SO₂ disproportionation for a sulphur-based thermochemical cycle studied in an operando Raman batch reactor

Abstract

In situ and operando Raman spectroscopy were used to study the iodide-catalysed SO₂ disproportionation reaction at temperature of 118 °C and total pressures up to 9 bar. The disproportionation of SO₂ is one of the three process steps constituting a solar-aided thermochemical sulphur-based cycle for producing unlimitedly storable sulphur capable for on-demand consumption. A suitably designed quartz-made batch reactor cell was used enabling a special containment of SO₂, capable for in situ monitoring of the reaction progress in the liquid I⁻/H₂O/SO₂ phase as well as for quantitative monitoring of the SO₂ pressure in the vapors thereof. The iodide catalyst content was varied in the 0.156–0.780 I⁻/H₂O mol% range. The incorporation of SO₂ into the I⁻/H₂O solution was facilitated by means of O₂S⋯I⁻ interactions resulting in formation of I(SO₂)_x⁻ adduct species. The rate of SO₂ consumption was accelerated with increasing I⁻ content and 5% average hourly rates of SO₂ consumption could be maintained after 6 h of reaction time. The mechanistic pathway of the iodide-catalysed SO₂ disproportionation was unravelled at the molecular level. Below a certain SO₂ threshold pressure and at high I⁻ content, formation of the undesired I₂ by-product takes place, which is known to severely complicate the post-batch product separation. The results offer insight into the SO₂ disproportionation and are discussed with relevance to implications for its technical integration into the sulphur based solar-aided thermochemical cycle.

---

1. Introduction

Concentrated solar power (CSP) technology, in combination with thermal energy storage (TES), helps to secure energy supply, reduce carbon emissions, and achieve sustainable development goals.^1–4 Special attention is addressed towards renewable energy CO₂-free conversion of solar energy into medium-to-high temperature heat and/or chemicals. To this end, a three-step solar-aided thermochemical cycle based on sulphur has been suggested by General Atomics, converting solar energy into storable elemental sulphur.³ Stored sulphur can be consumed on demand, hence the so-stored energy can be then converted back into high-temperature heat for power generation, significantly enhancing the efficiency, dispatchability, and cost-effectiveness of CSP plants.⁵ The processes comprising the cycle include:

(a) the decomposition of sulphuric acid followed by SO₃ splitting

H₂SO₄(aq) → H₂O(g) + SO₃(g) (450–500 °C) (1a)

SO₃(g) → SO₂(g) + ½O₂(g) (650–1000 °C) (1b)

(b) the disproportionation of SO₂ with H₂O into elemental sulphur and sulphuric acid

3SO₂(g) + 2H₂O(l) → 2H₂SO₄(aq) + S(s,l) (115–170 °C) (2)

(c) the combustion of elemental sulphur to produce high-temperature heat and SO₂

S(l) + O₂(g) → SO₂(g) (500–1500 °C) (3)

The key parameters for SO₂ disproportionation include pressure, temperature, H₂O : SO₂ ratio and the catalyst system.³ In thermodynamic grounds, the driving force of the reaction is inversely proportional to temperature, while elevated pressures are required to drive the reaction to completion, and the H₂O : SO₂ ratio needs to be controlled to avoid impeding the process. Specifically, high amount of water lowers the SO₂ partial pressure, while inadequate water content increases sulphuric acid activity at equilibrium. To the contrary, for intermediate H₂O : SO₂ ratios, the SO₂ disproportionation achieves stoichiometric formation of products (elemental sulphur and sulphuric acid), since the activity of water is high enough to drive the reaction forward, while still allowing SO₂ to remain sufficiently concentrated in the system.

The pressure dependence of the reaction remains also a key challenge, as SO₂ disproportionation only occurs above a certain threshold pressure, which varies with temperature. Experimental and modelling studies³ have shown that with increasing temperatures, the reaction requires higher pressures to proceed efficiently. The temperature–pressure relationship is also critical, as operating at lower temperatures prevents unwanted side-reactions, e.g. H₂S formation, while increasing pressure drives forward the disproportionation reaction. An SO₂ pressure of ca. 40 bar has been found as adequate for SO₂ disproportionation to proceed smoothly in a homogeneous iodide-based catalytic batch reactor at 165 °C.³ However, this temperature represents the upper limit for practical continuous reactor systems due to the increased viscosity of sulphur above this threshold, complicating extraction and pumping.⁶

Among the catalyst systems currently in use, the homogeneous iodide-based catalysts are effective in increasing the reaction rate but require complex and costly separation/recovery processing steps.³ Catalyst recovery, especially in the case of unwanted side-product iodine formation, requires several separation-purification steps, making the process inefficient for industrial application. Significantly, an aspect that complicates the scale-up process is balancing the pressure requirements across different steps of the thermochemical sulphur-based cycle. The sulphuric acid decomposition step (1) benefits from lower pressures for optimal SO₃ splitting, but the SO₂ disproportionation reaction (2) is favoured at higher pressures for effective SO₂ conversion to sulphur and sulphuric acid. Hence, a compromise in operating pressures is sought, as using a gas compressor in a corrosive environment is a major technical challenge.

Current knowledge of the molecular mechanisms underlying SO₂ disproportionation remains limited. Although there has been extensive research on sulphur combustion and sulphuric acid decomposition, particularly in relation to hydrogen production and concentrating solar power (CSP) applications,^7–9 the specific process of disproportionation – a fundamental reaction in the solid sulphur-based thermochemical cycle – has only been explored in laboratory scale.³ In an attempt for fundamental understanding of the disproportionation of aqueous sulphur dioxide into sulphuric acid and elemental sulphur at room temperature,¹⁰ chemical tests and visual observations of colour change were employed, focusing on the catalytic role of iodide ions on the reaction rate. According to this study,¹⁰ the spontaneous disproportionation of SO₂ is thermodynamically feasible, but the reaction proceeds at an imperceptible rate without a catalyst. It is proposed that the role of the iodide ion is one of acting as an electron-transfer medium, enabling redox reactions between the sulphur species. Briefly, iodide ions likely form complexes with SO₂ (clathrates), thereby creating intermediate compounds lowering the process activation energy. Once the I⁻ transfers its electrons to the adsorbed SO₂, virtual iodine (I₃⁻) forms, which is immediately consumed by excess SO₂. This process regenerates the iodide, allowing the catalytic cycle to continue.

The present work aims to explore the mechanism of the homogeneous iodide (I⁻/H₂O)-catalysed SO₂ disproportionation. In situ Raman spectroscopy is applied under controlled conditions at temperatures up to 125 °C, SO₂ partial pressures, P_SO2, up to 7 bar and total pressures up to 9.5 bar for monitoring the SO₂ consumption in the vapours and the reaction evolution in the liquid reacting system. In pursuit of this, a special containment technique has been conceived and developed to enable the spectroscopic monitoring of the process in an operando neoteric Raman batch reactor.

2. Experimental section

2.1 Materials, containment and sample preparation.

Sodium iodide, NaI, was from Fluka AG (p.a.). Sulphur dioxide (anhydrous, 99.98%) was from Union-carbide and nitrogen (99.999%) was from L’Air Liquide. Triply distilled water was used for preparing the liquid mixtures. Based on the documented results of a previous work, a H₂O/SO₂ molar ratio of 2.6 was used for ensuring smooth reaction progression.³

The reactor cell, shown schematically in Fig. 1(A) was made of quartz tubing and was suitably designed for in situ Raman monitoring of the SO₂ signal in the vapours and of the reaction evolution in the condensed phase (liquids, solid). The cell possessed a main bulb (∼3.0–3.5 cm long, 20 mm o.d. and 17 mm i.d.) and a bottom appendix (∼1 cm long, 6 mm o.d. and 4 mm i.d.) with a total volume of 7.3 cm³. The stem used for loading the quartz cell was ∼10 cm long (6 mm o.d., 4 mm i.d.). The main bulb was designed in such a way as to contain sufficient number of SO₂ moles capable of creating and withstanding an ascertained pressure of ∼12 bar at ∼125 °C after complete vaporization of SO₂, whilst the appendix had a sufficient volume to accommodate the I⁻/H₂O condensed phase at an amount that fulfilled the requirement for the H₂O/SO₂ mole ratio of 2.6. The cell design, dimensions and wall thicknesses were adequate for ensuring structural integrity up to 12 bar, while providing sufficient laser irradiation path lengths for the vapour and liquid phases (17 and 4 mm, respectively) and keeping the entire sealed optical cell within the uniform temperature zone of the vertical cylindrical core optical furnace.^11,12 Hence monitoring of in situ Raman spectra for both the vapour and liquid phases was enabled, the latter achieved by suitably adjusting the vertical position of the cell inside the optical furnace.

Fig. 1 (A) Schematic diagram of the quartz-made reactor cell. The actual cell dimensions are described in Section 2.1; (B) diagram of the T-shape construction showing the reactor cell containing the NaI/H₂O solution attached to the vacuum line. Valves 1 and 2 are also shown. The volume of the shaded area is 62 mL. Gaseous SO₂ introduced therein at a known pressure measured by the pressure indicator (PI) is – after opening valve V2-condensed in the cell, which is immersed in liquid nitrogen. The cell is then sealed by a propane-oxygen torch.

2.1.1 Containment and sample preparation. The iodide/water (I⁻/H₂O) catalyst system can be formulated by using water soluble iodide salts, such as NaI, KI and NH₄I or HI. Previous laboratory tests³ have shown that there is no significance in the iodide provenance, hence NaI was utilized. Weighed amounts of NaI were added to 1 mL of triply distilled H₂O to match the desired I⁻/H₂O mol% catalyst content (Table 1) and subsequently 100 μL of the solution were introduced into the quartz cell, hence filling its bottom appendix. Afterwards, the cell was attached in a vacuum/gas-addition line shown in Fig. 1(B) and evacuated while keeping the appendix containing the aqueous iodide solution immersed in liquid nitrogen. Valve 2 was then closed, to isolate the cell from the vacuum line that remained in static vacuum by closing the main valve of the vacuum pump. SO₂ was then introduced in the vacuum line up to a desired measured pressure and valve 1 was subsequently closed, thereby enclosing in the shaded T-shape glass assembly, which was confined between valve 1 and valve 2 an amount of SO₂ gas at known volume and pressure. Valve 2 is then opened and the contained SO₂ is condensed in the cell appendix which is maintained at 77 K (immersed in liquid nitrogen). Finally, valve 2 is closed and the cell is sealed-off by a propane-oxygen torch while maintaining its bottom immersed in liquid nitrogen. The amount of SO₂ added in the cell was adequate for establishing an initial pressure of after complete vaporisation of SO₂. Hence, at ∼125 °C the initial total cell pressure was up to ∼9.5 bar, accounting also for the respective vapor pressure of H₂O (∼2.3 bar). Table 1 compiles the characteristics of the various cells made. A reference cell containing only SO₂/H₂O was also made for comparing the Raman spectra of the I⁻/H₂O/SO₂ liquid systems with the counterpart spectra of iodide – free aqueous SO₂ solutions. Additionally, standard cells containing only SO₂ at known pressure were made for calibrating the ν₁(SO₂) Raman band intensity and determining the SO₂ pressure in the vapor phase over the reacting liquid mixtures. Fig. S1 shows the reference SO₂ cell used.

Table 1 Characteristics for I⁻/H₂O/SO₂ samples (iodide fraction in catalyst phase, initial pressure of SO₂)

Sample contents	Iodide fraction (I^−/H2O), mol%	^a
a Determined experimentally by exploiting the Raman band intensities of the ν1(SO2) band.^13
H2O/SO2	0	6.0
I^−/H2O/SO2	0.156	5.2
I^−/H2O/SO2	0.312	5.9
I^−/H2O/SO2	0.469	6.3
I^−/H2O/SO2	0.780	6.7
SO2	—	8.2

---

All sealed cells containing liquid SO₂ were subjected to a 30-min crash test at 125 °C in a vertical cylindrical core furnace to confirm that they can withstand the elevated pressures (up to ∼9.5 bar) without an explosion. This procedure was necessary for preventing a cell explosion from happening inside the Raman optical furnace.

2.1.2 Visual features of samples. Aqueous solutions of SO₂ at room temperature (298 K), under the SO₂ vapor pressure of ∼3.5–4 bar, are transparent and colourless and maintain their transparent colourless feature until vaporisation of SO₂ at 125 °C (P_SO2 = ∼5–7 bar). To the contrary, colourless aqueous iodide solutions obtain a bright transparent yellow colour upon SO₂ dissolution and at high iodide contents, e.g. at I⁻/H₂O mol% >0.5, split into two liquid phases in equilibrium to each other, as shown in Fig. 2. The lower phase is transparent pale yellowish, and the upper phase is transparent dark yellow. As shown below (vide infra), the lower phase is SO₂-rich, whilst the upper phase is H₂O-rich and contains dissolved iodide and SO₂. Presumably, an electrostatic association between iodide anions and polar SO₂ molecules results in formation of I(SO₂)_x⁻ clathrate like species,¹⁰ hence sulphur dioxide incorporation into the liquid reacting system is promoted/facilitated by the iodide ion catalyst. Upon temperature increase to 118 °C the lower SO₂-rich phase disappears due to evaporation of SO₂.

Fig. 2 Actual photograph of the sealed cell containing the I⁻/H₂O/SO₂ liquid under gaseous SO₂ at room temperature (P_SO2,293K = ∼3.5–4 bar) with I⁻/H₂O = 0.780 mol%, featuring the liquid separation into two conjugate phases. The green line portrays iconically the 532 nm laser irradiation of (a) the lower (SO₂-rich) phase; (b) the upper (iodide solution – rich) phase; and (c) the gas phase.

2.2 In situ Raman spectra and optical batch reactor

For obtaining the Raman spectra for the studied samples, each cell was placed in the optical Raman furnace^11,12 and irradiated with the green 532.0 nm line of a linearly polarized diode-pumped solid-state laser (Spectra Physics Excelsior) operated at a power level of 100 mW at the sample. The scattered light was collected at 90° through an edge filter and analysed with a ISA-Horriba IHR-320 JY monochromator. The signal was detected by a −56 °C thermoelectrically cooled CCD detector interfaced with Labspec software with a resolution of 2 cm⁻¹. With respect to the polarization configuration of the scattered light relative to the scattering horizontal plane, both VV (vertical-vertical) and VH (vertical–horizontal) configurations were applied. The observed wavenumbers’ precision and depolarisation ratios were calibrated by recording spectra of a CCl₄ standard sample. Further details concerning the Raman spectroscopy set up are described earlier.¹⁴ Notably, the use of a 100-mW power for the incident laser radiation does not raise concerns for sample overheating, because of the continuous movement of the molecules comprising the liquid and/or the gaseous phase, thereby resulting in a continuous renewal of the sample of molecules being irradiated. Hence, in each instant the temperature of the irradiated sample can safely be taken equal to the temperature of the container, like if a kind of rotating-sample technique¹⁵ was utilised.

Each cell, before being inserted in the Raman optical furnace, was first tested for 30 minutes at 125 °C in a tube furnace to confirm that it can withstand the inner pressure. Initial spectra were recorded at 25 °C, followed by heating up to 118 °C, where Raman spectra were obtained for both the vapours and the liquid phase, and the reaction progress was monitored spectroscopically for ca. 10 h. After stopping the reaction or after reaction completion (the latter evidenced by iodine formation), the cell was cooled to 25 °C, and Raman spectra of the vapor and liquid phases were obtained. Cells where the reaction had not been completed were stored at room temperature and progress in the SO₂ disproportionation could be monitored by intermittently recording Raman spectra at room temperature for several weeks, thereby observing accumulation of solid sulphur product in the cell appendix bottom.

Raman spectroscopy can be used for determining the pressure of a gas (i.e. SO₂, P_SO2|_cell i) in a reaction mixture sample contained in a Raman cell at a temperature T_i by comparing the intensity of a band (integrated peak area, I_SO2|_cell i) representing the species of interest (i.e. ν₁(SO₂)) with the intensity (integrated peak area) of the same band in a reference cell (I_SO2|_R) containing only SO₂ at a known pressure, P_SO2|_R, and temperature, T_R.¹³ The calculation is based in the following formula¹³

where h, c and k are Planck's constant, the velocity of light and the Boltzmann's constant, respectively. The reference measurement took place several times for SO₂ vapors in equilibrium with liquid SO₂ at 16 °C (T_R = 289 K) and the corresponding SO₂ vapor pressure used was 2.93 bar (P_SO2|_R = 2.89 bar).¹⁶

Hence, the vapor Raman spectra were exploited for quantitative monitoring of the partial pressure of SO₂ above the liquid catalyst/reaction mixture. To this end, during the present work the spectral recording conditions and the optical geometry remained unchanged. The quartz Raman cells inserted in the Raman furnace were kept in fixed position relative to the optics. Indicatively, the SO₂-containing standard cells used for relating the ν₁(SO₂) band intensity in the gas phase to the partial pressure of SO₂ above the liquid catalyst/reaction mixture could be removed from the furnace, cooled and subsequently reintroduced in the furnace and the measured intensities were reproduced to within 5%.

3. Results and discussion

3.1 Raman spectra of the I⁻/H₂O/SO₂ liquid system at room temperature

Fig. 3 shows Raman spectra of the I⁻/H₂O/SO₂ liquid catalyst/reaction mixture at room temperature obtained for various I⁻/H₂O mol% values in the 0–0.780 range in equilibrium with gaseous SO₂ (see Experimental section). Polarised (VV) and depolarised (VH) spectra are displayed for commenting on the polarisation characteristics of the featured bands. In absence of iodide ions, i.e. in the H₂O–SO₂ liquid system, SO₂ is known to form various hydrates with water¹⁷ that slightly perturb its symmetry. Hence, as shown in spectrum Fig. 3(a), SO₂ maintains mutatis mutandis its symmetric structural configuration with a polarised band at 1155 cm⁻¹ due to the symmetric ν₁ stretching mode, which exhibits a depolarization ratio of ρ = ∼0.10, due to the perturbations caused within the solution. When iodide ions are present, part of the SO₂ molecules form I(SO₂)_x⁻ clathrate-like or adduct species via O₂S⋯I⁻ interactions,^10,18 in which the SO₂ moiety is distorted in such a way as to cause a slight elongation of the S–O bonds, thereby justifying a red shift, and a further perturbation of its symmetry, by that means affecting the polarisation characteristics of its main stretch. Remarkably, due to the O₂S⋯I⁻ interactions, the ν₁ mode shifts to 1144 cm⁻¹ and its depolarisation ratio is increased (see e.g. spectrum (d) in Fig. 3). Hence, with increasing I⁻/H₂O mol%, more SO₂ molecules participate in O₂S⋯I⁻ interactions, as seen in spectra (b)–(d) shown Fig. 3. Moreover, for I⁻/H₂O mol%, >0.500, the I⁻/H₂O/SO₂ liquid splits into two conjugate phases in equilibrium (see Fig. 2), of which the lower-density upper phase is an iodide solution-rich one with a Raman spectrum shown in Fig. 3(d) and the higher density lower phase is an SO₂-rich one with a Raman spectrum shown in Fig. 3 inset, which also includes the Raman spectrum of the lower density upper phase for comparison. It turns out that the incorporation of SO₂ into the liquid catalyst/reacting mixture is facilitated by the presence of I⁻ ions by means of O₂S⋯I⁻ interactions, as evidenced by the Raman spectra.

Fig. 3 Raman spectra obtained for I⁻/H₂O/SO₂ liquid mixtures in equilibrium with gaseous SO₂ at room temperature (P_SO2,293K = ∼3.5–4 bar) with I⁻/H₂O mol% ratios in the 0–0.780 range as indicated by each spectrum. The inset shows the Raman spectra of the two liquid phases in equilibrium for the liquid mixture with I⁻/H₂O = 0.780 mol%. Laser wavelength, λ₀ = 532.0 nm; laser power, w = 100 mW; resolution, 2 cm⁻¹.

3.2 Monitoring the iodide – catalysed homogeneous SO₂ disproportionation by in situ Raman spectroscopy of the liquid reaction mixture at 118 °C

Fig. 4 pertains to the liquid–gas I⁻/H₂O/SO₂(l)–SO₂(g) system with I⁻/H₂O = 0.156 mol% and shows two snapshots of the liquid Raman spectra. Trace Fig. 4(a) is obtained for the liquid mixture at 25 °C, i.e. before heating up the Raman furnace for initiating the reaction. Notably, prior to inserting in the Raman furnace, the quartz reaction cell has been subjected to a 30 min “crash test” at 125 °C. Hence, the strongest HSO₄⁻ mode at 1054 cm⁻¹ can already be discerned in spectrum Fig. 4(a), thereby indicating the initiation of the SO₂ disproportionation reaction, already during the 30 min “crash test”. Spectrum Fig. 4(b) is obtained at 118 °C after 9.75 h of reaction time and exhibits: (a) the ν₁(SO₂) mode, which is weakened due to vaporisation (thereby increasing the cell pressure) and consumption of SO₂ due to the disproportionation reaction; and (b) additional emerging bands at 1195 cm⁻¹ (weak), 1054 cm⁻¹ (strong), 878 cm⁻¹ (medium), 592 cm⁻¹ (medium) and 425 cm⁻¹ (medium) due to HSO₄⁻(aq)^19,20 and at 475 and 220 cm⁻¹ due to elemental sulphur²¹ produced as a result of the SO₂ disproportionation reaction.

Fig. 4 Raman spectra obtained for the I⁻/H₂O/SO₂ liquid mixture with I⁻/H₂O = 0.156 mol%: (a) at 25 °C, just after the ex situ 30 min crash test at 125 °C and insertion into the Raman furnace; (b) At 118 °C, after 9.75 h of reaction time. Recording parameters: see the Fig. 3 caption.

Fig. 5 shows the evolution of the SO₂ disproportionation reaction in the liquid phase, monitored in the in situ Raman batch reactor at 118 °C for two different contents of iodide catalyst. Panel Fig. 5(A) shows the evolution in the time span 0.75–9.75 h for I⁻/H₂O = 0.156 mol% and Panel Fig. 5(B) shows the respective evolution up to 6 h for I⁻/H₂O = 0.780 mol%. As evidenced from the intensity of the 1054 cm⁻¹ HSO₄⁻ band relative to the 1154 cm⁻¹ “SO₂” band envelope, the disproportionation reaction proceeds much faster when increasing the content of the iodide catalyst. Notably, spectra pertaining to “t = 0.75 h” in Fig. 5 exhibit in both panels bands due to the HSO₄⁻ produced during the 30 min “crash test” taken place at 125 °C before insertion of each reactor cell into the Raman furnace.

Fig. 5 Sequential in situ Raman spectra obtained for I⁻/H₂O/SO₂ liquid reacting mixtures at 118 °C in the S–O stretching wavenumber region after reaction time duration as indicated by each spectrum. (A) I⁻/H₂O = 0.156 mol%; (B) I⁻/H₂O = 0.780 mol%. Recording parameters: see the Fig. 3 caption.

The SO₂ disproportionation reaction proceeds in the liquid phase via formation of HSO₄⁻(aq), catalysed by I⁻ in two steps: one of a reduction and one of an oxidation according to the following mechanistic scheme, which is slightly modified compared to a previously proposed one:¹⁰

SO₂ “Dissolution”:

3SO₂(g) → 3SO₂(aq)

Reduction step:

SO₂(aq) + 6I⁻(aq) + 4H₃O⁺(aq) → S(s) + 2I₃⁻(aq) + 6H₂O (5)

Oxidation step:

2SO₂(aq) + 2I₃⁻(aq) + 10H₂O(l) → 2HSO₄⁻(aq) + 6I⁻(aq) + 6H₃O⁺(aq) (6)

Total disproportionation reaction:

3SO₂(g) + 4H₂O(l) → 2HSO₄⁻(aq) + S(s) + 2H₃O⁺(aq) (7)

A follow-up step pertaining to HSO₄⁻(aq) ionization converts part of HSO₄⁻(aq) to SO₄²⁻(aq) upon cooling¹⁸ from the reaction temperature of 118 °C to 25 °C where both ions remain in equilibrium:

2HSO₄⁻(aq) + 2H₂O(l) ↔ 2SO₄²⁻(aq) + 2H₃O⁺(aq) (8)

The I⁻ ion is Raman silent, whilst the short-living I₃⁻ iodide intermediate, which has a characteristic Raman stretching vibration at 114 cm⁻¹,²² is not detectable, as its extent of presence is below the detection limit. Hence, neither elemental iodine, I₂, nor its I₃⁻ form are detected as long as the reaction proceeds. Actually, I₃⁻ is consumed at the very instant of its formation. The Bunsen reaction

SO₂(g) + 2H₂O(l) + I₂(s,l) ↔ 2HI(aq) + H₂SO₄(aq) (9)

remains shifted to the right as long as the residual SO₂ pressure remains above a certain threshold value, thereby preventing the undesired formation of I₂. Interestingly, in the following section, when increasing the I⁻/H₂O mol% ratio in the 0.156–0.469 range, the reaction takes place with an increasing rate of SO₂ consumption up to 9.75 h of typically applied daily reaction time, and is thereafter stopped by cooling the furnace. However, when a 0.780 I⁻/H₂O mol% ratio is used, the reaction proceeds much faster and is terminated abruptly because reaction (9) shifts to the left and formation of iodine, I₂(l,g) and I₃⁻(aq) is detected in the in situ Raman batch quartz reactor (vide infra). The progress duration of the disproportionation reaction before the undesired iodine (I₂) appearance depends on the initial pressure of SO₂ in the batch reactor cell as well as the iodide (I⁻) content, in accordance with the Bunsen reaction (9).

3.3 Operando Raman monitoring of the SO₂ pressure decay in the gas phase during SO₂ disproportionation catalysed by I⁻/H₂O at 118 °C and reaction outcome

3.3.1 Low-to-medium iodide catalyst content. Fig. 6 portrays the vapor Raman spectra obtained over the reacting H₂O/SO₂/I⁻ liquid and shows the decay of the main gaseous SO₂ symmetric stretching mode band, ν₁(SO₂), vs time by showing sequential snapshots of the vapor Raman spectra in the t = 0.75–10 h span for the reaction batches that used 0.156 I⁻/H₂O mol% (Panel Fig. 6(A)), 0.312 I⁻/H₂O mol% (Panel Fig. 6(B)) and 0.469 I⁻/H₂O mol% (Panel Fig. 6(C)). The corresponding dependence of the SO₂ pressure as a function of reaction time for each reacting mixture in the 0.156–0.469 I⁻/H₂O mol% range, calculated using eqn (4), is displayed in plots (a)–(c) shown in Fig. 7(A). It is evident that the SO₂ pressure decay as a function of reaction time is steeper when higher amounts of iodide catalyst are present. The SO₂ pressure data can furthermore be exploited to calculate the percent average hourly SO₂ consumption rates and the results are displayed in plots (a)–(c) shown in Fig. 7(B). Whereas an average 1% hourly SO₂ consumption rate is calculated for the reaction batch with the lowest 0.156 I⁻/H₂O mol% content, initial hourly rates of 4.0 to 4.7% are achieved with 0.312 and 0.469 I⁻/H₂O mol%. Significantly, a ∼3% hourly SO₂ consumption rate is maintained after ca. 9 h of reaction for an iodide content of 0.469 I⁻/H₂O mol%.

Fig. 6 Sequential in situ Raman spectra obtained for the vapors over I⁻/H₂O/SO₂ liquid reacting mixtures in the 0.75–9 h reaction time span. (A) I⁻/H₂O = 0.156 mol%; (B) I⁻/H₂O = 0.312 mol%; (C) I⁻/H₂O = 0.469 mol%. Recording parameters: see the Fig. 3 caption.

Fig. 7 (A) Plots of the SO₂ pressure determined by exploiting the Raman band intensities of the ν₁(SO₂) mode (see experimental section) as a function of reaction time for the reactor batches with: (a) I⁻/H₂O = 0.156 mol%; (b) I⁻/H₂O = 0.312 mol%; (c) I⁻/H₂O = 0.469 mol%; I⁻/H₂O = 0.780 mol%. (B) Plots of the average percent hourly SO₂ consumption rates as a function of reaction time for the reactor batches with: (a) I⁻/H₂O = 0.156 mol%; (b) I⁻/H₂O = 0.312 mol%; (c) I⁻/H₂O = 0.469 mol%; I⁻/H₂O = 0.780 mol%.

In all three cases of low-to-medium iodide catalyst content (i.e. 0.156–0.469 I⁻/H₂O mol%) the SO₂ disproportionation reaction proceeds with SO₂ consumption and HSO₄⁻ production as shown in Fig. 5(A) pertaining to the low 0.156 I⁻/H₂O mol% content as well as formation of S(s, l) identified by the characteristic 153, 222 and 475 cm⁻¹ bands due to sulphur as seen in Fig. S2 portraying the full wavenumber range of the Raman spectrum obtained for the H₂O/SO₂/I⁻/HSO₄⁻ reacting liquid mixture at 118 °C. Notably, part of the sulphur occurs in solid form and is settled in the bottom cell appendix where temperature is very close to the melting point of 116 °C, hence whereas the reaction progress can be followed by the gradual increase of the HSO₄⁻ bands there is no commensurate increase of the bands due to sulphur. In all above cases (i.e. 0.156–0.469 I⁻/H₂O mol%) the reaction proceeds without formation of the undesired side product iodine, I₂, that could be formed by the reverse Bunsen reaction (9) if the relative amounts of SO₂, I⁻ and HSO₄⁻ would allow reaction (9) to be shifted to the left. Hence, SO₂ is consumed, as shown in Fig. 6(A)–(C) and HSO₄⁻ is produced (Fig. 5(A)) without iodine formation for 10 h of reaction time, after which the experiments were terminated and the Raman batch reactor cooled to room temperature. Visual inspection of the reactor cell after cooling reveals the occurrence of a yellow solid chunk precipitate at the bottom of the cell appendix, as shown in Fig. 8(A). Fig. 8(B) shows the reactor cell through the scattering window inside the Raman furnace, where the solid chunk glows being irradiated by the 532 nm laser at a very low power of 5 mW. Hence, to validate that the yellow solid chunk is the desired sulphur product, a Raman spectrum of the solid chunk contained inside the H₂O/SO₂/I⁻/HSO₄⁻ liquid mixture was recorded and shown in Fig. 8(C). The spectrum shown in Fig. 8(C) consists exclusively of the well-known Raman bands at 222, 246, 439 and 475 cm⁻¹ due to S(s).²¹

Fig. 8 (A) Photograph of the reactor cell with I⁻/H₂O = 0.156 mol% after 9 h of reaction time taken at room temperature. The yellow solid chunk can be seen at the bottom of the cell appendix containing the H₂O/SO₂/I⁻/HSO₄⁻ liquid mixture. (B) Photograph taken through the window of he Raman furnace featuring the solid yellow chunk glowing due to focused irradiation by the 532.0 nm laser beam at a power of 5 mW. (C) Raman spectrum obtained for the solid chunk contained in the H₂O/SO₂/I⁻/HSO₄⁻ liquid mixture at T = 30 °C. Laser wavelength, λ₀ = 532.0 nm; laser power, w = 5 mW; resolution, 2 cm⁻¹.

Interestingly, when the reactor cell is cooled to room temperature (after ca 10 h of reaction at 118–123 °C), the SO₂ disproportionation continues to proceed. This is quite reasonable on account of the thermodynamics of reaction (10):

3SO₂(g) + 4H₂O(l) → 2HSO₄⁻(aq) + S(s) + 2H₃O⁺(aq) (10)

ΔG⁰_298K = −136.99 kJ mol⁻¹, ΔH⁰_298K = −312.53 kJ mol⁻¹

where the negative Gibbs free energy implies that the reaction is thermodynamically favoured at room temperature, albeit at low intrinsic rates. This can be seen in Fig. 9, showing Raman snapshots obtained at room temperature for the reactor cell with 0.312 I⁻/H₂O mol% right after the 30 min crash test (trace (a)), after 9.75 h of reaction at 118 °C (trace (b)) and after 32 days on shelf (trace (c)), which shows a very high extent of reaction, i.e. by exhibiting strong bands due to HSO₄⁻, a prominent band due to the strongest (symmetric stretch) of SO₄²⁻ and strong bands due to S(s) suspended in the liquid phase.

Fig. 9 Raman spectra obtained for the I⁻/H₂O/SO₂ liquid mixture with I⁻/H₂O = 0.312 mol%: (a) at 25 °C, just after the ex situ 30 min crash test at 125 °C and insertion into the Raman furnace; (b) at 25 °C, after 9 h of reaction time at 118 °C; (c) at 25 °C, after a further shelf time of 32 days. Recording parameters: see the Fig. 3 caption.

3.3.2 High iodide catalyst content. When a higher catalyst content of 0.780 I⁻/H₂O mol% was used, the SO₂ pressure decay is even steeper as displayed in plot (d) shown in Fig. 7(A) and the pertinent percent hourly SO₂ consumption rates are shown in plot (d) in Fig. 7(B). Initial hourly rates of 6% could be achieved, decaying to 4.7% after ca. 7 h of reaction time. Significantly, as shown in Fig. 10(A), after 7 h of reaction time, the SO₂ signal fell abruptly and new strong bands emerged in the vapor Raman spectrum exhibiting vibrational-rotational structure, which is reminiscent of bands due to iodine vapors, I₂(g). The occurrence of the deep purple coloured iodine vapours justifies the abrupt fall of the SO₂ signal in plot (d) shown in Fig. 7(A). Indeed, recording of the Raman spectrum of vapors in the 150–1600 cm⁻¹ after 9 h resulted in the well-known resonance Raman spectrum of iodine vapors²³ showing a progression of multiple overtones of the 214 cm⁻¹ vibrational fundamental with each overtone exhibiting fine vibrational–rotational structure and the spectrum is shown in Fig. 10(B) (red trace (a)). By lifting the cell inside the Raman furnace and irradiating the liquid phase contained in the cell appendix, the Raman spectrum of the liquid mixture after 10 h could be recorded (blue trace (b) in Fig. 10(B)). The cell was then removed from the Raman furnace for inspection and the characteristic purple color of iodine vapors could be observed (Fig. 10(C)). Moreover, the solid sulphur chunk at the bottom of the cell appendix had a very dark purple color due to its interaction with iodine.

Fig. 10 (A) Sequential in situ Raman spectra obtained at 118 °C for the vapors over I⁻/H₂O/SO₂ liquid reacting mixture with I⁻/H₂O = 0.780 mol% in the 0.75–8.5 h reaction time span. Recording parameters: see the Fig. 3 caption. (B) Raman spectra obtained at 118 °C for: (a) the vapors over the H₂O/SO₂/I⁻/HSO₄⁻ liquid mixture with I⁻/H₂O = 0.780 mol% after 9 h of reaction time; (b) the H₂O/SO₂/I⁻/HSO₄⁻ liquid mixture after 10 h of reaction time. Recording parameters: see the Fig. 3 caption. (C) Photograph of the reactor cell with I⁻/H₂O = 0.780 mol% after 10 h of reaction time at 118 °C taken at room temperature. The purple iodine vapors fill the gas phase, whilst the product sulphur chunk is impregnated with iodine. (D) Raman spectra obtained at 25 °C for the H₂O/SO₂/I⁻/HSO₄⁻ liquid mixture with I⁻/H₂O = 0.780 mol% after 10 h of reaction time at 118 °C. Recording parameters: see the Fig. 3 caption.

The fast disproportionation of SO₂ when a high catalyst content of 0.780 I⁻/H₂O mol% is used, results in stoichiometrically commensurate production of HSO₄⁻, which in combination with the high I⁻ content shifts the Bunsen reaction (9) to the left, thereby accounting for formation of iodine. After cooling the reactor cell to room temperature, the dark liquid phase is largely decolorised due to combination of I⁻(aq) with iodine to form the triiodide ion I₃⁻(aq), as shown in Fig. 10(D), showing the Raman spectrum of the liquid over the solid S(s)/I₂(l) chunk where a band at 114 cm⁻¹ characteristic of the I₃⁻(aq) vibrational mode²² is observed. Additionally, strong features due to HSO₄⁻(aq) formed according to reaction (7) are observed together with the most characteristic ν₁(A₁) stretching mode of SO₄²⁻(aq) formed according to equilibrium (8) and the ν₁ mode of the SO₂ remaining in the liquid mixture.

3.4 Implications for the sulphur-based thermochemical cycle

An innovative methodology for understanding the mechanistic route for the benchmark homogeneous iodide-catalysed SO₂ disproportionation based on in situ molecular spectroscopy and for monitoring the process is established. By in situ monitoring of the process, in particular the SO₂ pressure, prevention of the undesired iodine by-product formation can be achieved. The laboratory results indicate that whereas an increase of the I⁻/H₂O ratio accelerates on one hand the reaction, a high I⁻ content combined with a low threshold SO₂ pressure may on the other hand lead to undesired I₂ formation. Intermittent SO₂ replenishment may prolong the progress of the reaction, thereby avoiding the technically challenging post-batch sulphur-iodine separation. The laboratory tests show that for a I⁻/H₂O ratio below 0.500 mol% a SO₂ pressure of 4.7 bar is adequate for preventing formation of I₂.

The hourly SO₂ consumption rates of 5% after 6 h of reaction time at the mild conditions of 118 °C and total pressures below 9 bar are quite promising for achieving a technically sought compromise between the sulphuric acid decomposition step (reactions (1a) and (1b)), which precedes the SO₂ disproportionation step (reaction (2)) in the solar-aided sulphur-based thermochemical cycle, thereby avoiding the use of a gas compressor in a corrosive environment.

4. Conclusions

An unprecedented operando Raman spectroscopic study of the benchmark process of homogeneous iodide-catalysed SO₂ disproportionation has been undertaken at a temperature of 118 °C and pressures up to 9 bar. A suitably designed neoteric quartz batch reactor cell was used, enabling in situ monitoring of the reaction progress in the homogeneous liquid reacting phase as well as in situ quantitative monitoring of SO₂ consumption in the vapors thereof.

The mechanistic pathway of the I⁻-catalysed SO₂ disproportionation under mild conditions, specifically at 118 °C and total pressure (gaseous SO₂ and H₂O vapors) of ∼9 bar is deciphered at the molecular level. The incorporation of SO₂ in the I⁻/H₂O solution is facilitated by means of I(SO₂)_x⁻ clathrate-like or adduct species formed via O₂S⋯I⁻ interactions in the liquid state. The SO₂ disproportionation reaction mechanism consists of two main steps, one of a reduction resulting in formation of sulphur and of a short living I₃⁻(aq) intermediate and one of an oxidation resulting in formation of HSO₄⁻(aq), which upon cooling converts to sulphuric acid.

The rate of SO₂ consumption, measured by exploiting the Raman band intensity measurements, is accelerated with increasing content of the I⁻ catalyst, in the range 0.156–0.780 I⁻/H₂O mol%, up to a certain limit, beyond which the lowered reactant SO₂ presence, the high I⁻ content and HSO₄⁻ product accumulation result in formation of the undesired iodine by-product by means of the reverse Bunsen reaction, thereby terminating the SO₂ disproportionation reaction. There is an apparent SO₂ pressure low threshold, which in combination with a high I⁻ content terminates the reaction, implying that a best compromise between the two is to be sought. Therefore, a continuous monitoring of the SO₂ pressure and intermittent replenishments with SO₂ would allow the controlled duration of batch operations at technical level. By that means the formation of iodine could be avoided and hence the required number of post-batch purification steps could be reduced. Most importantly, the technically utmost problematic separation step, i.e. the one of sulphur–iodine separation could be avoided.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

The data supporting this work are included in the main article and its SI. See DOI: https://doi.org/10.1039/d5cp02071a

Acknowledgements

This project has received funding from the European Union's European Innovation Council and SMEs Executive Agency (EISMEA) programme under grant agreement no 101115538 – SULPHURREAL.

References

1. S. J. Wagner and E. S. Rubin, Economic Implications of Thermal Energy Storage for Concentrated Solar Thermal Power, Renewable Energy, 2014, 61, 81–95.
2. P. Denholm, Y. H. Wan, M. Hummon and M. Mehos, The Value of CSP with Thermal Energy Storage in the Western United States, Energy Procedia, 2014, 49, 1622–1631.
3. B. Wong, L. Brown, R. Buckingham, W. Sweet, B. Russ and M. Gorensek, Sulfur Dioxide Disproportionation for Sulfur Based Thermochemical Energy Storage, Sol. Energy, 2015, 118, 134–144.
4. O. Achkari and A. El Fadar, Latest Developments on TES And CSP Technologies – Energy and Environmental Issues, Applications and Research Trends, Appl. Therm. Eng., 2020, 167, 114806.
5. K. Yagi, R. Sioshansi and P. Denholm, Evaluating a Concentrating Solar Power Plant as An Extended-Duration Peaking Resource, Sol. Energy, 2019, 191, 686–696.
6. T. Doi, Rev. Phys. Chem. Jpn., 2016, 33, 1–23.
7. A. Noglik, M. Roeb, C. Sattler and R. Pitz-Paal, Experimental Study on Sulfur Trioxide Decomposition in a Volumetric Solar Receiver-Reactor, Int. J. Energy Res., 2009, 33, 799–812.
8. C. S. Kim, S.-D. Hong, Y.-W. Kim, J.-H. Kim, W. J. Lee and J. Chang, Thermal Design of a Laboratory-Scale SO₃ Decomposer for Nuclear Hydrogen Production, Int. J. Hydrogen Energy, 2008, 33, 3688–3699.
9. A. Terada, J. Iwatsuki, S. Ishikura, H. Noguchi, S. Kubo, H. Okuda, S. Kasahara, N. Tanaka, H. Ota, K. Onuki and R. Hino, Development of Hydrogen Production Technology by Thermochemical Water Splitting IS Process Pilot Test Plan, J. Nucl. Sci. Technol., 2007, 44, 477–482.
10. V. M. Petruševski, M. Bukleski and M. Stojanovsk, On the Catalyzed Disproportionation of SO₂ in Aqueous Solution of KI: A Marathon Classroom Demonstration, J. Lab. Chem. Educ., 2013, 1, 1–4.
11. S. Boghosian, Vibrational Modes and Structure of Vanadium(SS) Complexes in M₂SO₄–V₂O₅ (M = K Or Cs) Molten Salt Mixtures, J. Chem. Soc., Faraday Trans., 1998, 94, 3463–3469.
12. S. Boghosian, A. Chrissanthopoulos and R. Fehrmann, Structure of Vanadium Oxosulfato Complexes in V₂O₅–M₂S₂O₇–M₂SO₄ (M = K, Cs) Melts. A High Temperature Spectroscopic Study, J. Phys. Chem. B, 2002, 106, 49–56.
13. S. Boghosian and G. N. Papatheodorou, Evaluation of Stoichiometric Coefficients and Thermodynamic Functions of Vapor Complexes Using Raman Spectroscopy: The Systems ZrX₄–AlX₃ (X = Br, Cl), J. Phys. Chem., 1989, 93, 415–421.
14. A. G. Kalampounias and S. Boghosian, Distribution Of Tellurite Polymorphs in the xM₂O–(1 − x)TeO₂ (M = Li, Na, K, Cs, and Rb) Binary Glasses Using Raman Spectroscopy, Vib. Spectrosc., 2012, 59, 18–22.
15. S. Boghosian, G. A. Voyiatzis and G. N. Papatheodorou, Gas-Phase, Liquid and Solid Complexes in the POC1₃–FeC1₃ System, J. Chem. Soc., Dalton Trans., 1996, 3405–3410.
16. D. R. Stull, Vapor Pressure of Pure Substances. Organic and Inorganic Compounds, Ind. Eng. Chem., 1947, 39, 517–540.
17. N. N. Greenwood and A. Earnshaw, Chemistry of the Elements, Butterworth-Heineman, Oxford, UK, 2nd edn, 1998, pp. 700–701.
18. F. Dankert, A. Feyh and C. von Hänisch, Chalcogen Bonding of SO₂ and s-Block Metal Iodides Near Room Temperature: A Remarkable Structural Diversity, Eur. J. Inorg. Chem., 2020, 2744–2756.
19. C. E. Lund Myhre, D. H. Christensen, F. M. Nicolaisen and C. J. Nielsen, Spectroscopic Study of Aqueous H₂SO₄ at Different Temperatures and Compositions: Variations in Dissociation and Optical Properties, J. Phys. Chem. A, 2003, 107, 1979–1991.
20. C. B. Knudsen, A. G. Kalampounias, R. Fehrmann and S. Boghosian, Thermal Dissociation of Molten KHSO₄: Temperature Dependence of Raman Spectra and Thermodynamics, J. Phys. Chem. B, 2008, 112, 11996–12000.
21. A. T. Ward, Raman Spectroscopy of Sulfur, Sulfur-Selenium, and Sulfur-Arsenic Mixtures, J. Phys. Chem., 1968, 72, 4133–4139.
22. K. R. Loos and A. C. Jones, Structure of the Triiodide Ion in Solution. Raman Evidence for the Existence of Higher Poly Iodide Species, J. Phys. Chem., 1974, 78, 2306–2307.
23. W. Kiefer and H. J. Bernstein, Vibrational-Rotational Structure in the Resonance Raman Effect of Iodine Vapor, J. Mol. Spectrosc., 1972, 43, 366–381.

---