Integrated plug flow synthesis and crystallisation of pyrazinamide

We report the integration of flow chemistry with plug flow crystallisation. Catalytic flow hydration of pyrazinecarbonitrile to pyrazinamide was performed in a packed bed column of MnO2. The effluent of this flow reactor was directly linked to a tri-segmented tubular crystalliser (KRAIC), providing a seamless transition from flow synthesis to crystallisation, with control over solid form and particle characteristics.

Pumping and temperature control unit for flow synthesis -Vapourtec R2+ and R4 Air and carrier fluid pumps -Ismatec gear pump REGLO-Z Circulating baths for tubing jackets -Grant TC-120-R1  During periods of blocking a heat gun was used to partially dissolve the material blocking the tubing.

Controlled nucleation
These experiments were identical to the uncontrolled nucleation experiments excepting a solution concentration of 0.28M and the inclusion of a 1.2 m length of 3.2 mm ID FEP jacketed tubing between the outlet of the water bath and the main crystalliser tubing.A second circulating bath was used to control the temperature within the tubing jacket (set to 10 °C).A zero volume union was used to connect this tubing to the main crystalliser tubing.No blocking occurred within this set-up.

Synthesis only experiments
In order to ascertain the % conversion a flow synthesis only run was performed with identical parameters to the integrated runs but collecting the product immediately after the BPR.In three runs with 20 ml of 0.28 M aqueous pyrazinamide solution, the solvent was evaporated giving an average isolated yield of 0.6871 g (91.7 %) measured.NMR and DSC (see below) confirmed the purity of the product.

Slug size analysis
Slug volume was measured through 11 collections of 10-28 slugs.N.B.The actual volume of the slugs will be slightly lower than the calculated volume of a cylinder due to wetting effects at the front and rear of the slug.The measured average volume of 0.075 ml is therefore in very good accordance with the calculated 0.08 ml.

Particle size analysis
Due to the large size, undesirable particle attributes and low yield of the uncontrolled nucleation runs, the particle sizing was performed through image analysis similar to the slug size analysis.Table S1 details the average length and breadth of 30 crystals analysed from four collections using a microscope grid slide with minor divisions of 0.1 mm.The surface area was calculated assuming rectangular shapes.Particle size analysis of the controlled nucleation runs was performed using laser diffraction on a Malvern Mastersizer 3000.The particles were suspended in a saturated solution of pyrazinamide in isopropanol to avoid particle dissolution.A manual Malvern Hydro SM dispersion cell (stirring rate 2000 rpm) was used to disperse the sample as the needle like particles have a tendency to aggregate.After 30 mins, the majority of these aggregates had dispersed (Figure S6) without damage to the particles.This methodology was successful for disaggregation of the larger aggregates (>1000 μm) but smaller aggregates remained.All measurements discussed further were taken once sample stability was achieved (numbers 20 to 50).The Malvern Mastersizer 3000 software version 3.36 fits the diffraction data to an irregular sphere.

Figure S5 Images used for particle size analysis of pyrazinamide crystals resulting from uncontrolled nucleation runs
As the particles here are needle shaped, it can be presumed that a combination of all orientations of the needle crystals with respect to the laser beam will be represented in the data.A higher proportion of the long axis parallel to the flow (perpendicular to the laser) can be expected for needle crystals in flow.
The lower size range shown in Figure 7 corresponds to particles presenting parallel to the laser, hence projecting only their width, measured between 3-20µm.All other orientations of the particles are detected in the range of 20-300µm, maximum population expression at 100µm.The maxima at 100 μm most likely corresponds to the particles aligned with the flow in the cell, showing their actual length, this is in accordance with optical microscopy observations (Figure S3).
The sizes detected over 300µm correspond in part to occasional long particles, but mainly to aggregates breaking and reforming over the 30 measurements.One can indeed observe that the regions of the size distributions corresponding to the particles themselves are superimposed, showing their stability, while this region shows a very high variability between records, which is explained by the fact these aggregates are not stable and will break down then reform randomly over time.These results are confirmed by SEM analysis from two separate controlled nucleation experiments using a Jeol JSM-6610LV instrument.Figure S8 below shows an average range of particles between 100-300 μm with some larger particles present.2.6 1 H NMR All NMR analysis was carried out using a 400 MHz Bruker Advance

Controlled nucleation product
Prior to analysis, the recovered solid product from a controlled nucleation run was dried overnight.

Synthesis only product
The product obtained after evaporation of solvent was oven dried before being prepared for NMR analysis without purification.N.B. the excess water in the sample is due to wet d6-DMSO.

1. 1 1 Figure
Figure S1 Custom water bath used for the segmentation unit

Figure
Figure S2 Images from an uncontrolled nucleation run.N.B. major gridlines are 1 mm, minor gridlines 0.1 mm

Figure
Figure S4 Images of slugs in coils 1 and 2 used for average slug size analysis, red arrows indicate the lengths measured

Figure
Figure S6 D10, D50 and D90 in microns as a function of the experiment number.Each record takes around 20 seconds to acquire.Large aggregate dispersal and thus sample stability is achieved at 30 min (record 20).

Figure
Figure S7 Volume density measurements as a function of size class, superimposing records 20 to 50.

Figure
Figure S8 SEM analysis on pyrazinamide crystals produced from controlled nucleation experiments

Figure
Figure S10 DSC analysis of products obtained from synthesis only ('filtrate'), controlled nucleation and uncontrolled nucleation ('blocked run') experiments.N.B. due to the small percentage of α-pyrazinimide present in the 'blocked run' sample the heat flux has been magnified 10x to show the endothermic recrystallisation event)

Figure
Figure S11 NMR spectrum of the product from a controlled nucleation crystallisation run

Table of Contents
Apparatus: Table S2 slug volumes of captured solvent slugs for average slug volume calculations Slug lengths were calculated through measuring images of the slugs within coils.The pixel number to the tube size was normalised for every slug to account for perspective.

Table S3
slug length calculations from pixel widths

Table S5
Particle size analysis from pixel width calculations from images with microscope grid slide