Synopsis : Crude API sample containing >100 ppm of palladium, introduced by a Pd(0)-catalysed coupling reaction performed at the final stage of the synthetic process. Conventional metal scavenging protocols were not readily adaptable to this compound due to its very poor solubility.

Key Issue : Speed to initial 100 g quantities of API and then kg quantities for testing.

Process Target : Isolation of the API without any significant loss and with Pd content reduced to within regulatory guidelines (<5 ppm).

Process Step : Treatment of the existing process stream, consisting of a solution of the crude API sample in an organic-aqueous solvent mixture at >80 oC for 1 hour in slurry mode with PhosphonicS™ multi-functionalised

silica-based thiol scavenger SPM32 (unoptimised, 0.1 weight equivalents compared to the API used).

Process Result : API isolated with minimal loss of material, without the introduction of any further impurities (as determined by full gradient HPLC) and with regulatory-acceptable Pd content of <5 ppm.

Process Scale-up : API produced on 16 kg scale using this process step.

Synopsis : Crude API sample prepared by Suzuki and Sonogashira couplings containing up to 3000 ppm of palladium.

Key Issue : Speed to initial kg quantities of API for testing.

Process Target : Isolation of the API without any significant loss and with Pd content reduced to within ICH guidelines (<5 ppm).

Scavenger Screening : PhosphonicS™ recommended and previously successfully screened and utilised scavenger employed directly.

Process Step : Treatment of the crude API sample in methanol at reflux for 1 hour in slurry mode with PhosphonicS™ multi-functionalised silica-based thiol scavenger SPM32 (0.5 weight equivalents compared to the API used).

Process Result : API isolated with minimal loss of material and with regulatory-acceptable Pd content of <5 ppm.

Process Scale-up : API produced on multi-kg scale using this process step.

Synopsis : Crude Suzuki reaction product containing 7855 ppm of palladium.

Key Issue : Isolation of the API with reduced Pd content.

Scavenger Screening : PhosphonicS™ multi-functionalised silica-based thiol scavenger SPM36 screened alongside activated charcoal and polymer-based adsorbents.

Screening Protocol : Treatment of the Suzuki reaction product in toluene at room temperature (PhosphonicS™ SPM36) or up to reflux (as recommended by the individual adsorbent supplier) for 2 hours in slurry mode (at 1 weight equivalents compared to the Suzuki product).

Screening Results : Pd removal with both activated charcoals and polymer-based adsorbents was found to be relatively poor (<20% removal) even at the higher dose rate, at which PhosphonicS™ SPM36 reduced the Pd content to just 6.3 ppm (>99.9% removal).

Synopsis : Crude heterocyclic addition reaction product containing multiple functional groups. Initial metal content : up to 686 ppm of rhodium in organic solution.

Key Issue : Reduction of metal content to set specification within timescale (24 hours) allowed for this process step.

Process Target : Within the 24-hour timescale allowed for this process step, Rh content of the solution was required to be <100 ppm (<75 ppm in the intermediate), allowing further reduction to within ICH guidelines upon subsequent synthetic steps.

Scavenger Screening : PhosphonicS™ multi-functionalised silica-based thiol scavenger SPM32f was identified as the optimum scavenger for this process stream following a metal adsorbent screening campaign including activated carbons/charcoals, polymer-based adsorbents and competitor silica scavengers.

Process Step : PhosphonicS™ SPM32f metal scavenger (30 wt% to starting material) was packed into a standard filter housing and the process stream at 60 oC was circulated through the filter bed, with periodic monitoring of the Rh content. Rapid and effective Rh removal was achieved quickly and the process was allowed to continue for between 20 and 24 hours per batch to maximise metal removal.

Process Result : Metal content of the isolated API-intermediate was found to be reduced down to <50 ppm in solution (72 ppm in the intermediate) within the 24-hour process step timescale, with no loss of intermediate indicated by appropriate analytical assays. Metal removal was consistent for each of the four 30 kg batches run. The API-intermediate was advanced to the next process step, without any delays, allowing the API batch to be delivered on time and within regulatory guidelines for metal content.

Process Scale-up : API produced on 5 kg scale using this process step.

Synopsis : Crude highly-functionalised heterocyclic compounds containing high 100s of ppm of palladium formed by Suzuki reactions, using Pd(PPh3)4 as catalyst in DME/water.

Key Issue : Isolation of compounds with reduced Pd content for initial screening purposes.

Unsuccessful Metal Removal Protocols : The samples were treated with activated charcoal (~20 weight% compared to the reaction product), subjected to an aqueous work-up and finally passed through unfunctionalised silica gel (~500 weight%). Pd removal with variations of this protocol was found to be almost completely ineffective, demonstrating the strong ability of the target compounds to bind Pd.

After such treatments, compound 1 still contained 1197 ppm Pd and compound 2 still contained 867 ppm Pd.

Successful Metal Removal Protocol : Treatment of the Suzuki reaction products in DCM/methanol at reflux for 1 hour with PhosphonicS™ STA3 (50 weight%) in slurry mode resulted in extremely impressive reductions in the metal content of the samples – compound 1 contained 110 ppm Pd (>90% Pd removal) and compound 2 contained just 27 ppm Pd (>96% Pd removal). In both cases, chemical purities remained unchanged before and after treatment with PhosphonicS™ STA3, and compound recoveries were high.

Summary : A single, unoptimised treatment of these Med Chem scale-up samples with PhosphonicS™ STA3 for 1 hour reduced the Pd content by >90%, providing samples of the desired quality for screening. Traditional metal removal protocols featuring activated charcoal, unfunctionalised silica gel and aqueous extractive work-ups proved ineffective for Pd removal from these compounds containing combinations of functional groups which tightly chelate palladium.