TiO₂–reduced graphene oxide for the removal of gas-phase unsymmetrical dimethylhydrazine

Abstract

Unsymmetrical dimethylhydrazine (UDMH) contaminated waste gas and related intermediates pose a great threat to human health. TiO₂–reduced graphene oxide aerogel (rGA) samples with different graphene content levels were synthetized and characterized for the degradation of UDMH. The effects of GO content, humidity, and temperature were investigated under UV and VUV light, with highest UDMH conversion values of 68% and 95%, respectively. Compared with pure TiO₂, the enhanced degradation activity of TiO₂–rGA under UV light can be attributed to a synergetic effect between absorption and photocatalysis, while the high UDMH conversion under VUV light relies on photolysis and ozonation. The high oxygen-containing group content, rather than a high SSA, and electron trapping by graphene are key factors determining the outstanding performance of TiO₂–rGA with 80 mg of GO. The prepared TiO₂–graphene aerogels are promising for the degradation of gas-phase UDMH.

---

1. Introduction

Unsymmetrical dimethylhydrazine (UDMH) is a commonly used propellant in rocket and missile engines because of its high specific impulse. The waste liquid and gas produced during its use and storage can cause serious environmental pollution. UDMH is also seriously toxic to organisms, and it can cause different degrees of damage to the human central nervous system, liver, and kidneys. The treatment of UDMH waste has long been a research hotspot around the world. Traditionally, waste UDMH gas was treated via water absorption, high-altitude emission, and direct combustion methods. However, these methods often have the disadvantages of needing a lot of energy, being labor intensive, and having high cost. It is necessary to seek an effective, green, and convenient method for UDMH degradation. Photocatalysis and absorption are the most effective means that can be applied to remove UDMH contaminants. Gao¹ used CdS-sensitized TiO₂ nanorod arrays decorated with Au nanoparticles to degrade UDMH wastewater with an efficiency of 51.51%. However, little research has been done on the degradation of gas-phase UDMH using photocatalysts.

Photocatalysis, as an advanced oxidation technology, has been used in the field of waste-gas treatment, such as for formaldehyde,^2,3 benzene,^4,5 and toluene.^6–8 Titanium oxide (TiO₂) is considered to be a distinguished photocatalyst for remediating environmental pollution due to its unique features of being low-cost, non-toxic, and environmentally friendly.⁹ TiO₂ can create electron–hole pairs under illumination, followed by the generation of free radicals that are able to undergo secondary reactions. For practical applications, the problem of high photoinduced electron and hole recombination rates exists.¹⁰ To overcome this problem, a Schottky junction can be fabricated on a surface via loading with conductive materials with higher work functions than TiO₂.¹¹

Graphene, consisting of super strong sheets of carbon, conducts heat and electricity extremely well. TiO₂ and graphene can be integrated to effectively transport photoinduced carriers and prolong carrier lifetimes. What is more, TiO₂/graphene can easily form chemical bonds in the process, and the close interaction between TiO₂ and graphene is conducive to charge transfer and band-gap reduction.¹² During the in situ preparation of TiO₂/graphene, hydrophilic groups, including hydroxyl and carboxyl groups on GO, facilitate dispersion in water and the heterogeneous nucleation and growth of titanium dioxide. A. Trapalis¹³ explored the use of TiO₂/graphene composite photocatalysts made from pure graphene (G) and graphene oxide (GO) to remove NO_x. The results showed that TiO₂/rGO prepared via a solvothermal method exhibited superior efficiency to TiO₂/G. For TiO₂/rGO, the type and number of remaining oxygen-containing groups on rGO vary as a result of different degrees of reduction during the hydrothermal reaction. In addition, it is easy to carry out the chemical absorption of UDMH molecules on these groups. Therefore, chemical absorption and charge-carrier separation are proposed as key factors for improving the degradation performance of TiO₂/rGO. In a previous report, Ebrahimi¹⁴ prepared P25–rGO with P25 nanoparticles uniformly well-decorated on graphene, and the efficiency of the degradation of gaseous acetaldehyde reached about 70% at a velocity of 17 mL min⁻¹ under visible light.

Nowadays, another challenge hindering the widespread application of granular nanoparticles is that most photocatalysts are difficult to gather and recover. TiO₂/reduced graphene oxide aerogel (rGA) obtained via freeze-drying is easy to recycle. Also, the high surface area and countless bonding sites of 3D aerogels are favorable for physical and chemical adsorption.

In this study, we combine reduced graphene oxide with TiO₂via a one-pot hydrothermal method. The degradation of gaseous UDMH by TiO₂/rGA was investigated under ultraviolet (UV) light and vacuum ultraviolet (VUV) light in flowing gas. The phase structure, texture, and optical properties were characterized to understand the underlying reasons behind the improved UDMH conversion. Also, a possible degradation mechanism and proposed pathway are discussed based on the degradation products.

2. Experimental

2.1. Preparation of TiO₂/rGA

Graphite oxide was prepared according to the modified Hummers’ method.¹⁵ Briefly, 3 g of natural graphite was mixed with 70 mL of concentrated H₂SO₄ under stirring, then 1.5 g of NaNO₃ and 9 g of KMnO₄ were slowly added to the mixture in an ice bath. Afterwards, the solution was removed from the ice bath and kept between 35 and 40 °C for 90 min under strong stirring. Then, 140 mL of deionized (DI) water was added and the solution was stirred at 95 °C for 15 min. After that, 500 mL of DI water and 20 mL of 30% H₂O₂ were added. Finally, the solution was filtered and washed with HCl (1 : 10) and DI water several times. The supernatant was then removed via centrifugation to obtain a brown slurry. Then, the slurry was diluted and subjected to ultrasonic treatment for 1 h before further use. To prepare TiO₂/rGA, titanic sulfate (Ti(SO₄)₂) was chosen as the precursor, and glucose acted as the crosslinking agent and reducing agent. A certain amount (20 mg, 40 mg, 80 mg, 120 mg, or 160 mg) of GO, 0.03 g of glucose and 0.72 g of Ti(SO₄)₂ were magnetically stirred for 30 min and placed into a Teflon-lined autoclave, which was maintained at 180 °C for 12 h. The prepared TiO₂/rGO hydrogel (rGH) samples were washed and dehydrated via freeze-drying to form TiO₂/rGA samples, referred to as TiO₂/rGA-1, TiO₂/rGA-2, TiO₂/rGA-3, TiO₂/rGA-4, and TiO₂/rGA-5. In this work, TiO₂/rGA represents TiO₂/rGA-3 unless otherwise specified. For comparison, rGA (without adding Ti(SO₄)₂) and TiO₂ (without adding GO) were also synthesized via the same procedures.

2.2. Characterization methods

The phase structures of the samples were analysed via X-ray diffraction (XRD) with Cu Kα radiation (λ = 0.15401 nm) using a generator voltage of 40 kV and a current of 40 mA. Analysis was performed over a 2θ range of 5.0–85.0° at a scanning rate of 0.02° per second. The surface morphologies were characterized using a scanning electron microscopy (SEM) system (VEDAIIXMUINCN) equipped with an energy-dispersive X-ray spectroscopy (EDS) system. Raman spectra of the samples were taken using Ar⁺ (532 nm) laser excitation (inVia Reflex, Renishaw, England) in the range of 100–3200 cm⁻¹. The sample was loaded on a silica wafer and focused using a 50× objective. The surface functional groups were analysed via Fourier-transform infrared (FT-IR) spectroscopy (NEXUS, Nicolet, USA) in transmittance mode with a spectral range of 400–4000 cm⁻¹. X-Ray photoelectron spectroscopy (XPS) characterization was performed (Escalab 250Xi, Thermo Scientific) with an Al Kα X-ray source. All binding energy values were corrected via calibrating to the C 1s peak at 284.8 eV. Brunauer–Emmett–Teller (BET) specific surface areas and porosities of samples were evaluated based on nitrogen adsorption isotherms measured at 77 K using gas adsorption apparatus (TriStar II 3020, Micromeritics, USA). All samples were degassed at 180 °C before nitrogen adsorption measurements. The BET surface areas were determined using adsorption data in the relative pressure (p/p₀) range of 0.06–0.2. PL measurements were performed at room temperature using a Hitachi F-4700 fluorescence spectrophotometer.

2.3. UDMH degradation experiments under UV and VUV light

The photocatalytic degradation of UDMH was investigated in a continuous fixed-bed reactor, as shown in Fig. 1. UDMH was put in a refrigerator purged using nitrogen at a flow rate of 1 mL min⁻¹. Water vapor was generated via bubbling purified air provided by an air compressor. The flows of UDMH, water vapor, and purified air were adjusted using a mass flowmeter and then mixed in a buffer bottle. The total flow rate is controlled at 0.6 L min⁻¹. The gas from the buffer bottle then flowed into the reactor.

Fig. 1 A schematic diagram of the reaction system.

A cylindrical reactor made of stainless steel with a working volume of 500 mL was used in the experiments. UV and VUV lamps and a catalyst quartz tube were fixed vertically in the reactor. The UV and VUV lamp tubes used in the experiment were both 4 W lamps (Cnlight) with irradiation peaks at 254 nm and 185 nm, respectively. The distances between the lamps and the quartz tube were both 2 cm. The as-synthesized nanocomposite was loaded in the tube, through which the gas would be detected by the analysis system.

The concentration of UDMH was measured via a gas chromatograph (Clarus 680, PerkinElmer, USA)-mass spectrometer (Clarus SQ 8T, PerkinElmer, USA). The gas was periodically transferred into a sampling loop (500 μL) via a ten-way valve and separated through a capillary column (Elite-WAX, PerkinElmer) with an inner diameter of 0.25 mm and a length of 2 m. The oven temperatures were programmed as follows: the temperature was initially set to 50 °C and maintained for 1 min, then increased to 100 °C at a rate of 20 °C min⁻¹ and maintained for 1 min, and then increased to 180 °C at a rate of 10 °C min⁻¹ and maintained for 1 min. The measurement time for each sample was 3 hours. UDMH conversion was defined as follows:

UDMH conversion (%) = (UDMH_inlet − UDMH_outlet)/UDMH_inlet × 100%

where UDMH_inlet and UDMH_outlet are the concentrations of UDMH at the inlet and outlet, respectively. Each experiment was duplicated under the same conditions three times.

3. Results and discussion

3.1. Phase structure and chemical composition

Fig. 2 displays the XRD patterns of GO, rGA, TiO₂, and TiO₂/rGA. The diffraction peaks at 10.9° and 26.5° represent GO and rGA. The diffraction peak at 26° in the GO pattern is attributed to unoxidized graphite in small amounts. The d-spacing values of the graphite layers calculated using Bragg's law (2d sin θ = nλ) are 0.804 nm and 0.336 nm. The d-spacing of GO was broadened compared with natural graphite due to the introduction of functional groups between the graphite layers. The preparation of rGA involves the reverse process, namely the removal of functional groups from GO. The wide peak in rGA represents the high stripping of graphene and the formation of few-layer rGO sheets. The peaks at values of 25.3°, 37.8°, 48.1°, 54.4°, 62.9°, 69.9°, and 74.9° in the pure TiO₂ and TiO₂/rGA patterns can be respectively indexed to the (101), (004), (200), (211), (204), (220), and (215) planes of the TiO₂ anatase phase (JCPDS-21-1272). Based on the Scherrer equation, the crystallite sizes of TiO₂ in the pure TiO₂ and TiO₂/rGA samples were 11 and 14 nm, respectively, via assuming a shape factor of 0.9. Graphene and GO can hardly be seen in TiO₂/rGA due to the low content and the decreased layer-regularity of graphene nanosheets.¹⁶

Fig. 2 XRD patterns of the samples.

The surface components and chemical states of TiO₂/rGA were investigated via XPS. Fig. 3a exhibits the full spectrum of TiO₂/rGA, which demonstrates the coexistence of Ti, O, and C. The O 1s spectrum of TiO₂/rGA is exhibited in Fig. 3b. The O 1s XPS spectrum can be divided into three peaks at 530.5, 532.25 and 533.55 eV, representing Ti–O–Ti, C–O–Ti, and C–O, respectively, in TiO₂/rGA. In Fig. 3c, the Ti 2p peaks, located at binding energies of 459.2 and 465 eV, in the TiO₂ spectrum, with a separation of 5.8 eV, can be ascribed to pure anatase phase Ti⁴⁺.¹⁷ The peak at 472.7 eV is a satellite peak of Ti 2p_1/2. The Ti 2p peaks are positively shifted to 459.35 and 465.15 eV in TiO₂/rGA, which is caused by electrons migrating away from TiO₂ to graphene, because the Fermi level of TiO₂ is higher than that of graphene. A comparison of the C 1s spectra of GO and TiO₂/rGA is presented in Fig. 3d. The XPS peaks at 284.8, 286.3, and 288.95 eV can be assigned to the C–C/C=C, C=O, and O=C–O bonds of TiO₂/rGA and GO. The peak intensities of the oxygen-containing groups were reduced after the hydrothermal reaction, indicating that GO had been reduced. Notably, a few C=O and O=C–O bonds can still be found in TiO₂/rGA.

Fig. 3 XPS spectra: full spectrum of TiO₂/rGA (a), O 1s spectrum of TiO₂/rGA (b), Ti 2p spectra of TiO₂/rGA and of TiO₂ (c), and C 1s spectra of TiO₂/rGA and GO (d).

Fourier transform infrared spectra of the samples are shown in Fig. 4. The characteristic peaks from carbonyl C=O stretching vibrations, molecular water or C=C vibrations, C–OH bending vibrations, and C–O stretching vibrations or epoxy C–O–C vibrations in GO appear at 1734, 1633, 1200, and 1080 cm⁻¹, respectively. The abundant oxygen-containing groups on the surface indicate the good hydrophilicity of GO,¹⁸ easing its dispersal in water and bonding with other elements. During hydrothermal treatment, overheated supercritical water can play the role of a reducing agent, leading to decreases in the intensities of C=O stretching vibrations and C–O/C–O–C vibrations in TiO₂/rGA. The changes show the reduction of GO, as shown by XPS, and previous research produced similar results.¹⁹ The C–OH bonds exhibit more intense peaks in TiO₂/rGA than GO due to interactions between graphene and hydroxyl groups after the hydrolysis of titanous sulfate. The wide peak between 500 and 900 cm⁻¹ could be ascribed to Ti–O–Ti and Ti–O–C, confirming the chemical interaction between TiO₂ and graphene.^20,21 Comparing TiO₂/rGA with different GO content levels, the C–OH and C–O/C–O–C peaks at 1200 and 1080 cm⁻¹ in TiO₂/rGA-3 are relatively high. This is due to a combination of the hydrolysis of titanous sulfate and the reduction of GO, which vary at different GO content levels and for different structures.

Fig. 4 FTIR patterns of the samples.

Fig. 5a presents the Raman spectra of GO, rGA, TiO₂, and TiO₂/rGA. The peaks at about 1350 cm⁻¹ and 1590 cm⁻¹ represent the D band and G band of graphene, respectively. In detail, the D band reflects edge carbon or the disordered sp³ structure,²² while the G band represents the ordered sp² carbon structure. With a higher I_D/I_G ratio, the average size of the in-plane sp² domains and the conductivity of the composite will decrease, while the degree of reduction becomes higher. The I_D/I_G ratios of GO, rGA, and TiO₂/rGA-3 are 0.89, 1.04, and 0.91, respectively. The increased I_D/I_G ratio of rGA confirms the reduction of GO, as well as the presence of sp³ defects.^18,23 Furthermore, bare TiO₂ exhibits five intensive peaks from anatase. The peaks at 144 and 636 cm⁻¹ can be assigned to the E_g mode of TiO₂ (symmetric stretching vibrations of O–Ti–O bonds), and the peaks at 396 cm⁻¹ and 517 cm⁻¹ correspond to the B_1g mode of TiO₂ and an A_1g/B_1g doublet (symmetric and asymmetric bending vibrations of O–Ti–O).⁷ As seen in the enlarged view shown in Fig. 5a, the peak of TiO₂ at 144 cm⁻¹ is shifted positively to 150 cm⁻¹ in TiO₂/rGA, which indicates that the symmetric O–Ti–O stretching vibration in TiO₂ has been destroyed to some extent due to new bonds being formed between TiO₂ and graphene. As shown in Fig. 5b, the I_D/I_G values of TiO₂/rGA-1, TiO₂/rGA-2, TiO₂/rGA-4, and TiO₂/rGA-5 were all greater than 1, with only TiO₂/rGA-3 being less than 1. TiO₂/rGA-3 has the lowest reduction degree and some oxygen containing groups remain on the surface, which is similar to the FTIR spectra results.

Fig. 5 (a) Raman spectra of rGA, GO, TiO₂, and TiO₂/rGA. (b) Raman spectra of TiO₂/rGA with different amounts of graphene.

3.2. Microstructure and texture properties

The morphologies and constitutions of GO, rGA, TiO₂, and TiO₂/rGA were observed clearly via SEM and EDS studies, as seen in Fig. 6. As shown in Fig. 6a, the surface of GO is wrinkled and cross-linked, which can provide enough active sites. After the hydrolysis reaction, the three-dimensional porous structure of rGA can be seen in Fig. 6b. The pure TiO₂ nanoparticles are seen to clump together in Fig. 6c. The SEM images of TiO₂/rGA are quite different due to the different amounts of graphene. The morphologies of TiO₂/rGA-1 and TiO₂/rGA-2 are close to that of titanium dioxide, as shown in Fig. 6d and e. Upon increasing the graphene content, the skeleton of the graphene aerogel can be clearly seen in the next three images, and the TiO₂ nanoparticles are evenly coated on graphene due to the crosslinking effect of glucose and chemical interactions between TiO₂ and graphene. As shown in Fig. 6i, the elements Ti, O, and C can be detected in the EDS spectrum.

Fig. 6 SEM images of (a) GO, (b) rGA, (c) TiO₂, and (d–h) TiO₂/rGA, and the EDX spectrum (i) of TiO₂/rGA.

The N₂ adsorption–desorption isotherm curves for TiO₂/rGA and TiO₂ are shown in Fig. 7, and the corresponding data are summed up in Table 1. TiO₂ displayed a type-IV physisorption isotherm (based on IUPAC classification) with a H2 hysteresis loop. The pores are mainly in the shape of an ink bottle, and they are possibly caused by the accumulation of homogeneous titanium dioxide. TiO₂/rGA displayed type-IV physisorption isotherms, with a combination of H2 and H3 hysteresis loops. The hysteresis loops are triangularly shaped, mainly due to pores arising from aggregated nanoparticles and slit-shaped mesopores due to the stacking of rGO sheets.²⁴ The hysteresis loops of TiO₂/rGA are closer to TiO₂-type at low graphene content and closer to graphene-type at high graphene content, which agrees with the SEM results. From the BJH spectra, it can be seen that TiO₂ and TiO₂/rGA showed uniform aggregated mesopores in the range of 3–8 nm. The specific surface areas of the samples are shown in Table 1. TiO₂/rGA-1 and TiO₂/rGA-2 exhibited higher specific surface areas than pure TiO₂ and rGA, with surface areas of up to 116 m² g⁻¹ and 140 m² g⁻¹, respectively. The addition of graphene in small amounts contributed to the overall specific surface area and the crystallite growth. When the graphene content was higher, the surface area sharply decreased to 13 m² g⁻¹. As shown in Fig. 6, the morphology of TiO₂/rGA changed from granular to lamellar. Compared with Fig. 6b, the graphene sheets are stacked together and more closely linked to titanium dioxide.

Fig. 7 BET spectra of (a) TiO₂/rGA-1, TiO₂/rGA-2, and TiO₂, and (b) TiO₂/rGA-3, TiO₂/rGA-4, TiO₂/rGA-5, and rGA.

Table 1 Specific surface area and pore structure values obtained via the multipoint BET method

Composite	S BET (m^2 g^−1)	V pore (cm^3 g^−1)	D pore (nm)
rGA	97.5722	0.081287	1.63237
TiO2	34.6566	0.105611	4.58944
TiO2/rGA-1	116.7444	0.185195	4.7602
TiO2/rGA-2	140.3755	0.177772	4.2374
TiO2/rGA-3	13.0532	0.024687	5.1295
TiO2/rGA-4	32.5060	0.051381	4.8364
TiO2/rGA-5	38.9673	0.067024	5.8281

---

3.3. Optical properties and band gaps

The PL intensity can reflect the photogenerated hole and electron separation efficiency.⁶ The PL spectra of TiO₂ and TiO₂/rGA are shown in Fig. 8. There are strong emission peaks for TiO₂ at around 460 nm and 425 nm, which can be ascribed to self-trapped excitons in TiO₂.²⁵ TiO₂ and TiO₂/rGA show similar PL signal shapes, however, the TiO₂/rGA signals were obviously weaker than that of TiO₂. This is mainly because the addition of graphene can decrease the radiative recombination of carriers at the rGO–TiO₂ interface, henceforth quenching the PL intensity. The excited electrons are transferred from the conduction band of TiO₂ to the rGO sheets, prolonging the carrier lifetimes.

Fig. 8 PL spectra of TiO₂/rGA-1, TiO₂/rGA-2, TiO₂/rGA-3, TiO₂/rGA-4, TiO₂/rGA-5, and TiO₂.

3.4. The photocatalytic degradation of UDMH under UV light

3.4.1 Humidity. Photocatalytic experiments involving the gas-phase degradation of UDMH were performed at room temperature (25 °C) and under ambient pressure. UDMH conversion over bare TiO₂, rGA, TiO₂/rGA-1, TiO₂/rGA-2, TiO₂/rGA-3, TiO₂/rGA-4, and TiO₂/rGA-5 at different humidity levels (a: dry air; b: 50% H₂O) under UV light is shown in Fig. 9. The UDMH conversion levels when using TiO₂/rGA lie in the range of 19–28%, which are higher than the reference catalyst of pure TiO₂ in dry air. Pure rGA shows better performance than TiO₂/rGA-1 and TiO₂/rGA-2 due to its high absorption capacity. The relatively high efficiency of TiO₂/rGA is attributed to a synergetic effect between photocatalysis and absorption. On one hand, TiO₂ can produce electron–hole pairs under UV irradiation, and graphene can efficiently transport the carriers, as verified by the Ti 2p XPS spectra. On the other hand, the Schottky junction formed between TiO₂ and rGA contributes to the degradation of UDMH. The Schottky barrier serves as an efficient electron trap, thus leading to a high partial electron density around graphene for photocatalysis and preventing the unexpected migration of electrons from graphene back to the semiconductor.²⁶ In addition, the upshifted Fermi level of TiO₂ under UV light irradiation could decrease the TiO₂–rGA barrier and benefit electron transport.²⁷

Fig. 9 UDMH conversion under UV light in (a) dry air and (b) air at 50% humidity.

When the humidity was raised to 50%, the UDMH conversion of TiO₂/rGA-3 can reach 68%, which is much higher than in dry air. More reactive oxygen species can be produced in humid air. As a strong oxidizing agent, hydroxide radicals may be generated on the surface of TiO₂. The UDMH conversion levels of TiO₂/rGA with different GO content vary greatly, following the order: TiO₂/rGA-3 > TiO₂/rGA-4 > TiO₂/rGA-5 > TiO₂/rGA-2 > TiO₂/rGA-1 > rGA > TiO₂. The material with the highest efficiency is TiO₂/rGA-3, with a relatively low SSA and moderate carrier density based on the previous characterization results. From the obtained PL spectra, we can infer that TiO₂/rGA-5 has a relatively high catalytic effect, as it had the lowest fluorescence intensity. However, the degradation efficiency of TiO₂/rGA-5 is lower because the degradation of UDMH under UV light is affected by a combination of absorption and catalysis. In order to analyse the degradation mechanism, the FTIR patterns of fresh and used TiO₂/rGA-3 are shown in Fig. 10. The characteristic peaks of C=O, C–OH, and C–O–C decreased after degradation, and two new peaks due to C–H and N–H appeared. Degradation may arise due to chemical interactions between reducing UDMH and oxygen-containing groups. As shown in Fig. 4 and 5, the reduction degree of TiO₂/rGA-3 is weaker and the number of oxygen-containing groups on TiO₂/rGA-3 is higher than the other samples. It can be deduced that the increased activity of TiO₂/rGA-3 can mainly be accredited to chemical absorption on oxygen-containing groups.

Fig. 10 The FTIR patterns of fresh and used TiO₂/rGA.

3.4.2 Temperature. Fig. 11 demonstrates the effects of temperature on the degradation of UDMH in the range of 20–140 °C. UDMH conversion decreased as the temperature increased from 20 to 80 °C. The results suggest that UDMH adsorption on TiO₂/rGA is an exothermic process. Elevating the temperature can promote molecular diffusion, enhancing the adsorption rate.²⁸ Cation–π bonding and hydrogen bonding between UDMH and rGO also contribute to absorption.²⁹ When the temperature was further increased, the UDMH conversion to other organics slightly increased. These results are similar to a previous report.²⁸

Fig. 11 The effects of temperature on UDMH conversion under UV light.

3.4.3 Reaction mechanism and the proposed degradation pathway of UDMH under UV light. The UDMH degradation mechanism includes both chemical adsorption and a photocatalytic process. Fig. 12 schematically shows the degradation mechanism on TiO₂/rGA. Because most of the oxygen-containing groups contribute to the adhesion of hydrophilic VOCs onto the carbon surface, the adsorption of UDMH on graphene can provide more opportunities for TiO₂ to come into contact with UDMH. This allows further photocatalysis on TiO₂/rGA. Electron–hole pairs can be generated by TiO₂ when it absorbs photons with energy higher than the intrinsic bandgap energy. Graphene can effectively transfer the photogenerated electrons and prevent the recombination of h⁺ and e⁻ on TiO₂. Holes on the surface of TiO₂ can rapidly react with water to form ˙OH. Also, oxygen radical will be generated upon the interaction of electrons with O₂. The reactive oxygen species (ROS) can then react with UDMH.

Fig. 12 A schematic diagram of the photocatalytic degradation mechanism on TiO₂/rGA.

In order to further understand the possible UDMH transformation path, different intermediate products observed via GC-MS are listed in Fig. 13 and Table 2. The organic products vary considerably under different conditions. In dry air, the main components other than UDMH are formaldehyde dimethylhydrazone (FDMH), dimethylamine (DMA), and nitrosodimethylamine (NDMA).

Fig. 13 Gas chromatograms of the degradation products of UDMH under UV light.

Table 2 Products observed during the reaction of 1,1-dimethyldrazine with TiO₂–rGA under UV light

Product	Peak/min	CAS no.	Dry	50% H2O
FDMH	2.2	2035-89-4	+	+
UDMH	2.42	57-14-7	+	+
Butanedial	3.09	638-37-9		+
DMA	3.77	124-40-3	+	+
NDMA	5.19	62-75-9	+	+

---

The high yields of FDMH observed during the transformation of UDMH were initiated by H-abstraction from C–H bonds of UDMH. Firstly, –CH₂˙ is formed in the presence of oxygen. Then, oxygen inserts into R–CH₂˙ to form HCHO and methylhydrazine (MMH) (eqn (1)). Lastly, FDMH is formed when UDMH reacts with HCHO³⁰ (eqn (2)). MMH can further be attacked by ozone to form methyl radicals (eqn (5)). The formed methyl radicals account for the observed DMA groups. When the humidity of UDMH gas is adjusted to 50%, a new peak appeared at 3.09, which was correspondingly assigned to butanedial. The formation of butanedial may involve the participation of hydroxide radicals. The hydroxide radicals can react with methyl radicals (eqn (5)) to form aldehydes.

R₂–NH₂ + HCHO → R₃–N=CH₂ + H₂O (2)

O₂ + O˙ → O₃ (3)

In eqn (1) and (2), R₁ refers to (CH₃)NNH₂; R₁H refers to MMH; R₂ refers to (CH₃)(CH₃)N.

3.5. The photocatalytic degradation of UDMH under VUV light

3.5.1 Humidity. The UDMH conversion levels under VUV light in both dry and moist air are about 95%, as shown in Fig. 14. The differences in UDMH conversion are negligible at different GO content levels. Presumably, most UDMH molecules were degraded in the gas phase and not on the photocatalyst surface.³¹ The VUV light used in the experiment, with a wavelength of 185 nm, generates energetic photons that can activate oxygen and water vapor to produce numerous reactive species, such as O(1D), O(3P), hydroxyl radicals (˙OH), and O₃.³² VUV photolysis has been used to destroy VOCs like benzene and formaldehyde.^33,34 The bonds of UDMH are initially destroyed by VUV photolysis and are then oxidized by residual O₃ from photolysis. Due to the abundant reactive species generated in the gas phase, the oxidation degree of UDMH increases and the number of intermediates increases. When water was added, the UDMH conversion increased slightly, as more hydroxyl radicals (˙OH) were generated in the system.

Fig. 14 UDMH conversion under VUV light in (a) dry air and (b) air at 50% humidity.

3.5.2 Temperature. The UDMH conversion at different temperatures at 50% humidity is shown in Fig. 15. With an increase in temperature, the UDMH conversion is almost unchanged. Absorption is an exothermic reaction under UV light, while UDMH conversion increased slightly under VUV light. Therefore, adsorption does not play the main role and temperature has little effect on UDMH conversion, while photocatalysis and ozonation play the main role in the process.

Fig. 15 The effects of temperature on UDMH conversion under UV light.

3.5.3 Reaction mechanism and the proposed degradation pathway of UDMH under VUV light. The possible reaction mechanism of UDMH transformation is reliant on direct photolysis and photocatalysis reactions between UDMH and photogenerated ROS under VUV light. Different intermediate products are shown in Fig. 16, and the product details observed via GC-MS are listed in Table 3. The proportion of UDMH converted to other organics is almost 95%, much higher than under UV light. One of the degradation products, NDMA, has toxic mutagenic, teratogenic, and carcinogenic properties. It is noted that the yield of NDMA is obviously more than under UV light. Because more ozone can be produced under VUV light, the N–H bonds of UDMH could easily be attacked by ozone, and H-abstraction reactions can take place. Then, ozone attacks intermediates containing NH˙ free radicals to form NDMA³⁰ (eqn (6)). NDMA can also be synthesized via combining FDMH with ozone (eqn (7)). The high yields of NDMA are confirmed in EC Tuzon's study.³⁵ FDMH can also react with oxygen to produce (CH₃˙)₂NN and (CH₃˙)₂NN, which quickly combine to form 1,1,4,4-tetramethyl-1,2-tetrazene (TMT). Furthermore, NDMA and dimethylnitramine (NTDMA) can be formed in the presence of TMT upon adding ozone (eqn (8)). When water is involved in the reaction, the N=CH₂ species in FDMH will convert to ˙CN and then react with ˙CH₃ and UDMH to form (dimethylamino)acetonitrile.

Fig. 16 Gas chromatograms of the degradation products of UDMH under VUV light.

Table 3 Products observed during the reaction of 1,1-dimethyldrazine with TiO₂/rGA under VUV light

Product	Peak/min	CAS no.	Dry	50% H2O
FDMH	2.2	2035-89-4	+	+
UDMH	2.42	57-14-7	+	+
Butanedial	3.09	638-37-9		+
DMA	3.77	124-40-3	+	+
(Dimethylamino)acetonitrile	4.34	926-64-7		+
NDMA	5.19	62-75-9	+	+
Dimethylnitramine (NTDMA)	7.8	4164-28-7	+	+

---

3.6. Reusability of TiO₂/rGA under UV light and VUV light

TiO₂/rGA-3 was selected to investigate the stability and reusability of samples under UV light and VUV light. The sample was exposed to flowing UDMH gas for 3 h. Then it was washed three times in deionized water and dried at 110 °C prior to the next reaction cycle. As shown in Fig. 17, the UDMH conversion slightly decreased in UV and VUV light during the first cycle, due to partial sample loss. After five cycles, the samples still show stable UDMH conversion levels, implying good reusability during the degradation of UDMH.

Fig. 17 The stability of a typical TiO₂/rGA-3 sample under (a) UV and (b) VUV light.

4. Conclusions

In summary, a series of TiO₂/rGA samples was prepared using titanium sulfate and GO as precursors. The morphologies and number of surface oxygen-containing groups of the TiO₂/rGA samples varied considerably upon the addition of different amounts of GO, as determined through characterization studies. TiO₂/rGA-3 exhibited extraordinary UDMH degradation performance under UV light. Moist conditions were conducive to the removal of gas-phase UDMH, and absorption on TiO₂–rGA was an exothermic process. UDMH conversion could reach about 68% and 95% at 50% humidity under UV and VUV light, respectively. Compared with pure TiO₂ and graphene, chemical interactions and the Schottky junction formed in TiO₂–rGA accounted for the high charge separation and transmission efficiencies, which are helpful for improving the catalytic performance. TiO₂/rGA-3 showed higher UDMH conversion under UV light, which was mainly due to chemical absorption between reducing UDMH and oxygen-containing groups. The mechanism of UDMH degradation under VUV light was attributed to photocatalysis and ozonation. The high UDMH conversion in a continuous fixed bed reactor shows that this photocatalyst system is useful for removing gas-phase UDMH in practice.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This project is supported by the National Natural Science Foundation of China (21875281).

Notes and references

1. X. Gao, X. Liu, Z. Zhu, Y. Gao, Q. Wang, F. Zhu and Z. Xie, Sci. Rep., 2017, 7, 973–982.
2. J. Yang, Q. Shi, R. Zhang, M. Xie, X. Jiang, F. Wang, X. Cheng and W. Han, Carbon, 2018, 138, 118–124.
3. J. Wang, G. Zhang and P. Zhang, Appl. Catal., B, 2018, 239, 77–85.
4. L. Tian, L. Xing, X. Shen, Q. Li, S. Ge, B. Liu and L. Jie, Atmos. Pollut. Res., 2020, 11, 179–185.
5. A. J. Jafari, R. R. Kalantary, A. Esrafili and H. Arfaeinia, Process Saf. Environ. Prot., 2018, 116, 377–387.
6. H. Ye, Z. Wang, Y. Liu, S. Chen, H. Wang and Z. Wu, Catal. Lett., 2019, 149, 2739–2748.
7. J.-J. Li, B. Weng, S.-C. Cai, J. Chen, H.-P. Jia and Y.-J. Xu, J. Hazard. Mater., 2018, 342, 661–669.
8. W. Szeto, J. Li, H. Huang and D. Y. C. Leung, Chem. Eng. Sci., 2018, 177, 380–390.
9. X. Peng, M. Wang, F. Hu, F. Qiu, H. Dai and Z. Cao, J. Alloys Compd., 2019, 770, 1055–1063.
10. J. Vargas Hernández, S. Coste, A. García Murillo, F. Carrillo Romo and A. Kassiba, J. Alloys Compd., 2017, 710, 355–363.
11. P. Tan, A. Zhu, L. Qiao, W. Zeng, H. Cui and J. Pan, J. Colloid Interface Sci., 2019, 533, 452–462.
12. K. Alamelu and B. M. Jaffar Ali, New J. Chem., 2020, 44, 7501–7516.
13. A. Trapalis, N. Todorova, T. Giannakopoulou, N. Boukos, T. Speliotis, D. Dimotikali and J. Yu, Appl. Catal., B, 2016, 180, 637–647.
14. A. Ebrahimi and S. Fatemi, Clean Technol. Environ. Policy, 2017, 19, 2089–2098.
15. W. S. Hummers Jr and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
16. X. Chen, Q. Chen, W. Jiang, Z. Wei and Y. Zhu, Appl. Catal., B, 2017, 211, 106–113.
17. C. Peng, W. Wang, W. Zhang, Y. Liang and L. Zhuo, Appl. Surf. Sci., 2017, 420, 286–295.
18. N. Shehzad, M. Tahir, K. Johari, T. Murugesan and M. Hussain, Appl. Surf. Sci., 2019, 463, 445–455.
19. S. Pei and H.-M. Cheng, Carbon, 2012, 50, 3210–3228.
20. A. J. Wang, W. Yu, Y. Fang, Y. L. Song, D. Jia, L. L. Long, M. P. Cifuentes, M. G. Humphrey and C. Zhang, Carbon, 2015, 89, 130–141.
21. H. Zhang, X. Lv, Y. Li, Y. Wang and J. Li, ACS Nano, 2010, 4, 380–386.
22. K. N. Kudin, B. Ozbas, H. C. Schniepp, R. K. Prud’homme, I. A. Aksay and R. Car, Nano Lett., 2008, 8, 36–41.
23. Y. Fan, W. Ma, D. Han, S. Gan, X. Dong and L. Niu, Adv. Mater., 2015, 27, 3767–3773.
24. M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier, F. Rodriguez-Reinoso, J. Rouquerol and K. S. W. Sing, Pure Appl. Chem., 2015, 87, 1051–1069.
25. S. Shuang, R. Lv, Z. Xie and Z. Zhang, Sci. Rep., 2016, 6, 26670.
26. X. Li, R. Shen, S. Ma, X. Chen and J. Xie, Appl. Surf. Sci., 2018, 430, 53–107.
27. J. Liu, Z. Wang, L. Liu and W. Chen, Phys. Chem. Chem. Phys., 2011, 13, 13216–13221.
28. X. Zhang, B. Gao, A. E. Creamer, C. Cao and Y. Li, J. Hazard. Mater., 2017, 338, 102–123.
29. M. Maiti, M. Sarkar and D. Liu, Catal. Sci. Technol., 2020, 10, 2797–2809.
30. D. Huang, X. Liu, Z. Xie, X. Wang, X. Gao and Y. Yang, Symmetry, 2018, 10, 394–406.
31. T. Z. Xu, H. Zheng and P. Y. Zhang, Build. Environ., 2018, 142, 379–387.
32. T. M. Hashem, M. Zirlewagen and A. M. Braun, Water Sci. Technol., 1997, 35, 41–48.
33. H. Huang, H. Huang, Y. Zhan, G. Liu, X. Wang, H. Lu, L. Xiao, Q. Feng and D. Y. C. Leung, Appl. Catal., B, 2016, 186, 62–68.
34. P. F. Fu, P. Y. Zhang and J. Li, Int. J. Photoenergy, 2012, 2012, 174862.
35. E. C. Tuazon, W. P. L. Carter, A. M. Winer and J. N. Pitts, Environ. Sci. Technol., 1981, 15, 823–828.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0nj04617e

---