Introduction

The bis-allylphenol compound 2 (4,4′-isopropylidenebis(2-allylphenol)) is known as an important industrial raw material. It is mostly used as co-monomer for the synthesis of high temperature resistant polymers and as reactive diluent to improve processability and toughness of various resins, high-performance composites, and adhesives [1,2,3,4,5,6]. It is typically manufactured from the corresponding diallyl ether (1) via an aromatic Claisen rearrangement (Scheme 1) [7,8,9]. Such pericyclic reactions of allyl phenyl ethers are initiated at high temperatures and proceed quite slowly [10]. Further challenges are related to the exothermicity of the reaction that may lead to safety and operational issues at large scale owing to poor heat transfer and thermal gradient formation, or in extreme cases, even to thermal runaways [11]. These characteristics may also lead to the formation of various side products (e.g., isomers and oligomers) and entail general scalability issues [12].

Scheme 1
scheme 1

Large-scale batch synthesis of 2 via aromatic Claisen rearrangement

The ton-scale Claisen rearrangement of diallyl ether (1) is typically performed under solvent-free conditions at high temperatures [7,8,9]. The reaction requires careful thermal regulation due to the exothermic nature of the transformation (Scheme 1) [13]. Aromatic Claisen rearrangements have been shown to benefit considerably from continuous flow process conditions [14,15,16,17,18,19,20,21,22,23]. Such reactions can be chemically intensified applying a high-temperature/pressure flow regime [24,25,26], whilst the safety and selectivity of these transformations can often be improved owing to better heat transfer and control over key process parameters [27,28,29,30].

In this study, our aim was to develop a lab-scale continuous flow process for the synthesis of bis-allylphenol 2 via thermal Claisen rearrangement of the corresponding diallyl ether, 1. As the planned protocol should provide the basis to develop scale-up options capable for manufacturing purposes under flow conditions, we set numerous expectations regarding the process development. These were as follows: i) reducing batch reaction times to short residence times in order to ensure high productivity, ii) improvement of safety of the exothermic transformation whilst ensuring facile scalability, iii) maximizing chemical selectivity in order to obtain 2 in sufficient purity without the need for purification steps, iv) ensuring solvent-free conditions without precipitation and clogging of the reaction channels, and v) maintaining robust reaction conditions that are potentially amenable not only to lab-scale but also to potential manufacturing purposes.

Results and discussion

In order to acquire a basic understanding of the reaction parameters, preliminary batch experiments were performed first under microwave (MW) heating conditions (see Fig. S1 in the Supporting Information for heating profiles under batch MW conditions) [25]. Initially, reactions were carried out in 0.1 or 0.5 M solutions using toluene as solvent. Almost no product formation (14% conversion, 2% selectivity) was detected at a reaction temperature of 190 °C within 90 min of reaction time (Table 1, entry 1). However, by increasing the temperature to 240 °C, we could easily reach quantitative conversion and a selectivity of 74–88% (entries 2–4). The MW reaction was next attempted under neat conditions. Gratifyingly, by optimizing the reaction time and the temperature, it was possible to obtain complete conversion and selectivities of > 80% (entries 5–10). In general, at temperatures of 280–300 °C, the precise fine-tuning of the reaction time proved very important. For instance, at 280 °C and 30 min reaction time, a conversion of 93% was obtained, but the chemical selectivity of was only 23% (entry 5). In contrast, the same reaction gave similar conversion and 85% selectivity in case of 7 min reaction time (entry 7). The best MW batch results were achieved at 300 °C within 1 or 2 min of reaction time (entries 9 and 10). Compounds 1 and 2 are both oils at room temperature. However, possibly due to uncontrollable oligomerization, the formation of insoluble solids was observed in some of the batch reactions under neat conditions (entries 5 and 6; Fig. 1). This is an indication of potential technical difficulties for the flow method development employing extreme temperatures.

Table 1 Preliminary MW batch screening of reaction conditions
Fig. 1
figure 1

Physical appearance of reaction mixtures obtained under different MW conditions. A: Table 1, entry 5; B: Table 1, entry 6; C: Table 1, entry 7

For the continuous flow process development, a standard setup was assembled using a Phoenix™ reactor (ThalesNano) encompassing a heated stainless-steel reaction coil (ID: 0.8 mm, length: 11.0 m, internal volume: 5.5 mL). The system was equipped with an adjustable back pressure regulator (BPR), which was employed for most of the reactions at 75 bar to ensure safe operation within very high temperature ranges. The starting material was fed directly using an HPLC pump. Toluene was used as carrier or as washing solvent between flow experiments. (See Fig. S2 in the Supporting Information for a photograph of the flow system and also Fig. S3 for determination of steady state conditions.)

On the basis of MW batch data, reaction temperatures of 240, 280 and 300 °C were initially explored. At 240 °C, low conversion (up to 74%) and poor selectivity (up to 33%) was achieved either in toluene solutions (0.1 or 0.5 M) or under neat conditions applying different residence times (see Table S1 in the Supporting Information). Gratifyingly, at temperatures of 280 and 300 °C, the Claisen rearrangement proved much more efficient. As can be seen in Fig. 2, residence time had a very pronounced effect on the formation of the Claisen product even at such high temperatures. For example, while maintaining a substrate concentration of 0.1 M and a reaction temperature of 300 °C, the conversion rose from 50 to 96% and the selectivity from 15 to 89% with residence times increasing from 0.7 to 5.5 min. Despite the unimolecular nature of the reaction, concentration effects were also seen with higher concentrations typically accounting for an increase in reaction rate as well as in chemical selectivity. For example, while maintaining a temperature of 300 °C and 2.8 min residence time, an increase of substrate concentration from 0.1 to 0.5 M resulted in an improvement of conversion from 88% to > 99% and an increase of selectivity from 50 to 79%. To our delight, the continuous flow reaction proceeded well under neat conditions and exhibited a better performance when compared to diluted conditions. At 280 °C and 5.5 min residence time and at 300 °C and 2.8 min residence time, full conversion and 96% selectivity were achieved. Applying a prolonged residence time of 5.5 min at 300 °C resulted in precipitate formation in the reactor channels followed by a pressure increase due to possible overreaction, indicating a potential limitation of the solvent-free flow conditions. With respect to the nature of side reactions during process development, in addition to overreaction at higher temperatures and/or residence times resulting in isomerization and oligomerization, milder reaction conditions may lead to incomplete rearrangement and the formation of an intermediate with only one of the aromatic rings transformed. It is noteworthy that the performance of the flow reaction showed no detectable pressure dependence within the range of 10–100 bar (see Table S2 in the Supporting Information for details).

Fig. 2
figure 2

Investigating the effects of various substrate concentrations and residence times at 280 and 300 °C under flow conditions. (Blue bars: conversion and selectivity at 280 °C, orange bars: conversion and selectivity at 300 °C; conversion and selectivity were determined by HPLC area% at 270 nm.)

Despite the promising results attained at 300 °C, further parameter optimization was performed within the temperature range of 260–285 °C considering that in case of a manufacturing-scale process, consumption of energy should be minimized, whilst safety (heat release) should be maximized (Table 2, entries 1–10). Gratifyingly, with extension of residence times to 10 min, full conversion and a chemical selectivity of 94% was reached at 260 °C (entry 2). Similarly, temperatures of 280 and 285 °C in combination with residence times between 5 and 14 min ensured quantitative reactions with excellent selectivities of 94–97%, thereby providing a practically useful parameter window for the flow synthesis of 2 (entries 6–10). The viscosity of product samples prepared under different reaction conditions exhibit a significant dependence on the extent of oligomers formed as side products. For samples prepared under optimized flow conditions, viscosity values were found around 20 000 mPa s, which agrees well with the viscosity of commercially available samples of 2.

Table 2 Fine-tuning of the continuous flow reaction conditions.

In order to examine the practical limitations of the flow system from a more academic viewpoint, the solvent-free Claisen rearrangement was also attempted in the temperature range of 320–450 °C (Table 2, entries 11–16). With residence times of 20 or 40 s, the reaction furnished excellent selectivities of 92–95% (entries 11–14) in the temperature range of 320–360 °C. However, despite the short residence times, further increase of reaction temperature to 400 or 450 °C, resulted in a decrease in selectivity to 83% and 66%, respectively (entries 15 and 16). The tendency for side product formation as a function of temperature and heating time agrees well with the results of differential scanning calorimetry (DSC) measurements of the starting material and the product of the Claisen rearrangement (Fig. S4 in the Supporting Information). The DSC recording of 1 revealed two major peaks, which are likely correlated to Claisen product formation and overreaction (possibly isomerization and oligomerization), respectively. The DSC graph of 2 indicated one signal corresponding to thermal degradation in case of longer heating times, similar to what was observed in the MW batch experiment shown in Table 1 (see, for example, entries 5–7). In contrast, due to the effective heat transfer and well-defined short residence times, the flow process ensured high chemical selectivites even at higher temperatures.

To probe the stability and lab-scale preparative capabilities of the flow process, long-run experiments up to 5 h of continuous processing were implemented employing various sets of conditions shown in Table 2. To our delight, the flow system proved robust during all experiments attempted, no clogging or any other issues were observed. The processes furnished quantitative and highly selective Claisen rearrangements thus sufficiently pure batches of 2 were obtained without any work-up or purification required. Productivities were in the range of 25–45 g/h and space–time-yields (STYs) were obtained up to 8000 g/(L h) [31].

Conclusions

A continuous flow process was developed for the lab-scale synthesis of the valuable raw material bis-allylphenol 2. The process relied on the aromatic Claisen rearrangement of the corresponding diallyl ether precursor 1. Different temperature regimes and reaction times were initially screened under batch MW conditions. The findings of the preliminary batch study provided the basis for a thorough continuous flow process development, where reaction temperature and residence time were fine-tuned carefully in order to achieve a high yielding and selective formation of the Claisen product. In order to maximize productivity and to minimize environmental impact, concentration effects were also studied in detail. Gradual increase of the substrate concentration enabled an increased performance in reaction rate and in selectivity, and the best results were finally attained utilizing solvent-free conditions. The preparative capabilities and robustness of the process were explored during multiple long-run experiments yielding multi grams per hour of 2 directly “from tap” without any work-up or purification needed. Importantly, due to the effective heat transfer, the flow process enabled an improved safety profile, whilst ensuring a significant chemical intensification over traditional batch procedures due to the well-defined and short residence time. We therefore believe that the lab-scale protocol described herein, provides the foundation to develop scale-up options capable for manufacturing of 2 under flow conditions.