Manufacturing chemicals with light: any role in the circular economy?

Abstract

We outline how recent developments in photochemistry can contribute to the realization of the 1912 vision of the pioneering Italian scientist Giacomo Ciamician, namely world-wide chemical-using industry-based chemical plants fuelled solely by the Sun. We then show how a combination of organic photochemistry and flow chemistry could contribute to the circular economy by harnessing the ability of light to provide the energy to promote reactions without the need for some of the added reagents that are necessary in more traditional chemical routes, so-called 'reagentless' chemistry. Photochemistry has a long history but recently it has undergone a renaissance, particularly with the rise in interest in photoredox chemistry. Continuous photoreactors offer a route to scaling up such reactions to a productivity needed for smaller scale pharmaceutical manufacture. We describe some reactor designs from our own laboratory and outline some of their applications. We then relate these to the requirements of the circular economy and the need to conserve the stocks of the less abundant chemical elements.

This article is part of a discussion meeting issue ‘Science to enable the circular economy'.

Ever since the kick-start of green chemistry by the publication of the 12 Principles of Green Chemistry [1], there has been the realization that green chemistry needs a further paradigm shift to ensure that we can reduce or eliminate negative environmental impacts involving reduced waste products, non-toxic components and improved efficiencies. Indeed, in their comment on how green chemistry is currently practised, the Green Chemistry & Commerce Council (GC3) stated ‘Green chemistry incorporates every element of business, from product design to feedstock selection through manufacturing to finished products, including the ways that companies manage their businesses and engage their customers throughout the supply chain. While green chemistry is practiced primarily at the chemical discovery, development, and formulation levels, product developers, manufacturers, brands, and retailers all play an important role in its implementation. Several ways they do this are by changing design specifications, sourcing materials and products that incorporate green chemistry practices, changing manufacturing practices to substitute or reduce the use of hazardous chemicals, and developing and implementing policies that restrict chemicals of concern in the products they source, make, and/or sell. Green chemistry can be an iterative process where products are improved incrementally over time, or it can yield a disruptive innovation and offer entirely new technologies and approaches for making safer products' [2]. The final sentence in this quote perhaps describes the operation of two separate paradigms, namely, ‘an iterative process where products are improved incrementally over time' and ‘a disruptive innovation and offer entirely new technologies and approaches for making safer products'. However, both of these apparently contradictory approaches are crucial; both of them are equally important in moving society forward to implement the goals of green chemistry, because implementing green chemistry is a complex multi-dimensional problem which does not have a single unique solution. Success requires a combination of incremental change and radical paradigm shifts to address the huge variety of challenges which must be overcome before all chemical production becomes truly sustainable.

In this paper, we discuss two concepts, manufacturing with light and Moore's Law for Chemistry [3], both of which can operate under either paradigm but together offer a chance for transformative change both in green chemistry and the circular economy. Although the paradigm(s) of green chemistry in its modern form have been framed since the early 1990s, the concept of revolutionary changes for sustainable chemistry has a much longer history. For example, more than 100 years ago, the Italian chemist Giacomo Ciamician contrasted the harsh conditions of chemical synthesis in the laboratory with the mild syntheses carried out by green plants [4]. His ideas of harnessing sunlight and photochemistry for chemical reactions and chemical manufacturing can be framed in terms of (i) carrying out reactions under mild conditions; (ii) looking for activation conditions that allow for more direct reactions and thus improve atom economy; (iii) increasing the use of renewable feedstocks and/or reagents and (iv) minimizing the use of fossil energy. Even though this vision has been in place since 1912, we have yet to realize his vision of widespread implementation of photochemistry in industry to satisfy a significant fraction of the world's chemical needs.

There are several reasons for the slowness in implementation of Ciamician's vision. These include World War 1, which started only two years after his publication. In addition to this is the fact that the chemical industry used to source feedstocks largely from coal tar [5] as well as the rapid rise of petroleum-based feedstocks and the growth of the petrochemical industry, and more prosaically the engineering difficulties of implementing any sort of photochemistry at scale. This close alliance of chemistry and engineering has been recognized in recent years, particularly with the development of 12 Principles of Green Engineering to complement those of green chemistry [6]. The synergy between green chemistry and engineering has been further reinforced by the ACS Green Chemistry Institute which has published a pamphlet on ‘Design Principles for Sustainable & Green Chemistry & Engineering’ [7], particularly emphasizing the need for a more holistic systems approach to the whole problem of supplying chemicals in the future. Very recently, Anastas and Zimmermann have taken a bolder approach by proposing a comprehensive Periodic Table of Green & Sustainable Chemistry and Engineering [8]. At the same time, there has been a strengthening of the concept of stewardship of the elements. Unlike oil (which is burned), chemical elements cannot run out because they are not destroyed by chemical processing or usage. However, economically viable deposits of those elements can be exhausted and the elements themselves can be so widely dispersed across the environment that they are effectively lost to us. The European Chemical Society has highlighted the problem with a very striking periodic table, figure 1, which has aroused worldwide interest in the course of the 2019 International Year of the Periodic Table [9].

Figure 1.

The EuChemS periodic table of endangered elements, designed to mark the 2019 International Year of the Periodic Table. The area of each element gives a rough indication of its abundance on earth on a logarithmic scale; that means that the areas of the less common elements (e.g. Fr or At) are much larger than one might expect. Reproduced from the EuChemS website with permission. (Online version in colour.)

This raises the question of what is meant by sustainable chemistry. Although there is general agreement as to the definition of ‘sustainability' itself, surprisingly there is still discussion of what sustainable chemistry is and how it differs from green chemistry [10]. We are not going to add to the argument here apart from pointing out that sustainable chemistry clearly must respect economic issues and also potential societal impact. For example, green chemistry emphasizes the advantages of using renewable feedstocks based on biomass without explicitly considering the potential environmental, economic and societal issues associated with increased production and transport of the biomass needed to supply those feedstocks. Therefore, sustainable chemistry is clearly closer to the concept of the circular economy than is green chemistry itself.

Developments towards sustainable chemistry and particularly a circular economy need to take into account the effect on people, the planet and economics, i.e. the profitability. For sustainability to be effective for a circular economy, any interventions need to be economically viable. In other words, chemists need not only to think of economics as well as thermodynamics when implementing new sustainable processes, but also to be mindful of the societal impacts of their proposals. Many of these issues have now been incorporated into a proposed [11] set of 12 Principles of Circular Chemistry, in order to guide future thinking about sustainable chemical processes, figure 2.

Figure 2.

The recently proposed 12 Principles of Circular Chemistry [11], redrawn in the form of a clock-face, not least to symbolize that ‘the clock is ticking'.

Time will tell whether these Principles of Circular Chemistry are widely adopted but of particular relevance to this paper is the proposed Principle 10, CC10, ‘Sell service not product', which is essentially the underlying approach of the slightly misleadingly named UNIDO Chemical Leasing initiative [12]. Chemical Leasing promotes the idea that chemical producers should be selling a service or effect rather than a given amount of chemicals. Thus, an explosives company would be selling holes in the ground rather than a given weight of explosives. This then provides an incentive for chemical manufacturers to minimize the amount of chemical that is needed to achieve the desired effect because this will maximize their profit. Licence, Poliakoff and George recently built on this idea to propose a Moore's Law for Chemistry [3], by analogy with the well-known Moore's Law in electronics. Their proposal is that, over a given period (they suggested 5 years), the amount of chemicals needed to produce a given effect (cure a headache, dye a fabric, etc.) should be reduced by a factor of 2. Then, over a number of cycles, the amount of chemicals that are used will be substantially reduced with a correspondingly reduced burden on feedstocks and the environment. The potential advantage of this approach is that it proposes a goal which will benefit everyone, chemical producers, end users, environmentalists and eventually our planet.

Photochemistry has a definite role in the context of Moore's Law for Chemistry, as well as green chemistry and the circular economy, because photons can replace one or more reagents that would be necessary in a thermally promoted reaction. Furthermore, light can often promote reactions which are inaccessible via direct heating of a reaction mixture. Photochemistry also aligns well with at least two Principles of Circular Chemistry: CC4. Strive for energy persistence. Energy efficiency should be maximized; and CC5 - Enhance process efficiency. Innovations should continuously improve in- and post-process reuse and recycling, preferably on-site. So clearly Ciamician's vision is becoming highly topical in the context of the current sustainability agenda but what is making a real difference are technical advances in reactor engineering which remove many of the engineering barriers that have hamstrung larger scale photochemistry for so long. Indeed, the authors of the Principles of Circular Chemistry have themselves written, ‘There is a strong ambition towards developing chemically renewable energies, for example, solar-driven chemistry'.

Many of the problems of scaling up photochemical reactions are associated with the fact that a photochemical reaction requires absorption of light and this absorption limits the penetration distance of that light into the solution. Thus, there is a choice; either one must have a strongly stirred reactor so that all of the reaction mixture is swept past the light source or, alternatively, one can design a reactor in which the thickness of reaction mixture is small, for example, in a long narrow tube or within a falling film. Sunlight is highly directional and can be easily focussed onto a reaction vessel, although longer wavelength light needs to be filtered out to prevent overheating of the reaction mixture and, of course, the position of the sun changes throughout the day. Until recently, the majority of other light sources have been less directional with the result that it is often simplest to place the light source in the centre of a reactor which also means that the heat generated by the lamp must necessarily be removed from within the reactor, considerably complicating the engineering of the equipment. Nevertheless, many successful reactor designs have been proposed including falling film, [13,14], bubble column, [15] spinning disc, [16–18] slug flow, [19,20] high pressure, [21–27], vortex [28] FEP tubular, [29] parallel tubular, [30] rotating thin film, [31,32] annular thin film, [33] milling, [34] and one based on irradiating a nebulized liquid/gas mixture for singlet oxygen chemistry [35].

Perhaps the key question is why has photochemistry not been widely adopted/implemented in industrial processes? Surprisingly, this question was already being asked in a paper ‘Science of today—production tool of tomorrow Industrial Photochemistry', which was published in 1962 ca. 50 years after Ciamician propounded his original vision. In a section entitled ‘Why is photochemistry not used more widely?', the authors wrote [36], ‘A main problem in using photochemistry industrially is related directly to insufficient knowledge of the parameters involved and their effect on reactor design. Then added to this are other problems – inadequate flow of information between workers and a limited variety of equipment available. The solution to these two problems is not easy, but several steps in the right direction could be taken: (i) release of information on reactor design; (ii) develop and improve more adaptable manufacturing equipment and (iii) encourage chemical engineering to carry out graduate research (in the area)'. In recent years, there have been major advances which make the wider adoption of photochemistry by industry considerably more likely in the near future.

Three technical advances that are transforming the application of photochemistry should perhaps be highlighted in particular. These advances are (i) greatly improved light sources, particularly light emitting diodes (LEDs), which are not only more electrically efficient than many more conventional light sources but also are highly directional, allowing LEDs to be placed outside reactors without substantial loss of light; this, in turn, greatly simplifies heat management since much of the heat can be removed via the rear of the LEDs mounting and not through the reaction mixture; (ii) the rapid development of flow chemistry, where small reactors replace more traditional large batch reactors, thereby overcoming the limitations of poor light penetrations into large vessels, and allowing photoreactors to be incorporated into multi-step continuous manufacturing processes and (iii) innovative designs of continuous photoreactors which are permitting substantial process intensification with high productivity from compact equipment with a relatively small footprint.

The use of continuous flow chemistry has increased significantly in recent years and such approaches enable the use of small volumes of reagents and the safe handling of hazardous reagents eliminating the need to isolate toxic and/or hazardous intermediates. The large surface area to volume ratio, a result of the small dimensions, also provides efficient mass and heat transfer rates. Such advantages have meant that continuous reactors offer good pathways for larger scale processing either using a number of reactors in parallel (so-called ‘scaling out') or more simply by scaling up, namely increasing the size of the reactor. There are, however, limitations to the scale-up of photoreactors which arise from the increased pathlength of the reaction mixture as the size of the reactor increases, which means that the light fails to penetrate into the depths of the reactor. The result is that, depending on the design, the overall efficiency of the reactor may fall as it is scaled up.

The overall level of interest in this topic of manufacturing with light has been reflected by the large number of reviews [37–56] published recently in this area. There are many views on how to achieve widespread implementation of this approach and we are optimistic and place perhaps rather more emphasis on what remains to be done on the engineering side. The vision for multi-step chemistry is shown schematically in figure 3, namely, that process chemists will have a pool of flow reactors for photo-, electro- and thermo-chemistry which can be linked together like wagons on a train to deliver a multi-step continuous process. Once that process is finished, the reactors can be returned to the pool and another train can be assembled to deliver the next multi-step process that is required.

Figure 3.

A schematic representation of our vision for using different combinations of flow reactors from a central ‘reactor pool' to assemble trains of reactors to deliver particular continuous multi-step processes for manufacturing complex organic molecules. (Online version in colour.)

As described above, there have been many developments in reactor design and process development to try and facilitate more widespread use of photochemistry in academia and industry. Here, we describe recent developments of three reactors for flow photochemistry in our laboratory to highlight how novel reactor designs can open up new process possibilities namely, the Photovap, the Vortex and the UV excimers reactors and illustrate a photochemical daisy chain with the excimer reactor to give an integrated process which also exploits a more sustainable solid acid catalyst.

The PhotoVap [31–32] is a rotary evaporator (colloquially known as a RotorVap) adapted to act as a semi-continuous photoreactor, figure 4. The underlying principle is that the rotation of the flask creates a thin film of liquid on the inner surface of the flask, which permits good penetration by the light from a bank of high-powered LEDs. The reactor operates under computer control. An aliquot of reaction mixture is pumped into the flask which then begins to rotate. At the same time the light is switched on for the required reaction time. The rotation is stopped, the solution which now contains product is pumped out, the reaction flask is refilled and the process is repeated. A key feature is that, unlike more conventional flow reactors, the residence time and hydrodynamics can be adjusted totally independently because the hydrodynamics depend only on the rotation speed of the reaction flask. Although the PhotoVap is a semi-batch reactor, the reaction mixture is in motion during irradiation. Therefore, unlike a tubular flow reactor where increased irradiation time is achieved by slowing the flow, the motion of the fluid in the PhotoVap and, hence, the mixing, is independent of the irradiation time of the reaction mixture. We have tried using the PhotoVap as a continuously stirred tank reactor by pumping solution in and out while the LEDs are on and the reactor is turning, but we have found that higher yields and greater productivity are achieved by operating it in a semi-batch mode (i.e. using it repeatedly as a batch reactor with short irradiation times).

Figure 4.

Schematic of the ‘PhotoVap' reactor, based on a standard rotary evaporator. The components are labelled as follows: F reaction flask, the rotation of which creates a thin film of the reaction mixture on the inside surface of the flask; LEDs, light emitting diodes for promoting the reactions; R and P flasks containing the Reactant and Product solutions respectively. For further details, see [31,32]. (Online version in colour.)

The PhotoVap has been used extensively for reactions with singlet oxygen where its geometry gives excellent mass transfer of O₂ between the gas and liquid phases. It has also been used successfully for UV photochemistry, replacing the LEDs with high power high-pressure Hg lamps and using either borosilicate or quartz reactions flasks. For the reaction of Cookson's dione, productivity of greater than 200 gm h⁻¹ have been achieved, scheme 1. Even on the laboratory scale, the PhotoVap is scalable; we have used it with flasks ranging from 50 ml to 3000 ml volume depending on the cost of the starting material. The PhotoVap is inherently a small-scale reactor which perhaps is more suited to pharmaceutical rather than fine chemical production. One can envisage scale-up perhaps to 20 or even 50 L flasks in an academic environment. As the size of the flask increases its area : volume ratio will decrease. However, since we only use part of the surface for irradiation in the 3 L reactor, this reduced ratio could be compensated, at least in part, by using a larger proportion of the surface for irradiation. This problem is largely avoided in the Vortex reactor described below. In principle, commercial rotary evaporators are available up to 200 L volume so that further scale-up would not be difficult. The big attraction of the PhotoVap is that every organic chemistry laboratory owns a rotary evaporator which could easily be converted into a PhotoVap.

Scheme 1.

UV synthesis of Cookson's Dione, which has been carried out on a scale greater than 200 gm h⁻¹ in the UV version of the PhotoVap [32].

Chemical manufacture is unusual in the context of the circular economy because the products themselves are often consumed in use—for example pharmaceutical products, insecticides or even paints. Therefore, applying the principles of the circular economy also needs to focus on innovations that make the manufacturing process more efficient and minimize the use of elements that are or will be in short supply. As explained above, photochemistry has the advantage of reducing the need for additional reagents but innovative approaches are required to minimize the energy usage of photochemical processes and also to maximize the efficiency of the reactors, particularly when they are to be incorporated as one step in a train of consecutive continuous reactors for multi-step processes.

Therefore, there is a need to develop new reactor types that are versatile and agile in order to assemble trains of reactors to deliver particular continuous multi-step processes for manufacturing complex organic molecules combining processes that are initiated using either thermal, photochemical or electrochemical reactions in flow. As mentioned above, one of the properties of the PhotoVap is its ability to decouple flow-rate from residence time and this can be key in optimizing multi-step processes which inherently have different rates and optimal conditions. We describe one such approach, the Vortex Reactor, which also involves rotation in its operation, but the effect of that rotation is quite different from the PhotoVap, as explained in figure 5. The hydrodynamics which, in the Vortex Reactor, are controlled by the rotation speed of the central cylinder, are uncoupled from the residence time which is determined by the rate that the reaction mixture is pumped into the reactor. The design and optimization of the Vortex Reactor has involved a fruitful collaboration between computational modellers and experimental photochemists. This modelling strongly suggests that the Vortex Reactor should be scalable to a size large enough for commercial manufacture of pharmaceutical products. A great advantage of the Vortex Reactor is that, to a reasonable approximation, the ratio of (surface area) : (volume) of the reactor remains constant as the reactor is scaled up. Therefore, the productivity per unit volume per unit time should remain almost constant during scale-up.

Figure 5.

The Vortex Reactor. In this manifestation, the reactor consists of a smooth stainless steel cylinder inside a transparent jacketed reactor. When the cylinder is rotated rapidly, it creates toroidal Taylor vortices which ensure that all of the solution is moved close to surface of the reactor where it can be irradiated by light (see right-hand part of the figure). The reaction mixture is fed down the centre of the cylinder and exits from a pipe at the top. A particular advantage when working with singlet oxygen is that the gas phase is broken up into very small bubbles with a large total surface area which ensures good mass transport but a very small total volume for added safety. An extra benefit in the small-scale version of the reactor is that the rapid rotation of the cylinder draws in air from the laboratory so that a separate source of oxygen is not required [28]. On scale-up, however, a separate gas feed is required. (Online version in colour.)

Both the PhotoVap and Vortex reactors overcome a key problem that besets multi-step processes involving a sequence of continuous reactors, namely how to decouple flow-rate from residence time. Solving this problem can be key to optimizing and linking reactors together in multi-step processes, each step of which inherently has different rates and optimal conditions.

Our third reactor is targeted at addressing another major problem in optimizing photochemical processes, namely that broadband irradiation is not ideal because secondary photolysis can destroy the desired photoproduct or reduce selectivity. We have designed a reactor that provides a flexible platform on a smaller scale to establish the wavelength dependence of reactions and to identify whether secondary photolysis of photoproducts is an issue in a given reaction. In this design, we have exploited the use of an unusual annular design of exciplex/excimer UV lamp generates, nearly monochromatic radiation at a UV wavelength which depends on the particular gas mixture inside the lamp (e.g. KrCl 222 nm, XeBr 282 nm, XeCl 308 nm, etc.) (figure 6).

Figure 6.

Schematic diagram of a flow reactor for use in an annular exciplex/excimer lamp. The reactor itself consists of a stainless steel cold-finger with a co-axial feed tube running down its centre. The photoreactor itself consists of a quartz outer tube which allows the reaction mixture to flow through the annular volume between outer surface of the cold-finger and the inner surface of the quartz. The cooling is necessary because of the high temperature inside the bore of the lamp. The parts are labelled as follows: (I) 1/16′′ stainless steel tubing; (II) and (III) Swagelok™ T-pieces; (IV) Bola GL 32 PTFE T-Piece; (V) 1/4′′ stainless steel tubing; (VI) 12 mm stainless steel tubing, welded closed at the bottom; (VII) Quartz outer jacket; P: Org1) - Jasco™ PU-986 HPLC Pump; Cooling Bath - recirculating batch model Julabo^® F32-HE, T1) - manual valve. For further information see [33]. (Online version in colour.)

Each of these three reactors has its strengths and weaknesses which illustrate the general point that no one reactor will be able to satisfy the requirements of every reaction. Hence, there is a need for a selection of reactors as illustrated in figure 3 so that there will always be at least one reactor suitable for every application. In this way, one will be able to apply photochemistry to maximum effect in supporting the circular economy. Ultimately, reactions should be driven solely by sunlight either directly via light or indirectly via renewable energy and there is likely to be continuing improvements across the world in reactor design and process development towards this goal. Below, we give a relatively simple example of using the excimer reactor together with an abundant element, niobium, to give a greener process [33].

The excimer reactor was used as the first reactor in a train to carry out the multi-step process shown in scheme 2. In this process, the first step is the photodecaboxylation of a potassium salt, followed by cyclization. The product stream emerging from the reactor is highly alkaline because of the potassium ions liberated by the decarboxylation. These ions have to be neutralized before the acid catalysed dehydration can take place, which requires strongly acidic conditions. This neutralization was achieved using Amberlyst 15, an inexpensive acid resin, which is not a sufficiently strong acid to promote the dehydration. Clearly the Amberlyst 15 is consumed when used in this way but, since it is an ion exchange resin, it is easily regenerated. The dehydration step was then catalysed using NbOPO₄, a much stronger solid acid, which in the absence of the K⁺ ions does indeed function as a true catalyst. Although niobium has limited abundance, figure 1, it is considerably more abundant than many metals used in catalysis (e.g. the platinum group metals) and furthermore, in an application like this, the niobium could easily be recovered and recycled.

Scheme 2.

A simple multi-step process combining photochemistry and acid catalysis [33].

Photochemistry is undergoing a renaissance. Photochemistry can be considered 'reagentless' as the effect of light results in the same outcome as is achieved using chemical steps with solvents and reagents. For example, singlet oxygen is often generated from hydrogen peroxide and hypochlorite but, photochemically, it can be generated from O₂ and light via a reusable photosensitizer. The ‘reagentless’ nature of photochemical reactions offers a real opportunity for photochemistry to contribute to making chemical manufacture more sustainable and, hence, to contribute to the creation of the circular economy. As explained in this paper, bringing photochemistry into chemical manufacturing is being helped by recent developments in reactor design and higher efficiency light sources. It is also being greatly assisted by the changing nature of pharmaceutical manufacturing in the post-genome era. Increased understanding of the human response to medicines is leading to larger numbers of compounds being used to treat particular medical conditions. This in turn means the manufacture of more products, and new technologies including photochemistry will be needed to meet this increased variety of medicines. Furthermore, the role of photochemistry is also likely to increase because the low tonnage production of many of these new medicines will not require substantial scale-up from some of the latest reactor designs. Of course, the long-term goal of many photochemists, including ourselves, is the realization of Ciamician's 1912 vision, namely world-wide chemical-using industry-based chemical plants fuelled solely by the Sun.

Data accessibility

This article has no additional data.

Authors' contributions

M.W.G. and M.P. both contributed to the content and writing of the paper.

Competing interests

We declare we have no competing interests.

Funding

This study was supported by Engineering and Physical Sciences Research Council (grant no. EP/P013341).