44,45 Sensitization occurs when a dye molecule, instead, is excited by the incident photon and the excited electron transfers into the conduction band of the semiconductor.46,47,48,49 Dyes are generally strong visible light absorbers; they are commonly employed in solar cell applications to improve visible light absorbance of the system.50 However, in aqueous systems, like the work presented herein, the dye actually aids the photocatalyst in its ultimate self-demise in a phenomenon known as self-sensitization. This process has been studied for TiO2 nanoparticles with the dyes rhodamine B51,52,53 and eosin.54 These mechanisms rely on forming O2·- radicals which, in turn, form the remaining ROS through reactions with water; ultimately removing the need for a narrow band gap.

From the semiconductor perspective, a wide band gap, high conduction band energy, large specific surface area, and strong affinity for the sensitizing dye are preferred.50 The excited state potential of the dye molecule must exist at a higher energy than the conduction band minimum potential of the semiconductor, thus, providing a pathway for excited electrons to transfer from the dye to the semiconductor.51 Following band alignment, adsorption is critical for sensitization. The contact between the dye and semiconductor must allow for efficient electron transfer. Pan and coworkers concluded that a hydrated surface—with accessible bridging oxygen (Ti-O-Ti) surface sites for electrostatic interactions with the dyes—can improve the self-sensitization behavior.52,53

Nano- or low-dimensional materials have been the subject of many studies on the photocatalytic breakdown of MG, with particular emphasis on TiO2 as the catalyst. With a band gap >3 electron-volts (eV), unmodified TiO2 requires UV light to generate ROS and ultimately degrade MG.55 The two most common polymorphs of TiO2 are anatase and rutile, with band-gap energies of 3.2 and 3.0 eV, respectively. Evonik’s AEROXIDE® P25 is a combination of anatase and rutile and has widely been used as a standard throughout the TiO2 photocatalysis literature.56,57 The band gap of TiO2 can be lowered, through doping, such that MG degradation can be carried out with the visible part of the solar spectrum.58 Nano-ZnO has also been studied, similarly requiring modification since it has an even wider band gap (3.4 eV) than TiO2.59,60,61

Layered titanates, akin to this work, have also been exploited for the degradation of MG.62,63 Titanates consist of all titanium oxide compounds by which the oxygen to titanium atomic ratio (O:Ti) is >2, requiring additional cations to neutralize the charge. The focus of this work is related to the family of layered titanates by which the polymorph is lepidocrocite. These lepidocrocite titanates (LTs) are best described as sheets, with various widths, of distorted edge-sharing TiO6 octahedral units with a non-Ti cation existing in the interlayer space. Sasaki and coworkers made the discovery that acid treating bulk Cs-titanate, with a fully coordinated LT structure,64 resulted in the delamination and formation of individual protonic-LT 2D layers.65 Kasuga et al. was then able to hydrothermally produce alkali-LT nanotubes,66 which was a significant simplification over the existing protonic-LT fabrication process. However, alkali-LT products lack the hydrophilicity of protonic-LTs. Historically, LTs have existed as 2D sheets67 that, based on the cation, would scroll into what has been traditionally defined as 1D nanostructures, like nanotubes and nanowires.68,69,70

A recent discovery in our group has provided a much-simplified fabrication method for obtaining truly 1D LT nanofilaments.71,72 The minimal cross section of the base unit is just 5 × 7 Å2, or 2 × 2 TiO6 octahedra as shown in Figure 1. Various widths in the c direction have been observed, and all were <10 nm.73 These nanofilaments will be henceforth referred to as 1DLs. 1DL formation hinges on the use of tetramethylammonium hydroxide (TMAOH), a base with a large organic cation that allows for 1DLs to grow in the a-direction, while simultaneously capping growth in the other two orthogonal directions. Interestingly, when using alkali bases, we are unable to achieve similar results74 and instead formed nanotubes of alkali-LTs, similar to those reported by Kasuga et al.66Figure 1 Structure of primary building unit of 1DLs and MG

(A, B, and D) (A) a-c, (B) b-c, and (D) a-b planes of 1DL. Structure is composed of 2 edge-sharing TiO6 octahedra in the b and c direction. 1DLs stack in the b-direction and grow in a-direction. TEM images in Figure S1 best show this stacking order.

(C) Colloidal suspension of 1DLs showing Tyndall effect—indicative of a colloidal suspension.

(E) MG molecule.

(F) MG aqueous solution with a concentration of 10 mg/L. Note: (A), (B), and (D) are drawn to scale using van der Waals radii, (E) is not.

It is simple to obtain a variety of 1DL morphologies, including agglomerations of mesostructured microparticles,75,76 2D sheets comprised of misaligned filaments,77,78 and colloidal suspensions of the 1D building units. The drying of aqueous 1DL colloids results in the formation of the 2D sheets, best visible by transmission electron microscopy (Figure S1). These morphologies allow for a plethora of surface area-driven applications, including ion uptake and adsorption,75,76 sulfur anchoring for battery electrode materials,79 and self-sensitized photocatalytic dye degradation demonstrated on rhodamine 6G (Rh6G), a xanthene derivative, and crystal violet (CV), another triphenylmethane dye.80

Due to the extremely high adsorption uptake that we measured for both Rh6G and CV during our pilot study, we decided to expand our adsorption studies in this work to further probe our 1DL system. Some of the results we present here are counterintuitive. For example, reducing the concentration of the 1DL colloidal suspension actually increases the normalized—by 1DL mass—maximum uptake of dye. With regards to photocatalysis, the aim was to focus on a more widely studied dye system like MG in an effort to further understand the sensitization behavior of 1DLs. We demonstrate that 1DLs are again sensitized; however, in this instance, it is done by a traditionally non-sensitizing dye. To our knowledge, this is the first observation of this phenomenon for any titania/MG system. Additionally, MG is commercially available as a salt with chloride and oxalate anions. We show that anion choice does not impact the rate of decolorization, but does impact the total extent of mineralization.

Results

The experimental details are presented in the STAR Methods section at the end of this document.

Adsorption of MG on 1DL colloid leads to flocculation

When a 1DL colloidal suspension (Figure 1C) was mixed with an aqueous solution containing MG (Figure 1F), the color of the solution gets lighter and dark flocs form after several minutes. The flocs are visible to the eye (Figure 2A) and can easily be removed via filtration. Total flocculation only occurs, however, if the concentration of the MG solution is approximately greater than that of the 1DL colloid, on a per-mass basis. Conversely, if the 1DLs are in excess, some flocs will form, but the liquid will appear as if it is a colloidal suspension (Figure 2A); filtering this suspension through a syringe filter is quite difficult. It is also apparent that when the 1DL:MG ratio is >1:1, dye diffusion does not spontaneously occur (Figure S2) and must be forced by mixing. At lower 1DL concentrations, the dye readily diffuses through the 1DL suspension without mixing.Figure 2 Structural and morphological characterization of 1DL-MG oxalate flocs formed during adsorption

(A) Photographs of vials containing different mass ratios of 1DL to MG after shaking.

(B and C) SEM micrographs of 1DL-MG flocs, at different magnifications. Red box in (B) is to scale region associated with (C).

(D and E) XRD patterns of (D) crushed 1DL-MG flocs and (E) crushed 1DL filtered films. Insets of (D) and (E) show semi-log plots of same patterns with major peaks labeled. Initial concentration of MG in (A) in all vials is 10 mg/L with a total system volume of 10 mL.

A low-magnification scanning electron microscope (SEM) micrograph of a typical floc is shown in Figure 2B after washing with water and subsequent drying. On the micrometer scale, few features are visible—especially in regions that are not fractured surfaces. Nanoscale features are shown in Figure 2C. The random amalgam of waves and thin extrusions are a consequence of the one-dimensionality of the 1DLs. Under the TEM, some of these features are clearly 2D like (Figure S1).

The X-ray diffraction (XRD) pattern of the 1DL-MG flocs is shown in Figure 2D with the XRD of pure 1DLs shown in Figure 2E. The 1DL structural peaks of 002 and 200 still exist in the flocs, from which the a- and c-lattice parameters can be calculated. The results agree with our previous work and that in the 2D TiO2 literature. The 020, or stacking peak and its higher-order reflection peaks are significantly shifted to lower 2θ angles (labeled by asterisks in inset of Figure 2D) as compared to pure 1DLs (Figure 2E). Additionally, in the 1k0 region (2θ ≈ 28°) the peaks are significantly broadened. This region is also associated with the periodic stacking of the 1DLs. In summary, the MG-1DL flocs have significantly less ordered stacking than the pure 1DLs, as illustrated by the high relative intensity of the structural peaks (200, 002) compared to the peaks associated with the interlayer spacing (∗, ∗∗, ∗∗∗).

1DLs exhibit high uptake capacity of MG with little impact from competitive cations

The equilibrium and transient adsorption results (obtained at room temperature [RT] and unadjusted pH) for MG on 1DLs are shown in Figures 3A and 3B, respectively. During the exploration of the equilibrium adsorption experiments (Figure 3A), there was an increase in the normalized maximum adsorption capacity as the concentration of 1DLs decreased. The maximum uptake values for each sample are shown in the inset in Figure 3A. It must be mentioned that the molecular weight of MG+ (330 g/mol) was used in all calculations, intentionally omitting the weight contribution of the counter-anion—which in this case was oxalate. These maximum uptake results, and comments on the isotherm shapes, are discussed in the following.Figure 3 Adsorption of MG oxalate onto colloidal 1DLs under various conditions

(A and B) (A) Equilibrium and (B) transient adsorption of MG at RT for various 1DL concentrations. Dashed lines in (B) are Lagergren’s pseudo-first-order kinetic fits (Equation S1). Fitting parameters and details on Ho’s pseudo-second-order kinetic fit are available in Table S1.

(C) Impact on uptake by adding a competitive cation (Na+).

(D) Impact of pH on uptake. (B) conditions were [MG]0 = 100 mg/L with varying [1DL]. (C and D) conditions were each [MG]0 = 50 mg/L and [1DL] = 50 mg/L. All runs had a total volume of 10 mL and were mixed in the dark.

The kinetic data for a solution of 100 mg/L MG adsorbed onto the various concentrations 1DLs are shown in Figure 3B. Lagergren’s pseudo-first-order rate law (Equation S1) is the best model for these data over Ho’s pseudo-second-order rate law (Equation S2), with a rate constant (k1) of 0.11 min−1. Kinetic equilibrium is established after ≈ 1 h. As the 1DL concentration is reduced by half, to 50 mg/L, the rate of removal greatly improves with equilibrium uptake achieved in just 15 min. The reason for the increase in adsorption rate as the 1DL concentration decreases is unclear at this time but is the subject to further work. The lowest concentration sample of 10 mg/L does not fit well to either the pseudo-first- or pseudo-second-order rate laws. An additional set of kinetic data are shown in Figure S3 and Table S2 showing the system of 50 mg/L dye to 50 mg/L 1DL colloid in a 10 mL solution.

Determining the impact of competitive cations and pH is important for understanding any system intended for water treatment, due to the array of ions and pH values encountered. Adding Na+ ions, in the form of NaCl, showed very low impact on the overall uptake of MG on 1DLs (Figure 3C). If anything, it may have resulted in a slight increase (∼2%) in the uptake over the system with no Na+. When the pH was adjusted, using dilute NaOH or HCl, there was a noticeable reduction (>20%) in the removal rate by acidifying the system (Figure 3D). This trend is expected since at low pH the surface will be more positive, repelling the MG, a cationic species.

Photocatalytic decolorization results in peak shifts associated with MG degradation

Typically, dyes exhibit maximum absorbance wavelengths (λmax) that differ from their initial color. For example, MG exhibits a λmax of 617 nm (Figure 4A) corresponding to a red-orange hue, leading to the blue-green color observed (inset in Figure 4A). As the dyes are decolorized (e.g., via adsorption in the dark or degradation in the light), the overall intensity at λmax is reduced (Figure 4A). This is also exhibited by a fading in the color after adsorption and subsequent degradation (inset in Figure 4A).Figure 4 Monitoring the decolorization of MG oxalate over 1DLs

(A) Time dependence of typical UV-vis spectra of the MG/1DL system. Aliquots were taken at each time, indicated on panel, and filtered to remove 1DL flocs. Inset shows photographs of the filtered aliquots as a function of time.

(B) Peak shifts for λmax region of MG (black squares, left y axis) and relative intensity (to intensity at t = −30 min) increase at 366 nm (red triangles, right y axis). 60 min sample omitted in λmax shift due to absence of any peak in the 617 nm region at t > 30 min [MG]0 = 10 mg/L and [1DL] = 10 mg/L with a total system volume of 20 mL. Simulated solar spectrum with a 1 sun (100 mW/cm2) total irradiance, with the UV-vis mirror module (Figure S4, red). Shaded area in (B) denotes exposure in the dark.

Figure 4B (black squares) illustrates the shift in λmax as a function of time. There is no shift during adsorption in the dark (−30 to 0 min). However, a shift is apparent once the lamp is turned on at t > 0. The point at 60 min is absent because the system is completely decolorized at that point. Additionally, there is a peak at 366 nm that is absent in pure MG but emerges upon illumination. The change in peak intensity, relative to the pure dye sample, at t = −30 min, as a function of time is shown in Figure 4B (red triangles, right y-axis). This peak, its potential significance, and the observed hypsochromic shift upon illumination will be further detailed in the discussion section.

MG counter-anion has little impact on decolorization rate

As noted earlier, MG is encountered as both a chloride and a lower-cost oxalate salt. Figure 5A compares the rate of decolorization for the two salts. Nonlinear pseudo-first-order kinetic expression (Equation S3) fitting of the results yields k1 = 0.11 min−1 and k1 = 8.4 x 10−2 min−1 for MG oxalate (black squares) and chloride (blue circles), respectively. Based on the results, it is reasonable to conclude the anions have little effect of decolorization.Figure 5 Photocatalytic degradation of MG over 1DL as a function of time under 1 sun (100 mW/cm2) full simulated solar spectrum with varying counter-anions

(A) Decolorization of MG normalized to the concentration of MG after 30 min in the dark, to establish adsorption equilibrium. Runs were carried out in triplicate; error bars indicate standard deviation. Initial concentration of MG was 10 mg/L in a 1:1 mass ratio with 1DL, on a 20 mL scale. Analysis was completed by monitoring the decrease in intensity of λmax = 617 nm peak. Dashed lines indicate pseudo-first-order kinetic fits. Parameters listed in Table S3.

(B) Time dependence of MG mineralization measured by TOC. Error bars indicate standard deviation of duplicate analyses of single aliquots taken at each time. Boxed data points are those of pure dye samples. [MG]0 = 10 mg/L and [1DL] = 10 mg/L with a total system volume of 40 mL. Simulated solar spectrum with a 1 sun (100 mW/cm2) total irradiance, with the UV-vis mirror module (Figure S4, red). Shaded area in (B) denotes exposure in the dark.

With regards to mineralization, the total organic carbon (TOC) results (Figure 5B) provide additional insight. The pure MG oxalate (black squares) exhibits a slightly higher overall TOC with 3.9 mg/L compared to 3.2 mg/L for MG chloride (blue circles). This is reasonable, since the organic oxalate anion (C2O42−) will contribute to the TOC results, while Cl− will not. We also noticed an increase in the TOC (≈9 mg/L for both samples) when adding the 1DLs to the dye and allowing the systems to reach equilibrium in the dark (gray region in Figure 5B). This increase is most likely due to the TMA+, between the 1DLs, being replaced by the cationic MG. Once the lamp is turned on (clear region in Figure 5B), the TOC of the samples is reduced from 13 mg/L to 8.3 mg/L and 12 mg/L to 6.9 mg/L for MG oxalate (black squares) and chloride (blue circles), respectively.

Modifying irradiance wavelength range leads to changes in decolorization rate

Most photocatalysts require high-energy UV light to be active, which makes up a small fraction of the solar spectrum. Supplying artificial UV light is costly to employ at a large scale. A major thrust of our pilot study80 was the discovery that 1DLs decolorize both Rh6G and CV under visible light only. Figure 6 shows the decolorization of MG oxalate by both 1DL (black squares) and P25 (red circles). In an effort to maximize the potency of P25, we followed the procedure outlined by Pan and coworkers52,53 who immersed the P25 powder in water for 24 h prior to exposing it to the dye. This resulted in an increased level of adsorption, however, not quite to the 1DL level. If the P25 is left in water for a week, the particles lose much of their adsorption capacity (Figure S5) but still remained similarly active in decolorizing the MG when illuminated.Figure 6 Time dependence of photocatalytic decolorization of MG oxalate over 1DL and water-treated Aeroxide P25, under various light sources

Decolorization of MG over 1DL (black) and P25 (red) under 1 sun with, (A) full simulated solar spectrum, (B) UV-only spectrum; (C) and (D) visible-only spectrum. Green line in (D) shows absence of color change of MG, without catalyst under the visible-only spectrum. Dashed lines indicate either zeroth or pseudo-first-order kinetic fits (see Table S4). [MG]0 = 10 mg/L and [1DL or P25] = 10 mg/L with a total system volume of 20 mL. Analysis was completed by monitoring decrease in intensity at λmax = 617 nm. (A), (B), and (C) are normalized to the concentration of MG after 30 min in the dark. (D) same as (C), but normalized to MG initial concentration, before adsorption equilibrium was established. Error bars in (C) and (D) indicate standard deviation between runs completed in triplicate. Green data in (D) are those for MG without a catalyst present.

Under a full simulated solar spectrum (Figure 6A) with an irradiance of 1 sun, 1DLs outperform P25 with pseudo-first-order rate constants of k1 = 9.7 × 10−2 min−1 and k1 = 6.6 × 10−2 min−1, respectively. Due to the mirror module (ultraviolet-visible [UV-vis]), the solar simulator’s full spectrum for this study corresponds to the ≈300 to 600 nm range (Figure S4). In 30 min, 1DLs effectively decolorize the entire solution. When the visible light is removed with a shortpass filter (λ < 400 nm), so only UV light is present (Figure 6B), the 1DL decolorization changes to linear (zeroth order) kinetics (Equation S4), with a rate constant k0 = 1.3 x 10−2 min−1. With UV light alone, P25 outperforms the 1DLs (Figure 6B), maintaining its pseudo-first-order kinetics with k1 = 6.4 x 10−2 min−1, essentially quite similar to that under the full spectrum (Figure 6A).

By interchanging the mirror module of the solar simulator, the wavelength range can be adjusted to 385 to 740 nm (blue curve in Figure S4). To ensure only visible light is present, a longpass filter (λ > 400 nm) was also employed. 1DLs are able to decolorize MG under visible light only (Figure 6C, black squares), while P25 (Figure 6C, red circles) is not. Upon illumination, P25 desorbs dye into the solution. This is more easily visualized in Figure 6D, which includes the equilibrium absorption data for both 1DLs (black squares) and P25 (red circles). Figure 6D also shows that the color of MG, without the catalyst, is not impacted by visible light on the timescale of this experiment (green line in B). 1DLs decolorize MG under visible light according to the zeroth-order rate law with k0 = 2.2 x 10−2 min−1, nearly twice the rate of decolorization under UV light only (Figure 6B).

Discussion

The addition of colloidal 1DLs to an MG solution results in the formation of flocculated particles, or flocs. The latter are formed as a result of ion exchange between the TMA+ that exists between the 1DLs after processing and cationic MG. There is a slight change in the XRD patterns between as-synthesized 1DLs (Figure 2E) and the 1DL/MG flocs (Figure 2D), which is directly seen in the peaks associated with stacking, viz (0k0, 1k0). Due to the large reduction in relative intensity between the stacking peaks along the b-direction (0k0) and structural peaks (200, 002) between the as-synthesized 1DLs and the 1DL/MG flocs, it is reasonable to conclude that stacking is greatly reduced in the latter. This is important when considering the morphology shown in Figure 2C by which the flocs are composed of randomly oriented 1DLs with presumably MG adsorbed onto, or in between, their surfaces.

At ≈ 20 Å, the d-spacings of the flocs were significantly larger than the 11.5 Å of the initial 1DLs (Figure 2) demonstrating that the dye molecules readily replace the TMA+. The absence of a peak at the original 2θ suggests that the vast majority of TMA+ cations are replaced. Two adsorption modes—simply based on the shape and size of the MG molecule—are shown in Figure 7. Structurally, the most common resonance structures of MG will localize the positive charge interchangeably on the amine pendant groups81; thus, those groups would most likely coordinate to the negative 1DL surface. Further theoretical work is planned to explore the adsorption modes of cationic dyes onto 1DLs.Figure 7 Proposed adsorption modes of MG onto 1DL consistent with a d-spacing of 20 Å

Schematic is to scale. As-drawn shows ABA stacking configuration for 1DL; however, AAA is also possible.

The results shown in Figure 3A describe a fascinating, counterintuitive result by which reducing the overall 1DL concentration resulted in an increase in the maximum uptake. The most dilute sample, at 10 mg/L (blue triangles, Figure 3A), resulted in the highest maximum uptake of 3.5 mol/kg (1,100 mg/g). This high uptake is only achievable in a small range of equilibrium concentrations (≈10 μM), above which the uptake drops off severely. Increasing the colloidal concentration to 50 mg/L (red circles, Figure 3A), the system is slightly more stable, but with a lower maximum uptake of 2.4 mol/kg (780 mg/g) at ≈ 35 μM. Finally, the most concentrated colloid, 100 mg/L (black squares, Figure 3A), we tested resulted in an isotherm of a more traditional shape, with a broad maximum uptake of 2.1 mol/kg (680 mg/g) in a wider concentration range between 0 and 200 μM. This shape of this isotherm is quite similar to that seen for the adsorption of CV on 1DL in our pilot study, where the maximum uptake was also ≈2 mol/kg.80 From the results shown in Figure 3A, it is reasonable to conclude that the accessible surface area increases upon dilution. The simplest interpretation for this “less is more” observation is to assume the 1DLs go through a 1D-to-2D transition with increasing concentration. Said otherwise, the material behaves in a more 1D fashion with decreasing concentrations.

Each of these isotherms appear to follow a high affinity type at low concentrations—by which all the MG is adsorbed to a certain dosage. As the concentration increases, however, the behavior changes. Giles et al. discussed the presence of an isotherm that follows a similar shape to what we observe.82 They conclude this can occur when adsorbates, such as dyes, are highly concentrated; they associate in solution and form solvated micellar-like agglomerates. In simpler terms, the dye prefers itself over the surface of the adsorbent.

The 1DLs exhibit rapid adsorption kinetics at low concentrations. However, when the concentration of 1DLs is 100 mg/L in an equal concentration of MG the kinetics are reduced, where k1 = 0.11 min−1 (black squares, Figure 3B). To compare against our pilot study, we repeated the same kinetic experiment (50 mg/L dye to 50 mg/L 1DL colloid, in a 10 mL solution). The result is presented in Figure S3 and resembles the rapid kinetics observed with Rh6G and CV.80

As truly 1D LTs, they exist as almost entirely surface, where oxygen predominates—leading to strong negatively charged surfaces. The presence of TMA+ cations lasts only until MG is introduced, at which time ion exchange occurs. The Fourier transform infrared spectroscopy (FTIR) spectra of 1DLs and 1DL/MG flocs are compared in Figure S6, supporting the removal of TMA+ during adsorption. Ion exchange is common with clays,4,83 resins,84 and other layered materials like titanates.19 Flocculation, on the other hand, is a feature of organic systems, like polymers and surfactants.85,86 The maximum uptake of MG at RT for some of these materials, in addition to more commonly employed adsorbents, are compared against 1DLs in Figure 8. The 100 mg/L uptake was chosen for 1DLs since it is more stable over a wider range of equilibrium conditions, unlike the 50 mg/L and 10 mg/L 1DL systems.Figure 8 Maximum uptakes of MG on various adsorbents

From left to right: 1DLs (100 mg/L)—this work, activated carbon, AC,11 reduced graphene oxide (rGO),20 LT nanotubes,19 bentonite clay,17 ZIF-8,22 MOF MIL-100(Fe),24 Sn-pillared clay,18 Amberlyst 15 ion exchange resin,84 and anatase TiO2 nanoparticles.87

Figure 3C shows that the dye adsorption process is relatively insensitive to competitive cations, like Na+. Interestingly, this result is in direct contradiction to our work with Rh6G and CV; the reasoning behind this behavior is unclear to us at this point. The pH dependence (Figure 3D) of the adsorption indicates that the removal rate is the lowest at the lowest pH level of 3. Given that the point of zero charge for 1DLs is around pH = 4,76 it is not surprising that cationic MG is repelled by the positive charges on the 1DLs at low pH. As the pH is increased above 4, the 1DL surface becomes more negatively charged and more likely to adsorb the cationic MG, as observed.

A method of differentiating the removal contributions via adsorption vs. degradation is to monitor the shift in λmax as a function of time. In most dye systems, including MG, a hypsochromic—or blue—shift is observed during photocatalytic degradation. During the adsorption stage in the dark, since there is no chemical change in the MG, λmax for MG remains at 617 nm (black squares in gray region in Figure 4B). However, when the light is turned on, a hypsochromic shift is observed, and λmax shifts toward lower wavelengths. This has been widely attributed to the de-alkylation of external groups of the MG molecule,58,88,89 which is observed in degradation studies of other dyes as well.90,91 After 60 min of illumination, decolorization is complete to the eye, and there is no sign of a peak in the original λmax region (Figure 4A), signifying there is no triaryl character left to the molecule. However, as the sample is decolorized, there is an increase in activity around λ ≈ 366 nm. This region has been attributed to phenolic compounds92 and is comparable to existing mechanistic studies on MG degradation over titania.93

To this end, decolorization is not equivalent to total degradation. Mineralization is a common term in remediation and soil science. It is simply the complete degradation of organic compounds into their inorganic forms (i.e., CO2, NOx). Since it is an aqueous system, these will be released as gases during degradation and ultimately reduce the TOC of the solution.

As noted above, MG is crystallized as a cationic salt with both chloride and oxalate as counter-anions for neutralization. Both these anions have been demonstrated as hole scavengers94,95 and as potential competitors for adsorption.96,97 At this junction, our conjecture is that the majority of the adsorption of MG onto 1DLs is a result of cation exchange, so counter-anions should have little to no impact. As observed in Figure 5A, the oxalate and chloride anions do not differ significantly on the rate of decolorization (k1 = 0.11 min−1 and k1 = 8.4 x 10−2 min−1, respectively). However, the TOC (Figure 5B) is impacted by the anions. The organic oxalate anion contributes to a higher TOC overall (black squares, Figure 5B) than their chloride counterparts (blue circles, Figure 5B), but with similar rates of mineralization. Figure 5B shows the TOC of the pure dye (3.9 mg/L and 3.2 mg/L for oxalate and chloride, respectively), and, during adsorption, the TOC increased by ≈ 9 mg/L in both the cases. This increase is most probably due to the release of TMA+ during ion exchange.

Based on the results shown in Figure 6, there is little doubt that, like in our previous work, MG sensitizes 1DLs. However, it is important to determine if sensitization is feasible based on the redox potentials of both the dye and semiconductor. Figure 9 shows that the potentials align in a manner that would indeed allow for sensitization to occur in that the reduction potential of the excited state of MG (−1.24 V)98 is more negative than the conduction band energy edge of 1DL (-0.74 V).99 More details on the optical properties of 1DLs, including the band-gap determination and UV photoelectron spectroscopy (UPS) is available in Collin-Ulloa et al.99 This is a simple model that nevertheless gives some insight into the system at hand.Figure 9 Redox potentials of MG and 1DL

(Left) MG potentials are calculated from density functional theory, reproduced from Ng et al.98 (Right) 1DL reduction potential (3.26 V) and Fermi level (EF) were measured using UPS; oxidation potential is calculated using band-gap values determined from thin-film transmission; reproduced from Colin-Ulloa et al.99

With regards to photocatalysis, the display of self-sensitization is the most important result of this work. As far as we are aware, this is the first example of the sensitization of titania using MG. It was mentioned in a study on ZnO, however.59 MG, like CV, is not generally considered a sensitizing dye due to the absence of a binding group, like a carboxylic or cyanoacrylic acid.100 This work further cements 1DLs as materials with great promise in applications where sensitization is key, like dye-sensitized solar46,50 and photoelectrochemical cells.101 This will be the focus of further work in the near future. In summary, outside of outstanding adsorption capacity, the ability for 1DLs to degrade MG under visible light is core to their viability as a photocatalyst. 1DLs offer high levels of near-instantaneous MG adsorption coupled with visible light decolorization, due to self-sensitization, resulting in an effective way to remove MG from water and the environment.

Limitations of the study

The methods of monitoring the adsorption and photodegradation processes rely on taking aliquots from a main batch, or by taking discrete samples. Taking aliquots implies that each removal keeps the system at the same concentration.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER	
Chemicals, peptides, and recombinant proteins	
	
Malachite Green Oxalate	Thermo Scientific Chemicals	CAT#: A16186-14
CAS#: 2437-29-8	
Malachite Green Chloride	MilliporeSigma	CAT#: 38800
CAS#: 569-64-2	
Titanium Diboride (99.9%, ∼325 mesh)	Thermo Scientific Inc.	CAT#: 04455809
CAS#: 12045-63-5	
Tetramethylammonium Hydroxide (25% w/w in water, 99.9999%)	Thermo Scientific Chemicals	CAT#: 020932LV
CAS#: 75-59-2	
Ethanol (200 proof)	Decon Laboratories Inc.	CAT#: 2705
CAS#: 64-17-5	
Aeroxide® P25	Thermo Fisher Scientific	CAT#: 384290500	
Supelco Certipur® TOC standard solution	MilliporeSigma	CAT#: 1090170100	
	
Software and algorithms	
	
OriginPro – Data analysis and plotting	OriginLab	https://www.originlab.com/	
Fiji – Image Analysis	ImageJ	https://github.com/fiji/fiji	
	
Other	
	
UV VIS Spectrometer	Agilent/Shimadzu	Agilent - Cary 60 Shimadzu - UV-1700 Pharmaspec	
X-ray Diffractometer	Rigaku	Miniflex	
Scanning Electron Microscope	Carl Zeiss	Supra 50VP	
Solar Simulator	Asahi Spectra	MAX 350	
Thermal Power Meter and Console	Thorlabs	Head – S401C
Console – PM400	
400 nm Shortpass Filter	Edmund Optics	#84-702	
400 nm Longpass Filter	Edmund Optics	#62-981	
Total Organic Carbon Analyzer	OI Analytical	Aurora 1030w	
Monolayer Polypropylene Membrane	Celgard	#3501	
Syringe Filters (<0.45 μm, Hydrophobic PTFE)	Thermo Fisher Scientific	CAT#: 09719H	

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Michel W. Barsoum (barsoumw@drexel.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• All data reported in this paper will be shared by the lead contact upon request.

• This paper does not report original code.

• Any additional information required to re-analyze the data reported in this paper is available from the lead contact upon request.

Method details

Materials

All materials were used as-received without further modification. Malachite green oxalate; malachite green chloride; titanium diboride (TiB2); tetramethylammonium hydroxide (TMAOH); ethanol; hydrochloric acid (HCl); sodium hydroxide (NaOH); sodium chloride (NaCl); Aeroxide P25; phosphoric acid; sodium persulfate; potassium hydrogen phthalate (KHP); antipyrine. All experiments were conducted with ultrapure water (<18.2 mΩ/cm).

Synthesis of 1DLs

10 g of TiB2 was immersed in 87.5 g of TMAOH in a 250 mL polyethylene bottle that was heated and shaken in an incubator (Labnet International Shaking Incubator, NJ, US) at 200 rpm and 80°C for 4 days. The bottle was vented with a single 23-gauge needle. The weight of each reactant was chosen to keep a Ti:TMAOH mole ratio at 0.6:1, which is the ratio we have been working with since the discovery of 1DLs. This is by no means the optimal ratio, with respect to reaction kinetics, but it was used here so we can compare the results in this work with previous, and ongoing, work.

After the reaction, the resulting sediment was added to ethanol, vortex shaken, and centrifuged at 3,500 rpm for 2 min until a clear supernatant, that was discarded, and neutral pH, measured using pH strips, was achieved (usually 3 times). Water was added to the ethanol washed product, and the material was suspended by vortex shaking. After centrifugation, a highly stable colloidal suspension was obtained while unreacted TiB2 settled to the bottom. The colloidal 1DLs were diluted to an initial known concentration determined by vacuum filtering 2 mL of colloidal 1DL through a 25 μm thick microporous monolayer polypropylene membrane (Celgard 3501, Celgard, NC, US) over a fritted glass filter apparatus. After filtering, the solid was fully dried in an oven at 80°C and the weight of the residue was measured. Stock suspensions of colloidal 1DLs with a concentration of 1 g/L were used for all subsequent experimentation.

Primary characterization

When flocs were utilized for analysis, they were first allowed to dry, then washed 3 times with DI water, and filtered to remove the water. This removed any recrystallized dye from the system prior to analysis. The structure of the finely crushed filtered 1DL films and flocs were analyzed using an X-ray diffractometer, XRD (MiniFlex, Rigaku Corporation, Tokyo, JP) with a step of 0.02° and a 1 s hold time with Cukα radiation at 40 kV. The morphologies of the flocs were imaged using a field emission scanning electron microscope, SEM (Supra 50VP, Carl Zeiss AG, Jena, DE). Due to the non-conducting nature of the samples, they were sputtered for 30 s at 40 mV with a Pt/Pd sputter-coater (208HR, Cressington Scientific Instruments, Watford, UK).

Adsorption of MG on 1DLs

All adsorption experiments were conducted in the dark and at room temperature. The mixture of dye and 1DLs was added to a 20 mL glass scintillation vial and placed in a dark box on a horizontal oscillatory shaker (200 rpm). After the specified period of time, the vials were removed; an aliquot was taken and filtered through a <0.45 μm PTFE syringe filter. The resulting liquid was analyzed using a UV-VIS spectrophotometer (UV-1700 Pharmaspec, Shimazdu Corporation, Tokyo, JP or Cary 60, Agilent, Santa Clara, CA, USA) at 617 nm. If necessary, samples were diluted to obtain an absorbance value in the linear region of the calibration curve, then their concentration results were scaled appropriately.

For the equilibrium adsorption experiments (Figure 3A), samples were prepared according to Table S5. The mixtures were allowed to shake for 60 min prior to aliquot removal, filtering, and UV-VIS analysis.

For the kinetics of adsorption experiments (Figure 3B), samples were prepared according to Table S6. Samples were taken at various time intervals, aliquots were removed, filtered, and underwent UV-VIS analysis.

The effect of pH (Figure 3C) was studied at range of 3–9 using 0.1 M HCl/NaOH to adjust the pH of the MG oxalate solution prior to 1DL addition. Samples were prepared as follows: 0.5 mL of 1 g/L 1DL colloid, 1 mL of the pH adjusted 0.5 g/L MG, and 8.5 mL of water. The mixtures were allowed to shake for 60 min prior to aliquot removal, filtering, and UV-VIS analysis.

The impact of ionic strength began with making a 5 mL mixture of 1DLs, MG oxalate, and water prepared according to columns 2–4 in Table S7. The vials were shaken for 30 min. A NaCl solution was then added to each vial according to columns 5 and 6 in Table S7. These final mixtures were then allowed to shake for 30 min prior to aliquot removal, filtering, and UV-VIS analysis.

Photodegradation of MG

A 300 W Xe lamp (MAX 350 with a downward collimator, Asahi Spectra, Tokyo, JP) was positioned directly above a stir plate. The spectra of the lamp used in this study, and its associated mirror modules, are from the manufacturer and reproduced in Figure S4. Before each run, the lamp’s output irradiance was measured using a thermal power meter (S401C thermal power sensor head with a PM400 console, Thorlabs, Newton, NJ, US). In an effort to maintain consistency, an irradiance of 100 mW/cm2 – equivalent to 1 sun – was obtained and used throughout by modifying the output power of the lamp. The aqueous dye and catalyst mixture was placed in a beaker and constantly stirred at 400 rpm, with a PTFE-coated magnetic stir bar. To prevent stray light, the beaker was wrapped in opaque tape. Aliquots were taken at defined intervals, filtered, and analyzed via UV-VIS or TOC analysis.

In experiments where the visible or UV light was removed, a 400 nm shortpass (model #84-702) or 400 nm longpass (model #62-981) (OD 4.0, Edmund Optics, Barrington, NJ, USA) filter was used, respectively. Prior to putting the filter in the light path, an output irradiance of 110 mW/cm2 was obtained since the filters remove ≈10% of the incident light, regardless of wavelength. The lamp configuration for each figure panel is outlined in Table S8.

Dye concentrations were determined using the UV-VIS spectrophotometer at a scan rate of 300 nm/min. Spectra were obtained across the visible light range and into the UV range to monitor peak intensity reductions, peak shifts and/or formation of new peaks (Figure 4). Polymethyl methacrylate (PMMA) cuvettes were used – with a 1 cm path length – for measurements in the 800 nm–300 nm range.

Total organic carbon (TOC) analysis

Total organic carbon (TOC) in the form of non-purgeable organic carbon (NPOC) was measured using an Aurora 1030w TOC analyzer with 1088 autosampler (OI Analytical, College Station, TX, USA). Samples were diluted with water at a 1:20 mL ratio and transferred to 40 mL borosilicate glass vials (VWR TraceClean clear vials) which had undergone a multi-step cleaning process (20% HCl rinse and triple water rinse) and heated in a muffle furnace at 500°C for 4 h. NPOC was measured in accordance with Standard Method 5310C and EPA 415.3 by acidifying 3.0 mL aliquots of sample in a heated, 98°C glass reaction vessel with 2.0 mL of phosphoric acid (5% w/v) and sparging for 2.0 min with compressed air (Airgas Ultra Zero grade air) to strip off any purgeable organic and inorganic carbon. Sodium persulfate (10% w/v) was then injected into the sample, oxidizing the remaining carbon in the sample into CO2, which was detected by a non-dispersive infrared (NDIR) detector and reported in mass and concentration units. All samples were measured in duplicate and internal lab duplicates were performed every 5 samples.

The system is calibrated using a range of aqueous solutions of KHP, diluted from a stock NIST-traceable standard solution (Supelco Certipur TOC standard solution). The calibration range for these analyses ranged from 0.5 to 7.0 mg/L TOC. Calibration accuracy is verified by analyzing separate QC check solutions (3.5 mg/L and 7.0 mg/L antipyrine, Sigma) and lab fortified blanks (0.1mL 700 mg/L antipyrine in 20mL DI water). The detection limit for solutions is < 10 ppb.

Quantification and statistical analysis

All curve fitting was completed using the Levenberg–Marquardt (LM) iteration algorithm in the nonlinear curve fit tool in OriginPro. Any repeat measurements and their associated statistical analysis is as-described in figure captions.

Supplemental information

Document S1. Figures S1–S6, Equations S1–S4 and Tables S1–S8

Acknowledgments

XRD and SEM analyses were completed in the Materials Characterization Core at Drexel University. We appreciate the collaboration of The Academy of Natural Sciences of Drexel University and their expertise. We would also like to thank Dr. Yong-Jie Hu—of Drexel University’s Department of Materials Science and Engineering—for supplying the calculated 1DL structure. This work was funded by both the Ceramics Program of the 10.13039/100000001 National Science Foundation Division of Materials Research (DMR-2211319 ) and the Drexel Areas of Excellence & Opportunity (AEO) Pilot Award in Advanced Materials (284282 ).

Author contributions

Conceptualization, A.D.W., H.B., and M.W.B.; methodology, A.D.W. and H.B.; validation, A.D.W.; formal analysis, A.D.W. and H.B.; investigation, A.D.W., H.B., L.M.F., T.R., and T.C.; resources, M.H. and M.W.B.; writing – original draft, A.D.W.; writing – review and editing, H.B., M.H., T.C., M.A.C., and M.W.B.; supervision, M.H. and M.W.B.; funding acquisition, M.H. and M.W.B.

Declaration of interests

One of the authors, M.W.B., filed for a patent on the 1D titanate which is the subject of this work.

Supplemental information can be found online at https://doi.org/10.1016/j.isci.2024.110647.
==== Refs
References

1 Shedbalkar U. Jadhav J.P. Detoxification of malachite green and textile industrial effluent by Penicillium ochrochloron Biotechnol. Bioprocess Eng. 16 2011 196 204 10.1007/S12257-010-0069-0
2 Bansal P. Bhullar N. Sud D. Studies on photodegradation of malachite green using TiO2/ZnO photocatalyst Desalin. Water Treat. 12 2009 108 113 10.5004/DWT.2009.944
3 Amiri-Hosseini S. Hashempour Y. Photocatalytic removal of malachite green dye from aqueous solutions by nano-composites containing titanium dioxide: A systematic review Environ. Health Eng. Manag. 8 2021 295 302 10.34172/EHEM.2021.33
4 Raval N.P. Shah P.U. Shah N.K. Malachite green “a cationic dye” and its removal from aqueous solution by adsorption Appl. Water Sci. 7 2017 3407 3445 10.1007/s13201-016-0512-2
5 Foster F.J. Woodbury L. The use of malachite green as a fish fungicide and antiseptic Prog. Fish-Cult. 3 1936 7 9 10.1577/1548-8640(1936)318[7:TUOMGA]2.0.CO;2
6 Marking L.L. King E.L. Walker C.R. Howell J.H. Alien J.L. Luhning C.W. Harman P.D. Malachite green, its toxicity to aquatic organisms, persistence, and removal with activated carbon Investigations in Fish Control 16 1977 48
7 Srivastava S.J. Singh N.D. Srivastava A.K. Sinha R. Acute toxicity of malachite green and its effects on certain blood parameters of a catfish, Heteropneustes fossilis Aquat. Toxicol. 31 1995 241 247 10.1016/0166-445X(94)00061-T
8 Srivastava S. Sinha R. Roy D. Toxicological effects of malachite green Aquat. Toxicol. 66 2004 319 329 10.1016/J.AQUATOX.2003.09.008 15129773
9 Stammati A. Nebbia C. Angelis I.D. Albo A.G. Carletti M. Rebecchi C. Zampaglioni F. Dacasto M. Effects of malachite green (MG) and its major metabolite, leucomalachite green (LMG), in two human cell lines Toxicol. Vitro 19 2005 853 858 10.1016/J.TIV.2005.06.021
10 Gopinathan R. Kanhere J. Banerjee J. Effect of malachite green toxicity on non target soil organisms Chemosphere 120 2015 637 644 10.1016/J.CHEMOSPHERE.2014.09.043 25462308
11 Kumar K.V. Comparative analysis of linear and non-linear method of estimating the sorption isotherm parameters for malachite green onto activated carbon J. Hazard Mater. 136 2006 197 202 10.1016/j.jhazmat.2005.09.018 16730900
12 Gou Y. Yang S. Fu W. Qi J. Li R. Wang Z. Xu H. Adsorption of malachite green on micro- and mesoporous rice husk-based active carbon Dyes Pigments 56 2003 219 229 10.1016/S0143-7208(02)00160-2
13 Malik R. Ramteke D.S. Wate S.R. Adsorption of malachite green on groundnut shell waste based powdered activated carbon Waste Manag. 27 2007 1129 1138 10.1016/J.WASMAN.2006.06.009 17029775
14 Baek M.H. Ijagbemi C.O. O S.J. Kim D.S. Removal of malachite green from aqueous solution using degreased coffee bean J. Hazard Mater. 176 2010 820 828 10.1016/J.JHAZMAT.2009.11.110 20036052
15 Akar E. Altinişik A. Seki Y. Using of activated carbon produced from spent tea leaves for the removal of malachite green from aqueous solution Ecol. Eng. 52 2013 19 27 10.1016/J.ECOLENG.2012.12.032
16 Bosazza V.L. Adsorption of malachite green by clays and allied minerals Nature 146 1940 334
17 Ullah S. Ur Rahman A. Ullah F. Rashid A. Arshad T. Viglašová E. Galamboš M. Mahmoodi N.M. Ullah H. Adsorption of malachite green dye onto mesoporous natural inorganic clays: their equilibrium isotherm and kinetics studies Water 13 2021 965 10.3390/W13070965
18 Raghav L. Patanjali P. Patanjali N. Singh R. Adsorption of malachite green and chrysoidine-Y by Sn-pillared clay Env. Qual. Manag. 32 2023 181 193 10.1002/tqem.21888
19 Jiang R. Zan R. Zeng J. Wen X. Preparation of 3D pompon-like titanate nanotube microspheres and their adsorption properties on cationic dyes RSC Adv. 6 2016 97899 97906 10.1039/C6RA11562D
20 Gupta K. Khatri O.P. Reduced graphene oxide as an effective adsorbent for removal of malachite green dye: plausible adsorption pathways J. Colloid Interface Sci. 501 2017 11 21 10.1016/J.JCIS.2017.04.035 28431217
21 Arabkhani P. Asfaram A. Development of a novel three-dimensional magnetic polymer aerogel as an efficient adsorbent for malachite green removal J. Hazard Mater. 384 2020 121394 10.1016/J.JHAZMAT.2019.121394
22 Khoshnamvand N. Jafari A. Kamarehie B. Mohammadi A. Faraji M. Removal of malachite green dye from aqueous solutions using zeolitic imidazole framework-8 Env. Proc. 6 2019 757 772 10.1007/S40710-019-00384-9
23 Alqadami A.A. Naushad M. Alothman Z.A. Ahamad T. Adsorptive performance of MOF nanocomposite for methylene blue and malachite green dyes: kinetics, isotherm and mechanism J. Env. Manag. 223 2018 29 36 10.1016/J.JENVMAN.2018.05.090
24 Huo S.H. Yan X.P. Metal–organic framework MIL-100(Fe) for the adsorption of malachite green from aqueous solution J. Mater. Chem. 22 2012 7449 7455 10.1039/C2JM16513A
25 Omorogie M.O. Babalola J.O. Unuabonah E.I. Regeneration strategies for spent solid matrices used in adsorption of organic pollutants from surface water: a critical review Desalin. Water Treat. 57 2016 518 544 10.1080/19443994.2014.967726
26 Rial J.B. Ferreira M.L. Potential applications of spent adsorbents and catalysts: revalorization of waste Sci. Total Environ. 823 2022 153370 10.1016/J.SCITOTENV.2022.153370
27 Dutta K. Bhattacharjee S. Chaudhuri B. Mukhopadhyay S. Oxidative degradation of malachite green by Fenton generated hydroxyl radicals in aqueous acidic media J. Environ. Sci. Heal. A. 38 2003 1311 1326 10.1081/ESE-120021128
28 Oturan M.A. Guivarch E. Oturan N. Sirés I. Oxidation pathways of malachite green by Fe3+-catalyzed electro-Fenton process Appl. Catal., B 82 2008 244 254 10.1016/J.APCATB.2008.01.016
29 Hashemian S. Fenton-like oxidation of malachite green solutions: Kinetic and thermodynamic study J. Chem. 2013 2013 1 7 10.1155/2013/809318
30 Fenton H.J.H. LXXIII.—Oxidation of tartaric acid in presence of iron J. Chem. Soc.-T 65 1894 899 910 10.1039/CT8946500899
31 Lloyd R.V. Hanna P.M. Mason R.P. The origin of the hydroxyl radical oxygen in the Fenton reaction Free Radic. Biol. Med. 22 1997 885 888 10.1016/S0891-5849(96)00432-7 9119257
32 Thomas N. Dionysiou D.D. Pillai S.C. Heterogeneous Fenton catalysts: a review of recent advances J. Hazard Mater. 404 2021 124082 10.1016/J.JHAZMAT.2020.124082
33 Modirshahla N. Behnajady M.A. Photooxidative degradation of malachite green (MG) by UV/H2O2: influence of operational parameters and kinetic modeling Dyes Pigm. 70 2006 54 59 10.1016/J.DYEPIG.2005.04.012
34 Navarro P. Zapata J.P. Gotor G. Gonzalez-Olmos R. Gómez-López V.M. Degradation of malachite green by a pulsed light/H2O2 process Water Sci. Technol. 79 2019 260 269 10.2166/WST.2019.041 30865597
35 Baek M.H. Ijagbemi C.O. Kim D.S. Spectroscopic studies on the oxidative decomposition of malachite green using ozone J. Environ. Sci. Heal. A. 45 2010 630 636 10.1080/10934521003595779
36 Mills A. Le Hunte S. An overview of semiconductor photocatalysis J. Photochem. Photobiol., A 108 1997 1 35 10.1016/S1010-6030(97)00118-4
37 Fujishima A. Zhang X. Tryk D. TiO2 photocatalysis and related surface phenomena Surf. Sci. Rep. 63 2008 515 582 10.1016/j.surfrep.2008.10.001
38 Schneider J. Matsuoka M. Takeuchi M. Zhang J. Horiuchi Y. Anpo M. Bahnemann D.W. Understanding TiO2 photocatalysis: mechanisms and materials Chem. Rev. 114 2014 9919 9986 10.1021/cr5001892 25234429
39 Nosaka Y. Nosaka A.Y. Generation and detection of reactive oxygen species in photocatalysis Chem. Rev. 117 2017 11302 11336 10.1021/acs.chemrev.7b00161 28777548
40 Xu C. Ravi Anusuyadevi P. Aymonier C. Luque R. Marre S. Nanostructured materials for photocatalysis Chem. Soc. Rev. 48 2019 3868 3902 10.1039/C9CS00102F 31173018
41 Armaković S.J. Savanović M.M. Armaković S. Titanium dioxide as the most used photocatalyst for water purification: an overview Catalysts 13 2023 26 10.3390/catal13010026
42 Lee S.-Y. Park S.-J. TiO2 photocatalyst for water treatment applications J. Ind. Eng. Chem. 19 2013 1761 1769 10.1016/j.jiec.2013.07.012
43 Kumar S.G. Devi L.G. Review on modified TiO2 photocatalysis under UV/visible light: selected results and related mechanisms on interfacial charge carrier transfer dynamics J. Phys. Chem. A 115 2011 13211 13241 10.1021/jp204364a 21919459
44 Nosaka Y. Nosaka A. Understanding hydroxyl radical (·OH) generation processes in photocatalysis ACS Energy Lett. 1 2016 356 359 10.1021/acsenergylett.6b00174
45 Zhang J. Nosaka Y. Quantitative detection of OH radicals for investigating the reaction mechanism of various visible-light TiO2 photocatalysts in aqueous suspension J. Phys. Chem. C 117 2013 1383 1391 10.1021/jp3105166
46 Aduroja O. Jani M. Ghann W. Ahmed S. Uddin J. Abebe F. Synthesis, characterization, and studies on photophysical properties of rhodamine derivatives and metal complexes in dye-sensitized solar cells ACS Omega 7 2022 14611 14621 10.1021/ACSOMEGA.1C06772 35557707
47 Pino E. Calderón C. Herrera F. Cifuentes G. Arteaga G. Photocatalytic degradation of aqueous rhodamine 6G using supported TiO2 catalysts. A model for the removal of organic contaminants from aqueous samples Front. Chem. 8 2020 511915 10.3389/FCHEM.2020.00365
48 Hameed S. Kazmi S.A. Ahmad N. Khan W. Qurratulain Effect of different dyes on TiO2 based dye sensitized solar cell 2015 Annual IEEE India Conference 2015 INDICON 1 6 10.1109/INDICON.2015.7443767
49 Cassidy J.P. Tan J.A. Wustholz K.L. Probing the aggregation and photodegradation of rhodamine dyes on TiO2 J. Phys. Chem. C 121 2017 15610 15618 10.1021/acs.jpcc.7b04604
50 Hagfeldt A. Boschloo G. Sun L. Kloo L. Pettersson H. Dye-sensitized solar cells Chem. Rev. 110 2010 6595 6663 10.1021/cr900356p 20831177
51 Wu T. Liu G. Zhao J. Hidaka H. Serpone N. Photoassisted Degradation of dye pollutants. V. self-photosensitized oxidative transformation of rhodamine B under visible light irradiation in aqueous TiO2 dispersions J. Phys. Chem. B 102 1998 5845 5851 10.1021/jp980922c
52 Pan L. Zou J.-J. Zhang X. Wang L. Water-mediated promotion of dye sensitization of TiO2 under visible light J. Am. Chem. Soc. 133 2011 10000 10002 10.1021/ja2035927 21644522
53 Pan L. Zou J.-J. Liu X.-Y. Liu X.-J. Wang S. Zhang X. Wang L. Visible–light–induced photodegradation of rhodamine B over hierarchical TiO2: effects of storage period and water-mediated adsorption switch Ind. Eng. Chem. Res. 51 2012 12782 12786 10.1021/ie3019033
54 Zhang F. Zhao J. Shen T. Hidaka H. Pelizzetti E. Serpone N. TiO2-assisted photodegradation of dye pollutants II. Adsorption and degradation kinetics of eosin in TiO2 dispersions under visible light irradiation Appl. Catal., B 15 1998 147 156 10.1016/S0926-3373(97)00043-X
55 Ma Y. Ni M. Li S. Optimization of malachite green removal from water by TiO2 nanoparticles under UV irradiation Nanomater 8 2018 428 10.3390/NANO8060428
56 Ohtani B. Prieto-Mahaney O.O. Li D. Abe R. What is Degussa (Evonik) P25? Crystalline composition analysis, reconstruction from isolated pure particles and photocatalytic activity test J. Photochem. Photobiol., A 216 2010 179 182 10.1016/j.jphotochem.2010.07.024
57 Tobaldi D.M. Pullar R.C. Seabra M.P. Labrincha J.A. Fully quantitative X-ray characterisation of Evonik Aeroxide TiO2 P25® Mater. Lett. 122 2014 345 347 10.1016/j.matlet.2014.02.055
58 Lavand A.B. Bhatu M.N. Malghe Y.S. Visible light photocatalytic degradation of malachite green using modified titania J. Mater. Res. Technol. 8 2019 299 308 10.1016/J.JMRT.2017.05.019
59 Saikia L. Bhuyan D. Saikia M. Malakar B. Dutta D.K. Sengupta P. Photocatalytic performance of ZnO nanomaterials for self sensitized degradation of malachite green dye under solar light Appl. Catal. A 490 2015 42 49 10.1016/J.APCATA.2014.10.053
60 Meena S. Vaya D. Das B.K. Photocatalytic degradation of malachite green dye by modified ZnO nanomaterial Bull. Mater. Sci. 39 2016 1735 1743 10.1007/S12034-016-1318-4
61 Mohamed R.M. McKinney D. Kadi M.W. Mkhalid I.A. Sigmund W. Platinum/zinc oxide nanoparticles: enhanced photocatalysts degrade malachite green dye under visible light conditions Ceram. Int. 42 2016 9375 9381 10.1016/J.CERAMINT.2016.02.147
62 Prado A.G.S. Costa L.L. Photocatalytic decouloration of malachite green dye by application of TiO2 nanotubes J. Hazard Mater. 169 2009 297 301 10.1016/J.JHAZMAT.2009.03.076 19443110
63 Mohanty S. Moulick S. Maji S.K. Removal of malachite green from spiked pond water using titanate nanotubes Lect. N. Civil Eng. 207 2022 663 675 10.1007/978-981-16-7509-6_52
64 Grey I.E. Li C. Madsen I.C. Watts J.A. The stability and structure of Csx[Ti2−x4□x4]O4, 0.61 < x < 0.65 J. Solid State Chem. 66 1987 7 19 10.1016/0022-4596(87)90215-5
65 Sasaki T. Watanabe M. Michiue Y. Komatsu Y. Izumi F. Takenouchi S. Preparation and acid-base properties of a protonated titanate with the lepidocrocite-like layer structure Chem. Mater. 7 1995 1001 1007 10.1021/cm00053a029
66 Kasuga T. Hiramatsu M. Hoson A. Sekino T. Niihara K. Titania nanotubes prepared by chemical processing Adv. Mater. 11 1999 1307 1311 10.1002/(SICI)1521-4095(199910)11:15<1307::AID-ADMA1307>3.0.CO;2-H
67 Wang L. Sasaki T. Titanium oxide nanosheets: graphene analogues with versatile functionalities Chem. Rev. 114 2014 9455 9486 10.1021/cr400627u 24754464
68 Gao T. Fjellvåg H. Norby P. Crystal structures of titanate nanotubes: a Raman scattering study Inorg. Chem. 48 2009 1423 1432 10.1021/ic801508k 19143511
69 Sun X. Li Y. Synthesis and characterization of ion-exchangeable titanate nanotubes Chem-Eur J 9 2003 2229 2238 10.1002/chem.200204394 12772297
70 Esmat M. Farghali A.A. El-Dek S.I. Khedr M.H. Yamauchi Y. Bando Y. Fukata N. Ide Y. Conversion of a 2D lepidocrocite-type layered titanate into its 1D nanowire form with enhancement of cation exchange and photocatalytic performance Inorg. Chem. 58 2019 7989 7996 10.1021/acs.inorgchem.9b00722 31135144
71 Badr H.O. El-Melegy T. Carey M. Natu V. Hassig M.Q. Johnson C. Qian Q. Li C.Y. Kushnir K. Colin-Ulloa E. Bottom-up, scalable synthesis of anatase nanofilament-based two-dimensional titanium carbo-oxide flakes Mater. Tod. 54 2022 8 17 10.1016/j.mattod.2021.10.033
72 Badr H.O. Lagunas F. Autrey D.E. Cope J. Kono T. Torita T. Klie R.F. Hu Y.-J. Barsoum M.W. On the structure of one-dimensional TiO2 lepidocrocite Matter 6 2023 128 141 10.1016/j.matt.2022.10.015
73 Lagunas F. Bugallo D. Karimi F. Yang Y. Badr H.O. Cope J.H. Ferral E. Barsoum M.W. Hu Y.J. Klie R.F. Ion-exchange effects in one-dimensional lepidocrocite TiO2: a cryogenic scanning transmission electron microscopy and density functional theory study Chem. Mater. 36 2024 2743 2755 10.1021/ACS.CHEMMATER.3C02773
74 Schwenk G.R. Walter A.D. Badr H.O. Hassig M.Q. Kono T. Lagunas F. Montazeri K. Barsoum M.W. Synthesis of titanate-based lepidocrocite nanostructures by reacting TiC, TiB2, and TiN with NaOH or KOH at 95 °C under ambient pressure Ceram. Int. 49 2023 40001 40010 10.1016/j.ceramint.2023.09.328
75 Badr H.O. Cope J. Kono T. Torita T. Lagunas F. Castiel E. Klie R.F. Barsoum M.W. Titanium oxide-based 1D nanofilaments, 2D sheets, and mesoporous particles: synthesis, characterization, and ion intercalation Matter 6 2023 3538 3554 10.1016/j.matt.2023.07.022
76 Wang L. Badr H.O. Yang Y. Cope J.H. Ma E. Ouyang J. Yuan L. Li Z. Liu Z. Barsoum M.W. Shi W. Unique hierarchical structures of one dimensional lepidocrocite titanate with cation-exchangeable sites for extraordinary selective actinide capture for water purification Chem. Eng. J. 474 2023 145635 10.1016/j.cej.2023.145635
77 Sudhakar K. Karmakar A. Badr H.O. El-Melegy T. Hassig M.Q. Carey M. Masiuk S. Wu L. Qian Q. Kono T. One-dimensional, titania-based lepidocrocite nanofilaments and their self-assembly Matter 6 2023 2834 2852 10.1016/j.matt.2023.06.006
78 Schwenk G.R. Walter A.D. Barsoum M.W. Solvent-Driven Self-Assembly of One-Dimensional Lepidocrocite Titanium-Oxide-Based Nanofilaments Nano Lett 24 2024 7584 7592 10.1021/acs.nanolett.4c00921 38775805
79 Cardoza N.A. Badr H.O. Pereira R. Barsoum M.W. Kalra V. One-dimensional, titania lepidocrocite-based nanofilaments and their polysulfide anchoring capabilities in lithium–sulfur batteries ACS Appl. Mater. Inter. 15 2023 50973 50980 10.1021/acsami.3c03743
80 Walter A.D. Schwenk G.R. Cope J. Sudhakar K. Hassig M.Q. Ferrer L. Mininni A. Lindsay A.J. Barsoum M.W. Adsorption and self-sensitized, visible-light photodegradation of rhodamine 6G and crystal violet by one-dimensional lepidocrocite titanium oxide Matter 6 2023 4086 4105 10.1016/j.matt.2023.09.008
81 Xie B.-B. Xia S.-H. Liu L.-H. Cui G. Surface-hopping dynamics simulations of malachite green: a triphenylmethane dye J. Phys. Chem. A 119 2015 5607 5617 10.1021/acs.jpca.5b02549 25945894
82 Giles C.H. Smith D. Huitson A. A general treatment and classification of the solute adsorption isotherm. I. Theoretical J. Colloid Interface Sci. 47 1974 755 765 10.1016/0021-9797(74)90252-5
83 Bulut E. Özacar M. Şengil İ.A. Adsorption of malachite green onto bentonite: equilibrium and kinetic studies and process design Micropor. Mesopor. Mat. 115 2008 234 246 10.1016/J.MICROMESO.2008.01.039
84 Nandoost A. Bahramifar N. Moghadamnia A.A. Kazemi S. Adsorption of malachite green (MG) as a cationic dye on Amberlyst 15, an ion-exchange resin J. Environ. Public Health 2022 2022 4593835 10.1155/2022/4593835 35355855
85 Teixeira Y.N. Melo R.P.F. Fernandes M.R. Carmo S.K.S. Neto E.L.B. Malachite green removal using ionic flocculation Water Pract. Technol. 17 2022 1113 1128 10.2166/WPT.2022.054
86 Liu B. Lu H. Wu S. Wang Z. Feng L. Zheng H. Octopus tentacle-like molecular chains in magnetic flocculant enhances the removal of Cu(II) and malachite green in water Sep. Purif. Technol. 282 2022 120139 10.1016/J.SEPPUR.2021.120139
87 Abou-Gamra Z.M. Ahmed M.A. Abou-Gamra Z.M. Ahmed M.A. TiO2 nanoparticles for removal of malachite green dye from waste water Adv. Chem. Eng. Sci. 05 2015 373 388 10.4236/ACES.2015.53039
88 Maria Magdalane C. Kaviyarasu K. Judith Vijaya J. Jayakumar C. Maaza M. Jeyaraj B. Photocatalytic degradation effect of malachite green and catalytic hydrogenation by UV–illuminated CeO2/CdO multilayered nanoplatelet arrays: investigation of antifungal and antimicrobial activities J. Photochem. Photobiol., B 169 2017 110 123 10.1016/J.JPHOTOBIOL.2017.03.008 28314181
89 Afshar S. Samari Jahromi H. Jafari N. Ahmadi Z. Hakamizadeh M. Degradation of malachite green oxalate by UV and visible lights irradiation using Pt/TiO2/SiO2 nanophotocatalyst Sci. Iran. 18 2011 772 779 10.1016/J.SCIENT.2011.06.007
90 Rasheed T. Bilal M. Iqbal H.M.N. Hu H. Zhang X. Reaction mechanism and degradation pathway of rhodamine 6G by photocatalytic treatment Water Air Soil Pollut. 228 2017 291 10.1007/S11270-017-3458-6
91 Habib M.A. Muslim M. Shahadat M.T. Islam M.N. Ismail I.M.I. Islam T.S.A. Mahmood A.J. Photocatalytic decolorization of crystal violet in aqueous nano-ZnO suspension under visible light irradiation J. Nanostru. Chem. 3 2013 70 10.1186/2193-8865-3-70
92 Aleixandre-Tudo J.L. Toit W.D. The role of UV-visible spectroscopy for phenolic compounds quantification in winemaking Front. New Trends Sci. Fermented Food Beverages 1 2018 200 204 10.5772/INTECHOPEN.79550
93 Ju Y. Yang S. Ding Y. Sun C. Zhang A. Wang L. Microwave-assisted rapid photocatalytic degradation of malachite green in TiO2 suspensions: Mechanism and pathways J. Phys. Chem. A 112 2008 11172 11177 10.1021/JP804439Z 18841945
94 Iguchi S. Teramura K. Hosokawa S. Tanaka T. Effect of the chloride ion as a hole scavenger on the photocatalytic conversion of CO2 in an aqueous solution over Ni–Al layered double hydroxides Phys. Chem. Chem. Phys. 17 2015 17995 18003 10.1039/C5CP02724A 26096980
95 Gekko H. Hashimoto K. Kominami H. Photocatalytic reduction of nitrite to dinitrogen in aqueous suspensions of metal-loaded titanium(IV) oxide in the presence of a hole scavenger: an ensemble effect of silver and palladium co-catalysts Phys. Chem. Chem. Phys. 14 2012 7965 7970 10.1039/C2CP40729A 22539059
96 Janusz W. Matysek M. Coadsorption of Cd(II) and oxalate ions at the TiO2/electrolyte solution interface J. Colloid Interface Sci. 296 2006 22 29 10.1016/J.JCIS.2005.08.067 16212973
97 Krivec M. Dillert R. Bahnemann D.W. Mehle A. Štrancar J. Dražić G. The nature of chlorine-inhibition of photocatalytic degradation of dichloroacetic acid in a TiO2-based microreactor Phys. Chem. Chem. Phys. 16 2014 14867 14873 10.1039/C4CP01043D 24926853
98 Ng C.H. Ohlin C.A. Winther-Jensen B. Characterisation of a series of triarylmethane dyes as light harvesters for photo-electrochemical systems Dyes Pigm. 115 2015 96 101 10.1016/J.DYEPIG.2014.12.016
99 Colin-Ulloa E. Martin J.L. Hanna R.J. Frasch M.H. Ramthun R.R. Badr H.O. Uzarski J.R. Barsoum M.W. Grimm R.L. Titova L.V. Electronic structure of 1D lepidocrocite TiO2 as revealed by optical absorption and photoelectron spectroscopy J. Phys. Chem. C 127 2023 7275 7283 10.1021/acs.jpcc.2c06719
100 Zhang L. Cole J.M. Anchoring groups for dye-sensitized solar cells ACS Appl. Mater. Interfaces 7 2015 3427 3455 10.1021/AM507334M 25594514
101 Zhang S. Ye H. Hua J. Tian H. Recent advances in dye-sensitized photoelectrochemical cells for water splitting EnergyChem 1 2019 100015 10.1016/J.ENCHEM.2019.100015
