How to determine reversed-phase flash chromatography loading capacity

A flash column’s loading capacity varies depending on several factors but primarily the selectivity and resolution of the target compound from its closest eluting neighbors. While normal-phase flash chroma- tography loading capacity is commonly determined from thin-layer chromatography (TLC) separation data, reversed-phase loading capacity determination is typically an empirical process involving several repeated injections with increasing sample mass.

As a general rule, reversed-phase columns have lower loading capacity compared to normal-phase silica columns due to reversed-phase media’s lower available surface area (bonding C18 to silica reduces surface area) and different separation mechanism (partitioning vs. silica adsorption/desorption). If you check various flash column vendors you will find silica loading capacities up to 10% and even 20% of media weight. In contrast, typical published reversed-phase capacities are 1% - 2% of media weight. These are suggested maximum loads and not necessarily meant to be average or applicable for every purification. Again, it depends on how good your separation is along with the other criteria mentioned above.

In this document we will show how using a resolution value (Rs), calculated from small-scale empirical scouting runs, can be used to suggest the maximum sample load mass possible while maintaining a specific product purity goal.

Loading capacity

Column loading capacity is based on a number of factors…

1. Sample complexity (number of compounds)
2. Sample separation method
   
   1. Isocratic
   2. Linear gradient
   3. Step gradient
3. Sample separation (how well your target is separated from the by-products and impurities)
   
   1. Retention
   2. Selectivity
   3. Resolution
4. Solubility in…
   
   1. Mobile phase
   2. Dissolution solvent
5. Compound chemistry – neutral, acidic, basic, polar, lipophilic
6. Load technique
   
   1. Liquid
      
      1. Dissolution solvent choice
      2. Sample concentration
   2. Dry
      
      1. Sorbent used
      2. Sample/sorbent ratio
7. Purity and yield goals
   
   1. Higher purity = lower loading
   2. Higher yield = higher loading

For medicinal chemists, yield usually is more important than purity for intermediate compounds with 80+% purity deemed acceptable in many cases. Historically, medicinal chemists have utilized normal-phase flash chromatography for intermediate purification but have started to migrate towards reversed-phase as their synthetic products’ chemistry has become more polar and complex. Because of this migration to reversed-phase, understanding reversed-phase column loading capacity has increased in importance.

Method scouting

As with any chromatography, a proper method must be devel- oped that separates the targeted compound from its nearest eluting impurities. Method scouting is best performed with an analytical HPLC scaling column packed with the same media (particle size and chemistry) as the flash column (Bickler, J. Robert, 2015) or, if a scaling column and HPLC system are not available and you have adequate sample, methods can be developed on small flash columns.

For this whitepaper, the loading capacity of a 12-gram flash C18 column (Biotage® Sfär C18) was evaluated using four reaction mixtures, synthesized in-house using a Biotage® Initiator+.

After synthesis, each reaction mixture was evaporated in a tared scintillation vial providing weights of…

• Reaction 1 – 638 mg
• Reaction 2 – 545 mg
• Reaction 3 – 852 mg
• Reaction 4 – 862 mg

The contents of each vial were then dissolved in DMSO. Reactions 1 and 2 were dissolved in DMSO to a final volume of 5 mL while reactions 3 and 4 were dissolved in 2 mL. DMSO was chosen because of its broad solubility capability and very low retention in reversed-phase chromatography.

Reaction mixtures 1 and 2 were best separated using a 25–60% methanol gradient over 10 column volumes (10 CV) while reactions 3 and 4 needed a 35–85% methanol gradient over 10 CV.

Small aliquots (~10–20 mg) were injected into the 12-gram Sfär C18 column running the appropriate gradient to deter- mine how many compounds were present, where the target product eluted, and the degree of separation achievable, both selectivity and resolution. The data showed each reaction generated the product and by-products, Figure 1.
Figure 1. Scouting runs to determine resolution values. From top - reaction 1, reaction 2, reaction 3, reaction 4.

The scouting runs revealed that with reactions 3 and 4 the reaction product was the last to elute and very well separated from the earlier eluting by-products. However, the separations of reactions 1 and 2 showed that the major peak was centered between two other closely eluting peaks making target compound purification more challenging.

To determine each reaction mixture’s separation efficiency and maximum load, the target products and nearest eluting by-product peaks retention (in mL) and their peak volume (the difference in volume from the start of peak fractionation to the end of peak fractionation) measured at baseline, were obtained on the flash system (Biotage® Selekt) and tabulated, Figure 2.
Figure 2. Peak retention volume and elution volume are measured by expanding the chromatogram around the peak of interest. Elution volume is the difference between when fractionation begins and ends, retention volume is the peak apex.

This data was used to calculate the target compound’s resolu- tion from each neighbor, which was used to determine loading capacity, Table 1.

Table 1. Scouting run separation data.

Resolution is calculated as follows…
So, what does this provide in terms of determining loading capacity? Well, for starters, the larger the resolution, the higher the load. With a Rs of 2.76, the highest load (or best separation) should be seen with the reaction mixture 4. Since reaction mixtures 3 and 4 have the product eluting last, only the resolution between it and its faster eluting by-product were calculated. Reactions 1 and 2, with the target compound eluting between two by-products, required two resolution values be calculated. In these situations, the lower resolution value is the limiting factor so reaction 1, with a resolution of 0.91, should provide the lowest load. The resolution for reaction 2 is a similar 1.21.

Scaled purification runs

Since no direct conversion of resolution to load capacity exists, a series of separations with increasing load was performed to empirically determine maximum capacity. This work was done using reactions 1 and 2 injecting volumes 2x larger than the previous load until a complete resolution loss was achieved, Figure 3.
Figure 3. Scale-up chromatograms for Reaction 1 and Reaction 2 show decreasing resolution with increasing load.

Purity evaluation

The middle peak from each purification was evaporated using a Biotage® V-10 Touch, redissolved in DMSO, and analyzed for purity using the same flash method and column. The results showed increasing by-product amounts with increasing loads, a reasonable expectation.

To calculate peak purity, the peak height of each of the retained peaks were measured in mAU. Since the all-wavelength feature was used with a range encompassing the total UV absorption spectrum for each compound (200-300 nm), peak height ratios are valid measurement for comparison. The ratio of the target compound peak height to the cumulative total peak height was calculated and used as the purity measurement, Table 2.

For reaction 1 a load of 102 mg met the >80% purity goal making it the loading capacity for this crude mixture. However, for reaction mixture 2, the scaling study showed the maximum load is less than 171 mg but more than 86 mg, requiring another purification test at a load between 86 and 171 mg.

Table 2. Target compound purity results.

Calculating maximum load

To help determine the maximum load, a plot of purity vs. load shows a linear relationship useful for determining the load required to achieve the 80% goal, Figure 4.
Figure 4. Load vs. purity for reaction 2 shows that a load around 150 mg should achieve the >80% purity target.
The graph indicted a load about 150 mg would meet the purifica- tion goal. Unfortunately, only 139 mg of reaction 2 remained.

Though less than the calculated load amount, the resulting chromatography revealed a well-separated product from its adjacent by-products, Figure 5. Product purity was determined to be 89%, supporting the viability of a 150 mg maximum load, equating to a 1.3% load (crude weight/column media weight), Table 3.
Figure 5. Purification and analysis for reaction 2 with estimated maximum load. Top - 139 mg load purification shows the product, peak 7 (green), well separated from its adjacent by-products. Bottom - Fraction 7 purity analysis indicates high purity.

Table 3. Reaction 2, 139 mg load purity analysis.

Pushing the limits

The preceding research provides a pathway to determining maximum loading capacity for low resolution purifications.

Reactions 3 and 4, however, had far better resolution values and only one major by-product as a potential contaminant, a situation that enables increased loading and throughput.

Plotting the log of the load vs. the log of the resolution using the previous data’s maximum load and resolution (Rs 0.91=102 mg, Rs 1.21=150 mg) along with the resolution values and crude synthetic masses of reactions 3 and 4 generated a relationship with enough linearity (R2=0.9004) to test purification of each reaction’s total crude mix, Figure 6.
Figure 6. Log load vs. log resolution plot shows enough linearity to predict loading capacities for reactions 3 and 4.

Each crude mixture was loaded in its entirety (RxN 3 = 852 mg and RxN 4 = 862 mg) using the same 12-gram Sfär C18 column, Figures 7 and 8. The resulting purification yielded product purities of 86% for reaction 3 and 95% for reaction 4 indicating a loading capacity for both crude mixtures in excess of 7%, Table 4.
Figure 7. Reaction 3 purification (852 mg) and purity analysis. Top – crude reaction mix purification with product eluting in fractions 4 and 5 (blue peak). Bottom – analysis of fractions 4 + 5 indicates a high purity level.
Figure 8. Reaction 4 purification (862 mg) and purity analysis. Top – crude reaction purification with product in fractions 4 and 5 (pink peak). Bottom – Fraction 4 and 5 analysis showing very high product purity.

Table 4. Post-purification purity assessment for reactions 3 and 4.

As predicted, the highest load (862 mg) was achieved with the best resolution (2.76) and provided product purity of 95%.

During purification of reactions 3 and 4 an interesting phenomena was occurred - the product immediately began crystalizing in the collection vessels, an indication of high fraction purity, Figure 9.
Figure 9. Reactions 3 and 4 crystalizing during fractionation.

Conclusion

Reversed-phase flash chromatography loading capacity can be determined for crude mixtures from just a few small-scale purifi- cation runs by calculating the resolution of the target compound from its closest eluting by-products. High loading capacity is achievable with reversed-phase flash chromatography when resolution is maximized.

In the examples shown above, loads of ~7% were possible with two crude reaction mixtures when product resolution was >1.90 and DMSO was used as the reaction mixture dissolution solvent.

Literature Number: UI468