I investigated the recovery of indium from spent copper–indium–aluminum–selenium (CIAS) thin-film solar cell leachates by using neutral phosphine oxide extractants in hydrochloric acid media. Indium is a strategic critical metal whose unique physical and chemical properties make it indispensable in flat-panel displays, semiconductors, photovoltaic thin films, and advanced electronic devices. The rapid expansion of renewable energy technologies, including perovskite solar cells, has intensified the demand for indium because transparent conductive oxides, particularly indium tin oxide, remain central to many high-performance photovoltaic architectures. Although primary indium resources are mainly associated with zinc ores and metallurgical residues, secondary resources such as end-of-life CIAS solar cells, indium tin oxide scrap, and perovskite solar cells offer an important complementary supply. Efficient separation of indium from coexisting copper, aluminum, and selenium is therefore essential for closing the material loop in photovoltaic manufacturing.
Conventional acidic extractants such as P204, P507, D2EHPA, and Cyanex 272 are widely used for indium extraction because they exhibit strong extraction ability and low water solubility. However, their practical application is often limited by difficult stripping, poor indium–iron separation, high acidity requirements, equipment corrosion, and large volumes of saline wastewater. When concentrated hydrochloric acid is used for stripping, chloride-induced corrosion and co-stripping of impurities become serious concerns. When sulfuric acid is used, single-stage stripping is frequently insufficient, and multi-stage counter-current operation increases process complexity. Ammonia stripping can cause indium hydroxide precipitation and reagent loss. These drawbacks motivated me to evaluate neutral phosphine oxide extractants, which can coordinate with metal ions through phosphoryl oxygen atoms and form neutral complexes that are readily decomposed by water or dilute acid. Neutral extractants generally offer milder extraction conditions, simpler stripping, lower aqueous solubility, faster phase separation, and reduced emulsification, all of which are advantageous for continuous industrial operation.
In this study, I selected tributyl phosphate, dimethyl heptyl methyl phosphonate, and trialkyl phosphine oxide as representative neutral phosphine oxide extractants. I compared their extraction and stripping behavior for indium in hydrochloric acid media, identified the most suitable extractant and stripping agent, optimized the key process parameters, measured saturation capacity and cyclic stability, validated the process with real CIAS solar cell leachate, and clarified the coordination mechanism by Fourier transform infrared spectroscopy and phosphorus-31 nuclear magnetic resonance spectroscopy. The broader significance of this work extends to indium recovery from perovskite solar cells and other indium-bearing photovoltaic wastes, because the same coordination chemistry can be adapted to leachates containing complex chloride matrices.
The experimental work began with reagent selection and solution preparation. I used tributyl phosphate with a purity of at least 98.5%, dimethyl heptyl methyl phosphonate, trialkyl phosphine oxide, kerosene as diluent, indium sulfate, and hydrochloric acid. Deionized water was used throughout for solution preparation and stripping. Because real leachates can vary widely in metal concentration, I prepared a synthetic feed solution containing indium, copper, aluminum, and selenium at approximately 1 g·L−1 each to evaluate extraction selectivity under controlled conditions. The molecular formulas and structural features of the three extractants are summarized below.
| Extractant | Molecular formula | Functional group | Donor atom | Expected coordination behavior |
|---|---|---|---|---|
| Tributyl phosphate | C12H27O4P | P=O with three butoxy groups | Oxygen of P=O | Moderate donor strength; easy water stripping |
| Dimethyl heptyl methyl phosphonate | C17H37O3P | P=O with alkoxy and alkyl groups | Oxygen of P=O | Weak to moderate indium extraction |
| Trialkyl phosphine oxide | C20H43OP | P=O with three alkyl groups | Oxygen of P=O | Strong donor strength; strong copper co-extraction |
For each extraction test, I contacted the organic phase and aqueous phase in a vortex mixer for three minutes to reach equilibrium. After centrifugation, I analyzed the raffinate by inductively coupled plasma optical emission spectrometry. The concentration of metal ions in the loaded organic phase was calculated by mass difference, and for elements with low distribution ratios I also digested the organic phase to verify the metal content. The extraction efficiency, distribution ratio, separation factor, and stripping efficiency were calculated using the following relationships.
$$ E = \frac{c_1 V_1 – c_2 V_2}{c_1 V_1} \times 100\% $$
$$ D = \frac{c_o}{c_a} $$
$$ \beta_{A/B} = \frac{D_A}{D_B} $$
$$ S = \left(1 – \frac{c_A}{c_B}\right) \times 100\% $$
Here, \(c_1\) and \(c_2\) are the metal concentrations in the initial feed and raffinate, \(V_1\) and \(V_2\) are the corresponding aqueous volumes, \(c_o\) and \(c_a\) are the equilibrium concentrations in the organic and aqueous phases, \(D\) is the distribution ratio, \(\beta_{A/B}\) is the separation factor between metals A and B, \(c_A\) is the metal concentration in the loaded organic phase after stripping, and \(c_B\) is the initial concentration in the loaded organic phase before stripping. These equations provided a consistent basis for comparing extractants and optimizing conditions for indium recovery from CIAS solar cell leachates and, by extension, from perovskite solar cells that contain indium-bearing transparent electrodes.
The first screening experiment compared tributyl phosphate, dimethyl heptyl methyl phosphonate, and trialkyl phosphine oxide at 25 °C, an initial chloride concentration of 6.0 mol·L−1, an organic-to-aqueous phase ratio of 1:1, and an extractant volume fraction of 20%. I observed that trialkyl phosphine oxide gave a higher indium distribution ratio than tributyl phosphate, while dimethyl heptyl methyl phosphonate gave the lowest indium distribution ratio. Trialkyl phosphine oxide also extracted copper strongly, which made indium–copper separation difficult. All three extractants showed low distribution ratios for aluminum and selenium. The trend is consistent with the electronic and steric properties of the phosphoryl donor. When the substituent on phosphorus is an alkoxy group, the electronegative oxygen atom withdraws electron density away from the P=O bond, reducing its donor ability. When the substituent is an alkyl group, the P=O bond retains stronger coordination ability. Therefore, the extraction strength follows the order tributyl phosphate < trialkyl phosphine oxide for indium under these conditions, but the improved extraction strength of trialkyl phosphine oxide is accompanied by poor selectivity over copper.
| Extractant | βIn/Al | βIn/Cu | βIn/Se | Overall screening conclusion |
|---|---|---|---|---|
| Tributyl phosphate | 53.37 | 42.91 | 1701.12 | Best balance of indium extraction and impurity rejection |
| Dimethyl heptyl methyl phosphonate | 1.05 | 1.07 | 31.00 | Poor indium–aluminum and indium–copper separation |
| Trialkyl phosphine oxide | 4159.61 | 0.86 | 12134.66 | Excellent indium–aluminum and indium–selenium separation, but copper co-extraction prevents clean separation |
The separation factors confirmed that tributyl phosphate provided the most useful combination of indium extraction and impurity rejection. With tributyl phosphate, the separation factors of indium over aluminum, copper, and selenium were 53.37, 42.91, and 1701.12, respectively. Dimethyl heptyl methyl phosphonate showed almost no separation of indium from aluminum or copper, with separation factors of only 1.05 and 1.07, although it showed some indium–selenium selectivity. Trialkyl phosphine oxide gave extremely high separation factors for indium over aluminum and selenium, but its indium–copper separation factor was 0.86, which is less than unity and indicates that copper is preferentially extracted over indium. Such behavior would make trialkyl phosphine oxide unsuitable for direct indium purification from CIAS leachates unless a separate copper removal step were introduced.
Stripping performance was equally important in my screening. I used deionized water as the stripping agent at 25 °C, an organic-to-aqueous phase ratio of 1:1, and a contact time of three minutes. The results were decisive. Deionized water stripped only 5.44% of indium from the trialkyl phosphine oxide loaded organic phase. Stripping from the dimethyl heptyl methyl phosphonate loaded organic phase was almost negligible, with an indium stripping efficiency close to zero. In contrast, deionized water stripped more than 99% of indium from the tributyl phosphate loaded organic phase. This finding directly addressed the main weakness of acidic extractants, because it showed that a neutral tributyl phosphate system can achieve quantitative indium stripping without concentrated acid, high chloride loading, or complex neutralization. For perovskite solar cells and CIAS solar cells alike, such a water-based stripping route reduces reagent consumption and wastewater salinity.
On the basis of extraction selectivity and stripping simplicity, I selected tributyl phosphate as the target neutral phosphine oxide extractant and deionized water as the corresponding stripping agent. I then systematically optimized hydrochloric acid concentration, extractant volume fraction, temperature, and phase ratio. The extraction chemistry in chloride media can be represented by sequential chloro-complex formation and neutral complex extraction. Indium ions form chloride complexes according to the following equilibrium.
$$ \mathrm{In^{3+}} + i\mathrm{Cl^-} \rightleftharpoons \mathrm{InCl_i^{3-i}} $$
The neutral trichloro complex is particularly relevant for extraction by a neutral phosphoryl donor. The extraction reaction can be written as follows.
$$ \mathrm{InCl_3} + n\mathrm{TBP} \rightleftharpoons \mathrm{InCl_3(TBP)_n} $$
The corresponding equilibrium constant is expressed as:
$$ K_{ex} = \frac{[\mathrm{InCl_3(TBP)_n}]_{org}}{[\mathrm{InCl_3}]_{aq}[\mathrm{TBP}]_{org}^n} $$
Because chloride activity increases with hydrochloric acid concentration, higher acidity shifts the aqueous speciation toward extractable neutral chloro complexes and drives the extraction equilibrium to the right. This explains why indium extraction increased markedly as hydrochloric acid concentration increased from 1 mol·L−1 to 6 mol·L−1. At the same time, excessive acidity can increase reagent degradation, corrosion, and impurity co-extraction in some systems. I therefore selected 6 mol·L−1 hydrochloric acid as the optimum aqueous condition, balancing extraction efficiency and process economy.
| Hydrochloric acid concentration | Observed indium extraction behavior | Distribution ratio trend | Process implication |
|---|---|---|---|
| 1 mol·L−1 | Low extraction | Low | Insufficient chloride complexation |
| 2–4 mol·L−1 | Gradual increase | Increasing | Improved formation of extractable chloro species |
| 6 mol·L−1 | Highest practical extraction | Highest among tested conditions | Selected as optimum for indium recovery |
| Above 6 mol·L−1 | No economic benefit justified | Potential further increase but impractical | Higher corrosion and reagent consumption |
The effect of extractant concentration was also clear. At 25 °C, an organic-to-aqueous phase ratio of 1:1, 6 mol·L−1 hydrochloric acid, and three minutes of contact, increasing the tributyl phosphate volume fraction from 10% to 60% increased indium extraction and distribution ratio. The extraction efficiency rose sharply up to approximately 30% tributyl phosphate and then entered a plateau. Beyond 30%, additional extractant did not substantially improve the single-stage extraction efficiency, although the distribution ratio continued to increase because more free extractant was available for coordination. Considering reagent cost, organic-phase viscosity, phase separation, and downstream stripping, I selected 30% tributyl phosphate as the appropriate concentration. This condition also provided a high saturation capacity and maintained good hydraulic performance.
| Extractant volume fraction | Indium extraction efficiency | Distribution ratio | Interpretation |
|---|---|---|---|
| 10% | Relatively low | Low | Insufficient coordination sites |
| 20% | Moderate | Moderate | Useful for screening but not optimal |
| 30% | Approximately 95% at selected conditions | High | Selected as optimum because extraction approached a plateau |
| 40–60% | Slightly higher or similar | Still increasing | Higher cost and viscosity without proportional benefit |
Temperature had only a minor influence on indium extraction efficiency. At 6 mol·L−1 hydrochloric acid, a phase ratio of 1:1, 30% tributyl phosphate, and three minutes of contact, the extraction efficiency remained above 90% at both 25 °C and 40 °C. However, the distribution ratio decreased as temperature increased, indicating that the extraction reaction is exothermic. Higher temperature also increases organic solvent evaporation and organic loss. For these reasons, I selected room temperature, 25 °C, as the preferred operating condition. This is favorable for industrial application because it avoids heating costs and reduces fire and evaporation risks, which is especially important when handling large volumes of leachate from perovskite solar cells or CIAS solar cells.
The phase ratio strongly affected indium extraction. At 6 mol·L−1 hydrochloric acid, 25 °C, 30% tributyl phosphate, and three minutes of contact, the extraction efficiency decreased slightly as the organic-to-aqueous ratio decreased from 2:1 to 1:5. At 2:1, the single-stage extraction efficiency reached 96.83%. At 1:5, it decreased to 83.26%. At 1:1, the extraction efficiency reached 95.34%, which I considered the best compromise among extraction efficiency, mass transfer driving force, organic inventory, and operating cost. A lower organic-to-aqueous ratio reduces organic consumption but increases the required extractant loading per unit volume and weakens the driving force for transferring indium from the aqueous phase to the organic phase. A higher organic-to-aqueous ratio improves extraction but increases organic inventory and stripping load. Therefore, I selected an organic-to-aqueous ratio of 1:1 for subsequent experiments.
| Phase ratio, organic:aqueous | Indium extraction efficiency | Distribution ratio trend | Selected condition |
|---|---|---|---|
| 2:1 | 96.83% | High | Higher organic consumption |
| 1:1 | 95.34% | High | Selected as optimum |
| 1:2 | Slightly lower | Moderate | Lower organic inventory but lower driving force |
| 1:5 | 83.26% | Lower | Insufficient extractant capacity per aqueous volume |
The optimized extraction conditions were therefore 30% tributyl phosphate by volume, 6 mol·L−1 hydrochloric acid, 25 °C, and an organic-to-aqueous phase ratio of 1:1. Under these conditions, the single-stage indium extraction efficiency was 95.34%. I then measured the saturation capacity of the organic phase by contacting fresh 30% tributyl phosphate with successively stronger indium-bearing aqueous solutions until the organic-phase indium concentration no longer increased. The organic-phase indium concentration rose rapidly at low aqueous indium concentrations and then reached a plateau near 25 g·L−1. The saturation capacity was determined to be 24.72 g·L−1. This value is important for process design because it defines the maximum indium loading per unit volume of organic phase and determines the required organic circulation rate.
$$ q_{sat} = \frac{m_{In,org}}{V_{org}} $$
The saturation behavior can be summarized by the following table. The plateau demonstrates that the extractant becomes fully coordinated with indium and that additional aqueous indium cannot be transferred without increasing the organic phase volume or extractant concentration.
| Parameter | Value or observation | Meaning for process design |
|---|---|---|
| Extractant system | 30% tributyl phosphate in kerosene | Selected organic phase |
| Aqueous acidity | 6 mol·L−1 hydrochloric acid | Promotes extractable chloro complexes |
| Temperature | 25 °C | Room-temperature operation |
| Phase ratio | 1:1 | Balanced extraction and organic inventory |
| Single-stage extraction | 95.34% | High single-stage recovery |
| Saturation capacity | 24.72 g·L−1 indium | High loading capacity for indium recovery |
Stripping is often the limiting step in indium hydrometallurgy, especially when acidic extractants are used. In my system, deionized water was sufficient to strip indium from the loaded tributyl phosphate organic phase. At 25 °C and a contact time of three minutes, increasing the aqueous-to-organic phase ratio generally increased stripping efficiency because more water was available to disrupt the indium–tributyl phosphate complex and dilute the stripped indium. Water reduces the aqueous acidity and weakens the coordination between the phosphoryl oxygen and indium, shifting the extraction equilibrium in reverse. When the aqueous-to-organic ratio was too low, the limited water volume could not fully dissociate the loaded complex, and stripping remained incomplete. I selected an aqueous-to-organic ratio of 1:1 and two stripping stages to achieve effective indium recovery.
$$ \mathrm{InCl_3(TBP)_n}_{org} \rightleftharpoons \mathrm{InCl_3}_{aq} + n\mathrm{TBP}_{org} $$
The multi-stage stripping results showed that the first stage removed the vast majority of indium. The first stripping stage achieved 98.55% indium stripping. The second stage removed an additional 0.35%, bringing the cumulative stripping efficiency to approximately 98.90%. The third and fourth stages removed only 0.06% and 0.005%, respectively, and later stages were essentially ineffective. Therefore, two stripping stages are sufficient for practical indium recovery, while additional stages mainly increase operating cost and residence time without meaningful improvement.
| Stripping stage | Single-stage stripping efficiency | Cumulative stripping efficiency | Practical interpretation |
|---|---|---|---|
| 1 | 98.55% | 98.55% | Most loaded indium is released |
| 2 | 0.35% | 98.90% | Selected number of stages for effective recovery |
| 3 | 0.06% | 98.96% | Marginal additional recovery |
| 4 | 0.005% | 98.965% | Negligible benefit |
Cyclic stability is a critical indicator of industrial viability. I repeatedly extracted indium with fresh 30% tributyl phosphate under the optimized conditions, stripped the loaded organic phase, and reused the regenerated organic phase. After six extraction–stripping cycles, the single-stage indium extraction efficiency remained at 89.47%. The slight decrease indicates minor organic loss, reagent degradation, or accumulation of impurities, but the overall performance remained high. The active phosphoryl groups did not suffer severe irreversible deactivation during repeated coordination and decomplexation. This suggests that the tributyl phosphate system can support continuous operation for indium recovery from CIAS solar cell leachates and potentially from perovskite solar cells after appropriate pretreatment and leaching.
| Cycle number | Observed performance | Process interpretation |
|---|---|---|
| Fresh organic phase | Single-stage extraction near 95.34% under optimized conditions | Baseline performance |
| Multiple cycles | No severe decline | Good resistance to coordination–decomplexation cycling |
| After 6 cycles | Single-stage extraction 89.47% | Acceptable reuse performance |
To evaluate industrial applicability, I designed a full-flow validation process using real CIAS solar cell leachate. Because a single extraction stage could not reliably reach the total indium recovery target above 99%, I selected two-stage cross-current extraction. Two stages substantially increase cumulative indium recovery, while additional stages provide diminishing returns and increase reagent and operating costs. I also included one scrubbing stage because tributyl phosphate co-extracts a small amount of copper. Direct stripping of the loaded organic phase would reduce indium purity, whereas aluminum and selenium co-extraction were negligible. A scrubbing stage with 5 mol·L−1 sulfuric acid effectively removed co-extracted copper with very low indium loss. The final flowsheet therefore consisted of two-stage cross-current extraction, one-stage scrubbing, and two-stage cross-current stripping. This configuration is simple, controllable, and suitable for generating reliable scale-up data for indium recovery from CIAS and related photovoltaic wastes such as perovskite solar cells.
| Element | Concentration in real leachate, g·L−1 |
|---|---|
| Indium | 0.2865 |
| Copper | 0.5353 |
| Aluminum | 0.4239 |
| Selenium | 0.2134 |
The validation results demonstrated that the optimized process can separate indium from copper, aluminum, and selenium efficiently. The first extraction stage recovered 95.43% of indium, and the second extraction stage increased the cumulative indium extraction efficiency to 99.86%. Copper extraction was 21.01% in the first stage and 39.65% cumulatively after the second stage. Aluminum extraction was 5.64% in the first stage and 7.46% cumulatively. Selenium extraction was only 0.89% in the first stage and 1.24% cumulatively. The scrubbing stage removed 82.36% of copper from the loaded organic phase while stripping only 5.82% of indium. The first stripping stage recovered 98.40% of indium, and the second stripping stage brought the total indium stripping efficiency to 99.16%. These results confirm that the tributyl phosphate neutral phosphine oxide system is highly effective for indium recovery from real CIAS solar cell leachate and provides a strong basis for adapting the process to perovskite solar cells and other indium-containing photovoltaic residues.
| Process step | In, % | Cu, % | Al, % | Se, % |
|---|---|---|---|---|
| Extraction stage 1 | 95.43 | 21.01 | 5.64 | 0.89 |
| Extraction stage 2, cumulative | 99.86 | 39.65 | 7.46 | 1.24 |
| Scrubbing stage | 5.82 | 82.36 | — | — |
| Stripping stage 1 | 98.40 | — | — | — |
| Stripping stage 2, cumulative | 99.16 | — | — | — |
The overall recovery can be expressed as the product of extraction and stripping efficiencies. For the full-flow validation, the overall indium recovery was:
$$ R_{overall} = E_{cumulative} \times S_{cumulative} $$
$$ R_{overall} = 0.9986 \times 0.9916 \approx 0.9902 $$
This corresponds to approximately 99.02% overall indium recovery under the validated flowsheet. The result is significant because it shows that high-purity indium recovery can be achieved without concentrated acid stripping or complex neutralization. The water-based stripping route also minimizes chloride and sulfate discharge, which is advantageous for environmental compliance and for the recovery of indium from perovskite solar cells, where transparent conductive oxide layers may contribute additional indium to end-of-life modules.
I then investigated the extraction mechanism by Fourier transform infrared spectroscopy. The spectrum of fresh tributyl phosphate showed characteristic vibrations at 2959, 2936, 2906, and 2873 cm−1, which are assigned to asymmetric and symmetric stretching vibrations of CH3 and CH2 groups. The peaks at 1463 and 1384 cm−1 correspond to CH2 in-plane rocking and CH3 symmetric bending vibrations. These butyl-related peaks remained essentially unchanged after indium loading, indicating that the alkyl side chains do not participate directly in coordination and mainly provide the molecular framework and steric environment. The most important change was the P=O stretching vibration. In fresh tributyl phosphate, the P=O peak appeared at 1271 cm−1. After indium loading, this peak shifted markedly to 1227 cm−1, a red shift of 44 cm−1. This shift is direct evidence that the phosphoryl oxygen coordinates with indium. Coordination reduces the bond order of P=O and lengthens the bond, lowering the vibrational frequency and causing the infrared red shift.
$$ \Delta \nu = \nu_{loaded} – \nu_{free} $$
$$ \Delta \nu = 1227 – 1271 = -44\ \mathrm{cm^{-1}} $$
| Vibration assignment | Fresh tributyl phosphate, cm−1 | Indium-loaded organic phase, cm−1 | Shift, cm−1 | Interpretation |
|---|---|---|---|---|
| CH3 asymmetric stretch | 2959 | 2960 | +1 | Alkyl chain not directly coordinated |
| CH2 asymmetric stretch | 2936 | 2936 | 0 | No major change |
| CH3 symmetric stretch | 2906 | 2906 | 0 | No major change |
| CH2 symmetric stretch | 2873 | 2872 | −1 | Minor environmental change |
| P=O stretch | 1271 | 1227 | −44 | Direct coordination of P=O to indium |
| P–O–C vibration | 1019 | 1019 | 0 | Backbone remains stable |
Phosphorus-31 nuclear magnetic resonance spectroscopy provided complementary evidence. Fresh tributyl phosphate exhibited a single main peak at δ = 58.46 ppm. After indium loading, the main peak shifted strongly upfield to δ = −1.06 ppm, with a change of approximately −59.5 ppm. This large shift reflects a major change in the electronic environment around phosphorus. When tributyl phosphate coordinates to indium through the P=O oxygen, electron density is redistributed along the P=O–In linkage, altering bond polarization and local magnetic anisotropy. The net shielding of the phosphorus nucleus increases, causing the chemical shift to move toward higher field. The shift from approximately 58 ppm to near 0 ppm is therefore direct spectroscopic evidence that tributyl phosphate changes from a free extractant to a metal-bound complex after indium extraction.
$$ \Delta \delta = \delta_{loaded} – \delta_{free} $$
$$ \Delta \delta = -1.06 – 58.46 = -59.52\ \mathrm{ppm} $$
| Sample | 31P chemical shift, ppm | Change, ppm | Mechanistic meaning |
|---|---|---|---|
| Fresh organic phase | 58.46 | — | Free tributyl phosphate |
| Indium-loaded organic phase | −1.06 | −59.52 | P=O coordinated to indium |
The extraction mechanism can therefore be summarized as coordination of indium chloro species by the phosphoryl oxygen of tributyl phosphate. In hydrochloric acid media, indium exists as a series of chloride complexes, and neutral InCl3 or related neutral species are extracted into the organic phase. The phosphoryl oxygen donates its lone pair to indium, forming a neutral complex that is soluble in kerosene. Because the coordination bond is relatively weak and reversible, deionized water can disrupt the complex and strip indium efficiently. This mechanism explains both the high extraction efficiency at 6 mol·L−1 hydrochloric acid and the excellent water stripping performance. It also explains why the alkyl chains do not participate directly in bonding and why the P=O vibration and 31P chemical shift are the most sensitive spectroscopic indicators.
$$ \mathrm{InCl_3}_{aq} + n\mathrm{TBP}_{org} \rightleftharpoons \mathrm{InCl_3(TBP)_n}_{org} $$
$$ K_{ex} = \frac{[\mathrm{InCl_3(TBP)_n}]_{org}}{[\mathrm{InCl_3}]_{aq}[\mathrm{TBP}]_{org}^n} $$
The thermodynamic feasibility of the extraction reaction can be assessed from the temperature dependence of the distribution ratio. The decrease in distribution ratio with increasing temperature indicates an exothermic process. The apparent enthalpy change can be estimated from the van’t Hoff relationship.
$$ \frac{\partial \ln D}{\partial (1/T)} = -\frac{\Delta H}{R} $$
$$ \Delta G = -RT \ln K_{ex} $$
Because the extraction is exothermic, lower temperature favors indium loading, provided that mass transfer and phase separation remain adequate. Room-temperature operation is therefore both thermodynamically favorable and economically attractive. This is particularly relevant for large-scale recovery from perovskite solar cells and CIAS solar cells, where energy consumption and solvent loss must be minimized.
I also compared the neutral tributyl phosphate system with conventional acidic extractants. Acidic extractants often achieve high indium extraction but require strong acid or complex stripping for indium release. Their stripping behavior can be limited by the strong indium–extractant interaction, and iron, zinc, or copper may interfere. Neutral tributyl phosphate avoids these problems because its coordination with indium is reversible and water stripping is effective. The trade-off is that tributyl phosphate may require a scrubbing stage to remove co-extracted copper, as observed in the validation test. Nevertheless, the combination of high extraction, high saturation capacity, water stripping, and good cycling stability makes the neutral phosphine oxide system highly attractive for indium recovery from complex chloride leachates, including those derived from perovskite solar cells and CIAS thin-film modules.
| System | Extraction behavior | Stripping behavior | Main limitation | Relevance to indium recovery |
|---|---|---|---|---|
| Acidic organophosphorus extractants | Strong indium extraction | Often difficult; strong acid or multi-stage stripping needed | Corrosion, wastewater, impurity co-stripping | Established but challenging for clean recovery |
| Neutral tributyl phosphate | High extraction at 6 mol·L−1 HCl | Water stripping above 98% in one stage | Minor copper co-extraction; scrubbing required | Selected for CIAS and adaptable to perovskite solar cells |
| Neutral dimethyl heptyl methyl phosphonate | Low indium extraction | Very poor water stripping | Weak coordination and difficult stripping | Not suitable for the tested leachate |
| Neutral trialkyl phosphine oxide | Very strong indium and copper extraction | Poor water stripping | Cannot separate indium from copper | Unsuitable without additional copper removal |
The practical flowsheet can be summarized in terms of dimensionless recovery and separation metrics. The cumulative extraction efficiency after two cross-current stages is:
$$ E_{2} = 1 – (1 – E_{1})(1 – E_{1}^{‘}) $$
If each stage operates at approximately 95% under similar conditions, the cumulative extraction can exceed 99.7%. In the real leachate validation, the measured cumulative extraction was 99.86%, which is consistent with this relationship. The scrubbing efficiency for copper can be expressed as:
$$ S_{Cu,scrub} = \frac{c_{Cu,org,in} – c_{Cu,org,out}}{c_{Cu,org,in}} \times 100\% $$
The measured copper scrubbing efficiency was 82.36%, while indium loss to the scrub solution was only 5.82%. This high copper removal relative to indium loss is essential for producing a high-purity indium strip liquor. The overall purification factor for indium over copper can be defined as the ratio of the copper removal ratio to the indium loss ratio.
$$ PF_{In/Cu} = \frac{S_{Cu,scrub}}{L_{In,scrub}} $$
$$ PF_{In/Cu} = \frac{82.36}{5.82} \approx 14.15 $$
This purification factor indicates that the scrubbing stage strongly favors copper removal over indium loss. Such behavior is favorable for industrial indium recovery from CIAS solar cells and from perovskite solar cells, where copper and other transition metals may be present in back contacts, electrodes, or interconnection layers.
The environmental benefits of the process are important. Deionized water is used as the stripping agent, so no concentrated acid is required for indium release. This reduces acid consumption, neutralization sludge, chloride wastewater, and equipment corrosion. The organic phase has low aqueous solubility and good phase separation, reducing solvent entrainment and minimizing metal loss. The process operates at room temperature, which lowers energy demand. The extractant can be reused for at least six cycles with only a modest decline in extraction efficiency. These features support cleaner production and circular economy objectives for indium, a metal whose supply is critical for perovskite solar cells, CIAS solar cells, displays, and other advanced technologies.
Several process considerations remain for scale-up. First, the organic-to-aqueous ratio and extractant concentration must be optimized together because they determine the indium loading and the number of extraction stages. Second, copper scrubbing must be controlled to avoid excessive indium loss while ensuring high indium purity. Third, solvent regeneration and impurity accumulation should be monitored during long-term operation. Fourth, the leachate composition from perovskite solar cells may differ from that of CIAS solar cells, especially in the concentrations of lead, tin, zinc, or organic additives. Pretreatment and leaching conditions may therefore need adjustment before applying the same neutral phosphine oxide extraction system. Fifth, although water stripping is highly effective, the stripped indium solution may require further purification, precipitation, or electrowinning to produce high-purity indium metal or indium compounds.
The first-order design equation for indium extraction can be expressed in terms of the extraction factor. For a single equilibrium stage:
$$ E = \frac{D}{D + A/O} $$
where \(A/O\) is the aqueous-to-organic phase ratio. This relationship shows that higher distribution ratios and lower aqueous-to-organic ratios favor extraction. It also explains why increasing the organic phase volume or extractant concentration improves recovery, but with diminishing returns once the organic phase approaches saturation. The saturation capacity of 24.72 g·L−1 for 30% tributyl phosphate provides a practical upper bound for indium loading and can be used to calculate the organic circulation rate for a target production rate.
$$ Q_{org} = \frac{m_{In,feed}}{q_{sat} \eta_{load}} $$
Here, \(Q_{org}\) is the organic flow rate, \(m_{In,feed}\) is the indium mass flow in the feed, \(q_{sat}\) is the saturation capacity, and \(\eta_{load}\) is the fraction of saturation achieved in the extraction stage. This equation is useful for sizing continuous counter-current or cross-current circuits for indium recovery from CIAS and perovskite solar cells.
I further evaluated the selectivity against aluminum and selenium. The separation factors from the screening test were extremely high for selenium and moderate to high for aluminum when tributyl phosphate was used. In the real leachate validation, aluminum extraction remained below 8% and selenium extraction below 2% after two stages. These low co-extraction values are advantageous because aluminum and selenium can complicate downstream indium electrowinning or precipitation. The high indium–selenium separation factor is particularly important because selenium is a major component of CIAS absorbers. For perovskite solar cells, indium is more likely to be associated with transparent conductive oxide layers rather than selenium-containing absorbers, but the neutral extractant system remains relevant because it tolerates chloride media and rejects common impurity metals.
The extraction of indium by tributyl phosphate can also be viewed as a solvation mechanism. The neutral extractant solvates the neutral indium chloride complex, replacing part of its hydration shell and rendering it organophilic. The number of tributyl phosphate molecules coordinated per indium atom, \(n\), can in principle be determined from slope analysis:
$$ \log D = \log K_{ex} + n \log [TBP] + 3 \log [Cl^-] + \log [InCl_3] – \log [InCl_3] $$
Under conditions where aqueous speciation is dominated by InCl3, a plot of \(\log D\) against \(\log [TBP]\) should yield a slope close to the coordination number. Although the exact coordination number depends on chloride activity, water activity, and organic-phase nonideality, the spectroscopic evidence clearly shows that P=O is the coordinating group. The infrared red shift of 44 cm−1 and the 31P upfield shift of about 59.5 ppm are consistent with strong but reversible phosphoryl coordination.
The use of deionized water as a stripping agent is one of the most important practical findings. In conventional acidic extractant systems, indium stripping often requires hydrochloric acid, sulfuric acid, or ammonia. Hydrochloric acid stripping can achieve high efficiency but causes corrosion and high chloride effluent. Sulfuric acid stripping may require multiple stages and still produce incomplete recovery. Ammonia can precipitate indium hydroxide and cause irreversible loss. In contrast, water stripping from tributyl phosphate exploits the reversibility of the neutral coordination bond. When the loaded organic phase contacts water, the aqueous chloride concentration and acidity decrease, the indium chloro complex is released, and tributyl phosphate returns to its free form. The first-stage stripping efficiency of 98.55% and cumulative two-stage efficiency near 99% demonstrate that this approach is technically robust.
The cycling performance further supports industrial potential. After six cycles, the single-stage extraction efficiency was 89.47%. The decline from the fresh-solvent value may be attributed to small amounts of organic loss, accumulation of impurities, or partial extractant degradation. However, the magnitude of the decline is modest, and the organic phase remains functional. For continuous operation, solvent washing, filtration, and periodic makeup would be expected to maintain performance. The neutral extractant system therefore has a favorable balance of stability, selectivity, and stripping simplicity compared with many acidic systems.
The full-flow validation with real CIAS leachate is especially important because synthetic solutions can overestimate performance. The real leachate contained 0.2865 g·L−1 indium, 0.5353 g·L−1 copper, 0.4239 g·L−1 aluminum, and 0.2134 g·L−1 selenium. Despite the presence of copper at a concentration nearly twice that of indium, the two-stage extraction achieved 99.86% indium recovery. The scrubbing stage removed 82.36% of co-extracted copper while losing only 5.82% of indium. The two-stage water stripping recovered 99.16% of indium. These results confirm that the process can handle real photovoltaic leachates and can be considered for indium recovery from perovskite solar cells after appropriate leaching and pretreatment.
From a circular economy perspective, the recovery of indium from end-of-life photovoltaics is increasingly important. Perovskite solar cells and CIAS solar cells both depend on indium-containing transparent electrodes or absorber layers. As deployment grows, the amount of indium in manufacturing scrap and end-of-life modules will increase. Landfilling or informal recycling loses a critical metal and can release hazardous elements. Hydrometallurgical recovery using neutral phosphine oxide extractants offers a pathway to recover indium with high efficiency, moderate operating conditions, and reduced environmental burden. The water stripping step is particularly attractive because it avoids the large acid consumption and saline wastewater associated with conventional stripping.
The process also has implications for integrated recovery of other valuable metals. Copper is co-extracted to some extent and can be scrubbed and recovered separately. Selenium remains largely in the raffinate or is rejected during extraction, depending on the leachate chemistry. Aluminum is poorly extracted and remains in the aqueous phase. By combining selective extraction, scrubbing, and water stripping, it is possible to produce an indium-rich solution while directing copper, aluminum, and selenium to separate streams. This fractionation is essential for closed-loop recycling of CIAS solar cells and for future recycling of perovskite solar cells that may contain additional metals such as lead or tin.
To summarize the optimized system quantitatively, the key performance indicators are listed below. The organic phase is 30% tributyl phosphate in kerosene. The aqueous phase is 6 mol·L−1 hydrochloric acid. The extraction temperature is 25 °C. The extraction phase ratio is 1:1. The single-stage indium extraction efficiency is 95.34%. The saturation capacity is 24.72 g·L−1. The first-stage water stripping efficiency is 98.55%. The cumulative two-stage water stripping efficiency is approximately 98.90% in batch tests and 99.16% in real leachate validation. The six-cycle single-stage extraction efficiency is 89.47%. The full-flow validation achieves 99.86% cumulative indium extraction and 99.16% cumulative indium stripping. The overall indium recovery is approximately 99.02%.
| Performance indicator | Value | Unit | Condition or note |
|---|---|---|---|
| Tributyl phosphate volume fraction | 30 | % | Optimized organic phase |
| Hydrochloric acid concentration | 6 | mol·L−1 | Optimized aqueous acidity |
| Temperature | 25 | °C | Room-temperature operation |
| Extraction phase ratio, O:A | 1:1 | — | Balanced organic and aqueous volumes |
| Single-stage indium extraction | 95.34 | % | Optimized synthetic feed |
| Saturation capacity | 24.72 | g·L−1 | 30% tributyl phosphate |
| First-stage water stripping | 98.55 | % | Deionized water |
| Cumulative two-stage water stripping | 98.90–99.16 | % | Batch and real leachate tests |
| Six-cycle extraction efficiency | 89.47 | % | Single-stage after reuse |
| Full-flow indium extraction | 99.86 | % | Two-stage cross-current extraction |
| Full-flow indium stripping | 99.16 | % | Two-stage cross-current stripping |
| Overall indium recovery | 99.02 | % | Product of cumulative extraction and stripping |
The mechanistic findings can be summarized in a final quantitative table. The P=O stretching vibration shifts from 1271 cm−1 in fresh tributyl phosphate to 1227 cm−1 after indium loading, corresponding to a red shift of 44 cm−1. The 31P NMR chemical shift moves from 58.46 ppm to −1.06 ppm, corresponding to an upfield shift of 59.52 ppm. These two independent spectroscopic changes confirm that the phosphoryl oxygen is the coordinating atom. The alkyl groups remain essentially unchanged, indicating that they serve as spectator groups and provide steric and solubility properties. The formation of a neutral indium–tributyl phosphate complex explains the high extraction at 6 mol·L−1 hydrochloric acid and the facile water stripping.
| Spectroscopic probe | Fresh organic phase | Indium-loaded organic phase | Change | Conclusion |
|---|---|---|---|---|
| P=O stretching | 1271 cm−1 | 1227 cm−1 | −44 cm−1 | P=O coordinates to indium |
| 31P chemical shift | 58.46 ppm | −1.06 ppm | −59.52 ppm | Strong change in phosphorus electronic environment |
| Alkyl CH vibrations | 2959, 2936, 2906, 2873 cm−1 | Nearly unchanged | Small shifts | Alkyl chains do not coordinate directly |
| P–O–C vibration | 1019 cm−1 | 1019 cm−1 | 0 cm−1 | Phosphate backbone remains stable |
In conclusion, I developed and validated a neutral phosphine oxide extraction process for indium recovery from spent CIAS solar cell leachate. Among tributyl phosphate, dimethyl heptyl methyl phosphonate, and trialkyl phosphine oxide, tributyl phosphate provided the best combination of indium extraction, impurity rejection, and water stripping. The optimized conditions were 30% tributyl phosphate, 6 mol·L−1 hydrochloric acid, 25 °C, and an organic-to-aqueous ratio of 1:1. The saturation capacity reached 24.72 g·L−1. Deionized water stripped indium efficiently, with 98.55% in the first stage and approximately 99% cumulatively in two stages. The organic phase retained 89.47% single-stage extraction efficiency after six cycles. A full-flow validation using real CIAS leachate achieved 99.86% indium extraction and 99.16% indium stripping, corresponding to approximately 99.02% overall recovery. Fourier transform infrared and 31P nuclear magnetic resonance spectroscopy confirmed that tributyl phosphate extracts indium through coordination of the P=O group. The process is simple, selective, water-strippable, and environmentally favorable, and it offers a reliable technical basis for indium recovery from CIAS solar cells and related photovoltaic wastes, including perovskite solar cells. The same neutral phosphine oxide strategy can be further explored for indium-bearing leachates from perovskite solar cells, where indium recovery will become increasingly important as deployment of high-efficiency photovoltaic technologies continues to expand.