Julia
Urbiña-Alvarez†
a,
Daniela
Torres-Ruiz
a,
Camilo
Mahecha-Mahecha†
a,
Gabriel
Hernandez-Abdallah
a,
Mario
Macías
b and
Diego
Gamba-Sánchez
*a
aLaboratory of Organic Synthesis, Bio and Organocatalysis, Chemistry Department, Universidad de Los Andes, Cra. 1 No. 18A-12 Q:305, Bogotá 111711, Colombia. E-mail: da.gamba1361@uniandes.edu.co
bCrystallography and Chemistry of Materials, CrisQuimMat, Department of Chemistry, Universidad de Los Andes, Bogotá D.C. 111711, Colombia
First published on 17th October 2025
Here we describe a masked formylation strategy based on the activation of 1,3-dithiane oxide with oxalyl chloride. The reaction proceeds through a thionium ion intermediate, enabling efficient electrophilic aromatic substitution with electron-rich arenes and heteroarenes. The method shows good tolerance toward alkyl-, hydroxyl-, and alkoxy-substituted phenols, as well as indoles, while thiophenols react preferentially at sulfur, supporting a Pummerer-type pathway. In addition, we developed a practical oxalyl chloride-mediated dithiane cleavage protocol, affording aldehydes in good to excellent yields. Dithiolanes were also reactive but generally less efficient. Overall, this work introduces oxalyl chloride as a versatile reagent for masked formylation and dithiane cleavage, providing operationally simple and broadly applicable synthetic methodologies.
Among the most widely used protecting groups for aldehydes is dithiane.4 It is favored not only because its preparation is more straightforward than that of acetals, but also because it enables the well-known umpolung process, serving as a nucleophilic equivalent of the formyl group (Corey-Seebach reaction).5 Consequently, the development of new strategies for introducing formyl groups, as well as efficient methods for dithiane cleavage, remains highly appealing to the scientific community.
Both processes, however, present significant challenges. In the case of formylation, several methods have been reported in the literature, but most require harsh conditions, toxic reagents, or the in situ generation of highly electrophilic intermediates under specific conditions. Among the classical formylation reactions (Scheme 1.1) are some of the most traditional transformations in organic chemistry, such as the Rieche,6 Reimer-Tiemann,7 Gattermann,8 Vilsmeier-Haack,9 Gattermann–Koch,10 and Duff11 reactions. All of these rely on the in situ generation of highly electrophilic species to achieve formylation of strongly activated substrates, while only a few examples are effective with deactivated substrates. Recent developments include modifications of the Rieche reaction,12 the innovative use of MeOH13 or Me3N14 as formylating agents, adaptations of the Gattermann–Koch reaction,15 and the use of glyoxylic acid,16 among others.17 The relevance of these approaches has been discussed in recent reviews.3
![]() | ||
| Scheme 1 Comparison between (1) classic formylations (2) masked formylations and (3) the current work. | ||
The direct introduction of dithianes or acetals onto aromatic rings can be regarded as masked formylations, and in this context several noteworthy advances deserve mention, such as the use of thionium ions (Scheme 1.2a and b)18 and the Minisci reaction.19 All of these methods perform well with electron-rich aromatics, as they proceed through electrophilic aromatic substitution.
Regarding dithiane cleavage, although alternative strategies have recently been developed, the most traditional approach still relies on mercury salts.20 Other methods include the use of H2O2/I2,21 hypervalent iodine species,22 Selectfluor,23 and, more recently, a photoinduced iodine-based process.24 While these alternatives represent valuable contributions, there remains a strong need for simpler and more practical methodologies.
Motivated by these challenges, we present herein a masked formylation protocol based on 1,3-dithiane oxide, which is presumed to proceed via a classic Pummerer reaction25 (reaction of a thionium ion with a nucleophile) after activation with oxalyl chloride. In addition, we describe an alternative dithiane cleavage method that uses oxalyl chloride as the sole reagent under straightforward conditions.
| Entry | 1 (equiv.) | (COCl)2 (equiv.) |
3a b |
4a b |
2a b |
|---|---|---|---|---|---|
| a 2a was always the limiting reagent. b Proportions calculated by GC-MS of the crude mixture. c Yield of isolated pure product, only purified when its proportion was greater than 75%. d Reaction performed at room temperature. e Reaction performed at −10 °C. N.D the product was not detected by GC-MS of the crude reaction mixture. | |||||
| 1 | 1.00 | 1.00 | 60 | <5 | 14 |
| 2 | 1.00 | 1.20 | 75c | <5 | 8 |
| 3 | 1.00 | 1.50 | 25 | 17 | 7 |
| 4 | 1.00 | 1.80 | 24 | 35 | 5 |
| 5 | 1.00 | 2.00 | 4 | 41 | N.D |
| 6 | 1.10 | 1.32 | 72 | <5 | 9 |
| 7 | 1.20 | 1.44 | 55 | <5 | 7 |
| 8 | 1.26 | 1.51 | 59 | <5 | 6 |
| 9 | 1.35 | 1.63 | 85 | <5 | 6 |
| 10d | 1.35 | 1.63 | 62 | <5 | <5 |
| 11e | 1.35 | 1.63 | 65 | <5 | <5 |
We then initiated optimization studies by increasing the amount of oxalyl chloride (entries 2–5). Although a moderate improvement in the yield of 3a and in overall conversion was observed with 1.2 equivalents (entry 2), further increases in oxalyl chloride resulted in decreased formation of 3a and a gradual increase in 4a (entries 3–5). At this stage, the formation of 4a was only anecdotal, and we chose not to pursue its optimization, focusing instead on the isolation of 3a. Studies directed toward the synthesis of 4a are presented later in this manuscript.
Next, we slightly increased the equivalents of 1 while keeping the stoichiometry of oxalyl chloride constant and retaining naphthol as the limiting reagent (entry 6), which afforded results similar to those of entry 2. However, a systematic increase of both 1 and oxalyl chloride (entries 7–9) led to a significant improvement, providing pure 3a in up to 85% yield with nearly complete conversion and only traces of 4a detected (entry 9). Finally, we evaluated the reaction at room temperature (entry 10) and at −10 °C (entry 11), but no further improvements were observed. Consequently, we selected the conditions of entry 9 as optimal.
With the optimized conditions in hand, we next evaluated the reaction scope, and the results are summarized in Scheme 2. Using m-cresol (2b) as the substrate, we obtained a mixture of regioisomers arising from substitution at the 2- and 4-positions, with the major product 3b isolated in 21% yield and the minor isomer 3b′ in 14% yield. These results were somewhat disappointing, given that the optimization had previously afforded yields above 80%, and we anticipated better outcomes with activated phenols. The combined yields of compounds 3b and 3b′ indicate essentially the same reactivity as the other alkyl phenols such as o-cresol, which provided 3c in 30% yield without the formation of regioisomers.
To further enhance the reactivity, we examined more strongly activated phenols bearing additional alkyl substituents. In the case of phenol 2d, the corresponding product 3d was obtained in 34% yield, while gratifyingly, phenol 2e afforded 3e in 76% yield. At this stage, the method demonstrated limited applicability, as phenols with additional activating groups (through inductive effects) provided products in yields ranging from poor to good.
We then turned our attention to phenols bearing additional electron-donating groups such as OH or OMe. To our delight, catechol (2f) afforded product 3f in 54% yield, demonstrating that the introduction of additional electron-donating substituents positively influences the reaction outcome. Similarly, m-methoxyphenol (2g) gave 3g in 54% yield. Interestingly, only a single regioisomer was obtained, consistent with the directing effect of the more activating substituent, as expected in electrophilic aromatic substitutions. Improved yields were observed with more activated trisubstituted substrates (2h and 2i), while 1-naphthol (2j) provided better yield that its less activated methyl derivative (2k).
Next, we replaced the activating substituent with a nitrogen group. The corresponding product 3l was obtained in acceptable yield; however, when free anilines were employed, no reaction occurred (vide infra). Activated heterocycles also proved to be suitable substrates: for example, indole (2m) delivered 3m in good yield.
We then examined thiophenols as substrates and obtained markedly different results. In these cases, products 3n, 3o, and 3p were isolated, arising from nucleophilic attack of sulfur on the thionium ion rather than electrophilic aromatic substitution, interestingly the reaction proved completely chemoselective since no product of electrophilic aromatic substitution was observed. These findings provide strong evidence for the intermediacy of a thionium ion and support a Pummerer-type reaction pathway, regardless of the final product formed. The structures of 3n, 3o, and 3p were confirmed by X-ray crystallographic analysis (see SI for details).
Finally, certain substrates proved unreactive. Strongly deactivated aromatics did not undergo the transformation, which may also explain the lack of reactivity observed with free anilines. Since the reaction is performed under acidic conditions, sufficient HCl is present to fully protonate the amine, thereby deactivating the aromatic rin; in addition, apparent oxidation was observed, as the reaction mixture turned dark—likely due to the formation of highly conjugated species. Unfortunately, no isolable product was detected based on the 1H NMR analysis of the crude mixture. More complex structures, such as coumarins, were likewise unreactive.
| Entry | (COCl)2 (equiv.) | Solvent | 4q |
|---|---|---|---|
| a Reaction performed at 0 °C. b Reaction performed at −10 °C. N.D. As starting material was observed the reaction was not purified. | |||
| 1 | 1.00 | CH2Cl2 | N.D. |
| 2 | 1.50 | CH2Cl2 | N.D. |
| 3 | 2.00 | CH 2 Cl 2 | 83 |
| 4 | 2.20 | CH2Cl2 | 80 |
| 5 | 2.50 | CH2Cl2 | 80 |
| 6 | 2.00 | Toluene | — |
| 7 | 2.00 | MeCN | — |
| 8 | 2.00 | THF | — |
| 9a | 2.00 | CH2Cl2 | 82 |
| 10b | 2.00 | CH2Cl2 | 82 |
Optimization began by varying the amount of oxalyl chloride. With 1.0 or 1.5 equivalents (entries 1 and 2, respectively), no complete conversion was observed, and the product was not purified. Using 2.0 equivalents, however, the starting dithiane 3q was fully consumed within 30 minutes. Nevertheless, the isolated yield after purification was only ∼30%. We observed the formation of a polar byproduct with polymer-like characteristics. Allowing the reaction mixture to stir for 24 hours afforded product 4q in 83% yield (entry 3). Increasing the amount of oxalyl chloride did not improve the yield or shorten the reaction time (entries 4 and 5).
We next evaluated different solvents (entries 6–8), but in all cases the reaction failed to proceed. Lowering the reaction temperature (entries 9 and 10) gave essentially the same results. Therefore, the conditions described in entry 3 were selected as optimal.
Applying the optimized conditions (see Scheme 3 for the results), we first examined compound 3r, which bears a chlorine atom instead of the bromine used during optimization. As expected, the reaction proceeded smoothly, affording product 4r in 93% yield. In contrast, with the more substituted substrate 3s, the corresponding product 4s was obtained in only 40% yield—remarkably low for a deprotection reaction. We therefore shifted our focus to simpler aromatic aldehydes, and products 4t, 4u, and 4v were isolated in 64%, 83%, and 84% yields, respectively, demonstrating good compatibility of the method with simple substrates.
Similarly, compounds 4w, 4x, and 4y were obtained in good yields, further supporting the robustness of the protocol.
For broader applicability, an effective deprotection method must also be chemoselective in the presence of other protecting groups. We therefore tested substrates bearing common oxygen- and nitrogen-protecting groups. Gratifyingly, the method proved compatible with benzyl ethers (4x), esters (4aa), and silyl ethers (4ab and 4ac), consistently delivering the desired aldehydes in good to excellent yields. Furthermore, the protocol tolerated carbamate protecting groups, with both Cbz (4ad) and Boc (4ae) derivatives giving the expected products.
Regarding unreactive substrates, the reaction profile paralleled that of the masked formylation protocol: compounds bearing free amines or strongly electron-withdrawing substituents failed to react. Unfortunately, the method was also incompatible with dithianes derived from aliphatic aldehydes or ketones.
To conclude our study, we sought to evaluate the selectivity of our method against other functional groups commonly employed as aldehyde protecting groups. Unfortunately, the reaction proved to be non-selective, as we observed cleavage of both acetals 5 (see Scheme 4a) and dithiolanes 6 (see Scheme 4b).
The reactivity of acetals under these conditions can be attributed to the strong acidity of the reaction medium, while dithiolanes are expected to react analogously to dithianes.
To further explore this reactivity, we prepared several dithiolane derivatives 5 and tested them under the optimized conditions for dithiane cleavage. Gratifyingly, the corresponding aldehydes were obtained in all cases, albeit in slightly lower yields compared to those derived from dithianes (see Scheme 5).
Dithiolanes with halogenated rings (5q and 5s) afforded low yields of the corresponding aldehydes (4q and 4s). The yield of 4s was comparable to that obtained from dithiane 3s; however, the yield of 4q was much lower, suggesting that dithiolanes are generally less reactive than dithianes. This trend was also observed in other cases—for example, 4t (29% vs. 64%) and 4v (67% vs. 87%). Fortunately, for compounds 4x and 4z, the yields from dithianes and dithiolanes were virtually identical.
Providing an explanation for the marked differences in yields between dithiane and dithiolane cleavage is challenging, though the reaction mechanism likely plays a role. As noted above, the starting materials were always fully consumed within 15 minutes, but isolating and purifying the products at this stage gave significantly lower yields than when the reaction mixture was stirred for 24 hours. Attempts to identify the intermediate were unsuccessful; however, we assume that a polymer-like species is formed, which is more stable when derived from dithiolanes than from dithianes. Cleavage of this intermediate by water generates the final product, a process that appears to proceed more slowly in the case of five-membered rings. Further experiments, supported by computational studies, are ongoing in our laboratory and will be reported in due course.
In addition, we established a practical protocol for the cleavage of dithianes using oxalyl chloride as the sole reagent under simple reaction conditions. This transformation proved robust, chemoselective in the presence of several common protecting groups, and applicable to a broad range of aromatic dithianes. While dithiolanes were also reactive under the optimized conditions, they generally afforded lower yields, suggesting differences in intermediate stability and reactivity.
Taken together, these results expand the toolbox for formylation chemistry and deprotection methodologies, providing a straightforward and efficient protocol for both masked formylation and dithiane/dithiolane cleavage. Ongoing mechanistic and computational studies are expected to further elucidate the reaction pathway and guide future developments.
CCDC 2486107–2486109 (3n–3p) contain the supplementary crystallographic data for this paper.27a–c
Footnote |
| † Current address: Escuela de Medicina, Fundación Universitaria Juan N. Corpas, Bogotá 110311, Colombia. |
| This journal is © the Partner Organisations 2026 |