*2.4. Work-Up Investigations*

As the last point, we investigated some different work-up approaches, in order to (*i*) compare the results (also in terms of sustainability), and (*ii*) establish if part of the organocatalyst could be easily recovered. The first 100 mmol-scale aldol condensation (Table 6, 45 min long aldehyde addition) was stopped after 49 h and the reaction mixture (total volume = 79 mL) was divided in six portions, treated as described in Table 8.



1 Aldol condensation carried out on 100 mmol of **2d,** reaction conditions described in Scheme 2, reaction stopped after 49 h, reaction mixture (total volume = 79 mL) divided in six portions and treated with six different work-up methods. 2 Determined by 1H NMR on the crude mixture. EtOAc = ethyl acetate, Et2O = diethyl ether, DCM = dichloromethane.

The first portion of reaction mixture (18 mL, 22.8 mmol) was filtered through a short pad of silica to remove water and proline (method **A**, Table 8). EtOAc was used as mobile phase to elute product **3ad** (along with residual reagents **1a** and **2d**). Despite the significant polarity of EtOAc, a large amount of solvent was required to recover all the product and an undesirable high volume of organic solvent (242 mL) had to be evaporated under reduced pressure.

The second portion of reaction mixture (18 mL, 22.8 mmol) was subjected to a typical aqueous work-up to remove water and proline (method **B**, Table 8). NH4Cl (2 equivalents with respect to proline, solved in 20 mL of H2O) was employed to quench proline, the two phases were separated, and the aqueous phase was extracted two additional times with EtOAc, until complete recovery of the product (checked by thin-layer chromatography). The solution was dried with Na2SO4, which restrained a significant amount of aldol product **3ad**, so that it was necessary to wash it three times with EtOAc. A considerable volume of organic solvent (90 mL) had to be evaporated.

At this point, we tried to develop a work-up method that allowed us in a simple way not only to remove the catalyst, but also to recover it, at least partially. Exploiting the very low amount of protic polar solvents used in our MeOH/H2O/(*S*)-proline-based protocol, we envisaged that the addition of a small portion of organic solvent could make the reaction environment sufficiently lipophilic to trigger the catalyst precipitation. Four different organic solvents were tested: EtOAc (method **C**, Table 8), Et2O (method **D**), dichloromethane (method **E**), and *n*-hexane (method **F**). The minimum amount of solvent able to provide an opalescent solution was added to each portion (10 mL, 12.6 mmol) and the mixtures were stored at −15 ◦C for 36 h. In the portions treated with EtOAc and Et2O (methods **C** and **D**, respectively), a white precipitate, corresponding to proline, was clearly observed; therefore, it was filtered under vacuum and washed with a small amount of cold solvent. The filtered solutions were stored at −15 ◦C for further 36 h and a second portion of catalyst was recovered in both cases. Afterwards, the mixtures were dried with Na2SO4, which was required to be washed three times with organic solvent. In these two cases (methods **C** and **D**), a considerable amount of organic solvent also had to be evaporated (46 and 52 mL, respectively). In the portions treated with DCM and *n*-hexane (methods **E** and **F**, respectively), two liquid phases were observed and we decided to directly use Na2SO4 to remove the small water-based phase. Na2SO4 was washed with solvent until complete recovery of the product (three times for DCM, four times for *n*-hexane). In these two cases (methods **E** and **F**), the lowest amounts of organic solvent were employed (36 and 43 mL, respectively).

By comparing the results obtained with the tested work-up methods, we can infer the following: (1) no work-up approach adversely affects reaction conversion and diastereoselectivity, in fact all the crude mixtures showed comparable good results (Table 8). (2) The least suitable and sustainable method to remove water and proline seems to be the silica-pad (method **A**), owing to the large amount of solvent required to recover all the desired product; (3) when two organic and aqueous phases are formed (methods **B**, **E** and **F**), the simplest and cheapest work-up appears the dilution with a very small amount of DCM, cooling, and directly drying with Na2SO4 (method **E**). This approach is practicable only thanks to the very low amount of protic polar solvents used in our MeOH/H2O/(*S*)-proline-based protocol. (4) Among the tested work-up approaches, the most convenient are those allowing an easy recovery of a large part of the organocatalyst (methods **C** and **D**). In particular, method **D** employing Et2O reached 86% of proline recovery using an acceptable volume of organic solvent. Moreover, this result is obtainable on the basis of very low amount of protic polar solvents used in our protocol.

Although the tested work-up methods are not optimized and, therefore, can be further improved, they give a clear indication of the advantages that our protocol can offer.

### **3. Materials and Methods**
