1
Photochemical and Substrate-Driven CO 2 Conversion

Bart Limburg1, Cristina Maquilón1,2, and Arjan W. Kleij1,3

1Institute of Chemical Research of Catalonia (ICIQ), The Barcelona Institute of Science and Technology (BIST), Av. Països Catalans 16, 43007, Tarragona, Spain

2Universitat Rovira i Virgili, Departament?de Química Física i Inorgànica, Marcel·lí Domingo s/n, 43007, Tarragona, Spain

3Catalan Institute of Research and Advanced Studies (ICREA), Pg. Lluis Companys 23, 08010, Barcelona, Spain

1.1 Introduction

The use of carbon dioxide (CO2) as a raw material in molecular science has been the subject of many investigations. Obviously, the replacement of fossil fuel-based chemistries with those primarily based on CO2 cannot alleviate the challenges we are facing in terms of global carbon emissions and managing the carbon cycle. However, new technologies that can help to partially replace the nonsustainable feedstock into renewable and widely available ones will help to transition to a circular rather than a linear economy [1]. In this regard, technologies encompassing the use of catalysts have demonstrated that the valorization of CO2 is feasible, offering many opportunities in the areas of organic [2-5], polymer [6-8], and fuel-based chemistries [9, 10].

The use of CO2 as a reagent in nonreductive coupling reactions (i.e. after integrating the CO2 molecule into an organic substrate, the oxidation state of the carbon center, +4, remains unchanged) has been prominent in the wider area of CO2 catalysis. In this respect, the [3?+?2] cycloaddition reaction of CO2 and epoxides [11-15], and to a minor extent oxetanes [16, 17], has been among the most widely studied transformations. Conventionally, the formation of cyclic carbonates is carried out using phosgene as a reagent (Figure 1.1a), and obviously, finding more sustainable alternative routes has been the subject of intense studies over the past 20?years. Substantial progress has been noted in the synthesis of cyclic carbonates and the required catalysts for these [3?+?2] cycloaddition reactions, and nowadays, a variety of epoxides including terminal [18-20] and the more challenging internal ones with multiple substituents can be readily utilized (Figure 1.1b) [21-28]. Notwithstanding, there are still important issues to resolve in order to further advance the sustainability of these kinds of nonreductive CO2 conversions in terms of reaction conditions (preferably using ambient conditions) [29, 30], catalyst structures (preferably halide-free ones) [31], and expansion of the portfolio of cyclic carbonate compounds by using conceptually different approaches [32, 33].

Recently, various halide-free methodologies have been reported (Figure 1.1c) [31,34-40], which are important to reduce both operational cost and corrosion issues where typical binary catalyst systems (i.e. a combination of a Lewis acidic complex and a halide additive) are used. For instance, North and coworkers used a bimetallic, O-bridged Al(III)salen complex that is able to induce insertion of CO2 into one of the Al─O bonds, thereby forming an Al- carbonate intermediate [34]. This nucleophilic species further engages with the epoxide substrate to induce ring opening to eventually give the cyclic carbonate product essentially in the absence of any cocatalytic halide. In a more recent contribution, the same authors reported the use of a bis-phenol-type salen organocatalyst that is able to induce formation of cyclic carbonates from terminal epoxides and CO2, albeit at rather elevated reaction temperatures [35]. This work nicely builds on previous success in this area using multiphenolic (binary) organocatalysts as effective systems for cyclic carbonates derived from internal and terminal epoxides [41, 42]. Replacing salen diphenol with other types of H-bond activators such as a combination of DBU/L-histidine (DBU = 1,8-diazabicyclo[5.4.0]undec-7-ene) also enables a halide-free synthesis of these CO2-based heterocycles [36]. The design of other halide-free systems (with this particular design characteristic although still in its early stage) clearly demonstrates a shift toward the use of more sustainable catalysts in the valorization of CO2 into cyclic carbonates.

What should be the next step in the development of efficient catalysts for cyclic carbonate products (Figure 1.1d)? The recent literature testifies that the use of cyclic carbonates and related precursors to build more complex molecules [43-45] can only be carried out if the former can be prepared with a certain degree of substitution and functionality. Therefore, new approaches are desirable that can meet the growing need for a larger diversity in cyclic carbonate scaffolds amplifying their role as key substrates in a much wider variety of transformations compared to the recent state of the art. This chapter will thus focus on two rather new developments in the area, viz. the photochemical assisted conversion of CO2 into cyclic carbonates/carbamates and related heterocycles and the use of substrate-triggered CO2 conversion as efficient and novel strategies affording the heterocyclic targets through different manifolds and reactive intermediates (Figure 1.2). Apart from the main characteristics of the involved protocols, relevant details encompassing the key mechanistic intermediates will also be highlighted.

Figure 1.1 Evolution of the (catalytic) synthesis of cyclic carbonates and future perspective discussed in this chapter. (a) Conventional, (b) contemporary, (c) recent, and (d) current.

Figure 1.2 Different approaches to arrive at cyclic carbonates using functional (homo)allylic precursors and various substrate-involved activation strategies.

1.2 Iodine Activation of (Homo)Allylic Substrates

Molecular iodine (I2) is one of the most widely used iodinating reagents and particularly readily forms reactive iodonium species with alkenes that can afford various addition products in the presence of suitable nucleophiles [46]. Recently, various research groups have used I2 or other electrophilic I-containing reagents in reactions that utilize CO2 and comprise of a formal addition of an activated form of CO2 (typically being a carbonate/carbamate intermediate) to I+-activated double bonds present within the same substrate.

The first synthesis to use electrophilic, I+-directed double bond activation in the formation of either five- or six-membered cyclic carbamates (i.e. oxazolidinones and oxazinones) from CO2 and (homo)allylic amines follows a well-known strategy that has been previously used for the synthesis of linear and cyclic carbamate derivatives by Yoshida and Saito, respectively [47, 48]. The approach reported by Toda et al. [49] starts with the in situ formation of ammonium carbamate salts from allylic amines and CO2 at atmospheric pressure and temperature (Scheme 1.1). After the formation of ammonium carbamate, iodine is added to presumably form an iodonium intermediate, after which intramolecular cyclization by addition of the carbamate onto the activated allylic double bond results into an oxazolidinone product. The use of homoallylic substrates affords six-membered oxazinones in a similar manner (Scheme 1.1). Both primary and secondary amines were used, but the yields of the heterocyclic products were moderate. Improved yields of the products were achieved by long reaction times and a stoichiometric amount of cesium carbonate.

Scheme 1.1 Approach to five- and six-membered heterocycles using CO2 as a reagent and I2 as an olefin activator. Reported yields are those in the presence of Cs2CO3 as an additive.

Later, Muñoz and coworkers developed a similar methodology for the synthesis of five- and six-membered cyclic carbamates using (homo)allylic amines, CO2, I2, and a guanidine base (2-phenyl-1,1,3,3-tetramethylguanidine, PhTMG ), see Scheme 1.2 [50]. The utilization of this non-nucleophilic and strong base led to higher yields and significantly shortened reaction time compared to the work by Toda. The authors propose that PhTMG does not only act as a base but also stabilizes the intermediate carbamate anions. In addition to this, and as may be expected, the regioselectivity of the carbamate attack onto the I-activated double bond follows a preference for smaller sized ring heterocycles, i.e. five-membered carbamates are produced from allylic amine precursors, whereas six-membered products are formed from homoallylic substrates. In most of the reactions studied, addition of the carbamate to the double bond takes place in an anti-manner, thereby locking the final stereochemistry of the heterocyclic product. This was shown by elimination reactions induced by Ag2O and analyzing the resultant alkene products by nuclear Overhauser effect (NOE) NMR spectroscopy or in some of the cases by crystallographic...