Overcoming Stability Challenges Associated with Determination of Residual Phosphine Ligands in Active Pharmaceutical Ingredients and Their Intermediates Using Liquid Chromatography with Derivatization

Suzuki-Miyaura coupling reactions are a very well-established area of organic chemistry in the manufacture of active pharmaceutical ingredients (APIs), with the area of application lying in the synthesis of biphenyl structures.¹ Figure 1 depicts the general form of the Suzuki-Miyaura (Suzuki) reaction and the non-oxidized and oxidized forms of the phosphine ligands (X Phos) included in this study.

The mechanistic detail of the Suzuki reaction is illustrated in Figure 2.

The key steps of the reaction are as follows: oxidative addition, transmetallation, and reductive elimination. The reaction occurs in the presence of a base; the role of the base is in the formation of the palladium complex [R-Pd-OEt], activation of the boronate compound by formation of a more reactive species, and the reductive elimination, which yields the coupled compound.

The scope of the reaction partners has since expanded from aryls to include alkyls, alkenyls, and alkynyls. Potassium trifluoroborates and organoborates or boronate esters may be used in place of boronic acids. Some pseudohalides (for example, triflates) may also be used as coupling partners instead of the halide derivatives. Phosphine ligands are the most significant class of ligands for cross-coupling because their electronic and steric properties can be readily tuned.

Case Study

For this case study, the Suzuki reaction was conducted under oxygen-free conditions by thorough nitrogen sparging of reaction solvents and reaction vessels. The addition of the X Phos was performed under inert conditions to prevent oxidation. The oxidized form of the X Phos did not participate in the reaction. Following the isolation workup of the product material, oxygen extrusion measures would no longer be required. If any residual X Phos was present in the product material, then this would be susceptible to oxidation of the oxidized form of the ligand.

The Suzuki reagents are common additive/catalysts used in pharmaceutical manufacturing, whilst testing for the residual Pd is common, testing for residual levels of the phosphine ligands is less common. Residual ligands in the material product are process impurities, analogous to residual solvents. Residual phosphine ligands exhibit a degree of toxicity, and an accurate method for demonstrating control over their levels is essential. A liquid chromatography (LC) method is the method of choice for this purpose due to the non-volatility of the phosphine ligands coupled with their strong UV chromophores. Quantification of the non-oxidized and oxidized forms of the X Phos was challenging due to these reasons:

1. The UV chromophore of structures differ from each other significantly due to the additional conjugation present in the oxidized form of the ligand.
2. Solubilization of both compounds is challenging, but the relative solubility of the compounds differs significantly.
3. The conversion of the non-oxidized form to the oxidized form is not controllable, and the extent of conversion to the oxidized form has been observed to be highly variable.

The three above observations occur concurrently, which only serves to compound any uncertainty in the results for any attempt to quantify residual amounts of the non-oxidized and oxidized forms of the X Phos ligands vs. a prepared calibration standard of X Phos. The residual phosphine ligands can potentially be present in an unoxidized and oxidized form, which is not conducive to accurate and precise quantification via a chromatographic method. This method serves to convert any residual phosphine ligand present in the unoxidized form to the oxidized form. So, if any residual phosphine ligand is present, it will be present only in the oxidized form. As the residual phosphine ligand is present only in the oxidized form, this circumvents the difficulties based upon differing UV chromophores and solubility imparted by the residual ligands being present in two different forms.

Initial Strategy for Quantification by Liquid Chromatography

A variety of steps were taken to inhibit oxidation. These steps included sparging of the diluent with oxygen and addition of the antioxidants sodium thiosulfate and dithiothreitol. The conclusion from these initial studies was that oxidation was inadequately controlled, and the presence of the antioxidants impacted the chromatography to an extent where it was unusable. Following this, the strategy of method development was redirected from implementing preventative mechanisms of the X Phos oxidation to derivatization of the X Phos to the more stable oxidized form.

Alternative Strategy for Development

The basis of the presented alternative strategy was to derivatize all the residual X Phos to the oxidized form and quantify the total ligand as the oxidized form. The solubility of the X Phos non-oxidized, oxidized forms, and the Suzuki reaction product was obtained, and a solubility assessment was performed using methanol–water mixtures, dimethylformamide, and dichloromethane. From this study, dichloromethane was identified as a suitable solvent.

Experimental

Establishment of Liquid Chromatography and X Phos Oxidation Conditions
A phenyl stationary phase was chosen: 2.1 mm x 50 mm, 1.7-µm Acquity UPLC CSH Phenyl Hexyl column (Waters). C18 phases are strongly retentive of X Phos compounds, and the phenyl functionality may impart additional selectivity for the non-oxidized and oxidized forms. Water and methanol were assigned as weak and strong mobile phases with a UV detection at a wavelength of 220 nm on an H Class UPLC system with a UV detector (TUV model of detector) (Waters). Sample preparation was as per the specificity and recovery assessment section below.

Specificity and Recovery Assessment for Oxidized X Phos from Sample Material
X Phos Standard Preparation: X Phos Working Standard Preparation: 250 µL of the stock preparation was diluted to 50 mL with dichloromethane and mixed well (concentration 0.0025 mg/mL).
Blank Preparation:Stock Blank Preparation: 20 mL of dichloromethane was added to a 50 mL volumetric flask. 200 µL of 30% hydrogen peroxide was added and mixed well. It was allowed to stand for 20 min with occasional gentle swirling and then diluted to volume and mixed well.
Working Blank Preparation: 250 µL of the stock blank preparation was diluted to 50 mL with dichloromethane and mixed well.
X Phos Spiking Stock Solution Preparation: 25 mg of X Phos was weighed into a 50 mL volumetric flask, dissolved, and diluted to volume with dichloromethane and mixed well (concentration, 0.50mg/mL).
X Phos Spiking Solution Preparation: 5 mL of the X Phos Spiking Stock Solution Preparation was diluted 5 mL to 50 mL with dichloromethane and mixed well (concentration 0.05 mg/mL).
Sample Material Spiking Stock Solution (Solution A): 100 mg sample material was weighed into a 50 mL volumetric flask. It was dissolved in 20 mL dichloromethane, 5 mL of X Phos Spiking Solution Preparation added and 200 µL of 30% hydrogen peroxide. It stood for 20 min with gentle occasional swirling and then diluted to volume with dichloromethane and mixed well.
Spiked Sample Preparations (Prepared in Triplicate):5 mL of Solution A was added to a 20 mL volumetric flask. 10 mL of dichloromethane was added with 1 mL of the X Phos Spiking Solution Preparation and 200 µL of 30% hydrogen peroxide. It was allowed to stand for 20 min with occasional gentle occasional swirling and then diluted to volume with dichloromethane and mixed well.
Unspiked Sample (Prepared in Triplicate): 5 mL of Solution A was added to a 20 mL volumetric flask. 10 mL of dichloromethane was added, and then 200 µL of 30% hydrogen peroxide. It was allowed to stand for 20 min with occasional gentle swirling. It was then diluted to volume with dichloromethane and mixed well.

Results

Standardization and Accuracy: By ensuring that both the calibration standards and all sample solutions contain the phosphine ligand exclusively in the oxidized form, the method eliminates variability caused by the coexistence of non-oxidized and oxidized species. This standardization is critical for accurate and precise quantification, as it provides a consistent reference point for all measurements.
Complete Conversion and Irreversibility: The procedure guarantees full conversion of any residual phosphine ligand to the oxidized form. Since oxidation is irreversible, there is no risk of the unoxidized form reappearing and negatively impacting the method accuracy. This stability is essential for reliable quantification, especially in pharmaceutical analysis, where trace residual phosphine ligands must be tightly controlled.
Analytical Benefits: The approach overcomes the challenges posed by differing UV chromophores and solubility between the two forms of phosphine ligands. By converting everything to the oxidized form, the method simplifies the chromatographic analysis, improves sensitivity, accuracy, and precision.
Advantages: This derivatization strategy is novel because it circumvents the need to separately quantify both forms and prevents oxidation during sample preparation. It provides a robust solution for analysts working with Suzuki-Miyaura coupling products, where residual phosphine ligands are process impurities with potential toxicity.

A system precision check was performed using the calibration standard. The X Phos content of the triplicate unspiked and spiked sample preparations was determined vs. the mean response factor from the injections of the calibration standard. Representative chromatography is depicted in Figure 3. A summary of the quantitative method performance is listed here.

Summary of Method Performance

Calibration Standard Precision: Oxidized X Phos standard retention time precision (n = 6) – 0.14%; Oxidized X Phos standard response precision (n = 6) – 1.25%.
Oxidized X Phos Accuracy: Replicate 1 – 101.3%, Replicate 2 – 105.9%, Replicate 3 – 103.2%; Mean – 103.2%, % RSD 2.3%.

These values are based upon recovery of the X Phos from the test API material, with a correction for the sample intrinsic X Phos content being applied.
Method Sensitivity: Mean USP signal-to-noise ratio for the six oxidized X Phos standards was 87. Approximate limit of quantitation (LOQ) for X Phos was 0.06% w/w.

Conclusion

Quantification of the unoxidized and oxidized forms of phosphine ligands can present the analyst with challenges because of the differing solubilities of the forms, their different UV chromophores, and the difficulties in preventing and controlling oxidation. The method described ensures that all residual phosphine ligands are quantified in their oxidized form, providing a stable, accurate, and precise analytical procedure. This innovation addresses the core challenges of solubility, UV chromophore variability, and oxidation control, making it a valuable advancement for pharmaceutical quality control.

Reference

1. Miyaaura, N. Nickel or Palladium-catalyzed coupling between an organoboronic acid and alkenyl halide. Tetrahedron Letters 1979, 20, 3437–3440.