Oral administration of APIs necessitates the stability of the API and its respective acidic and basic components in the digestive tract and effective permeation of cell membranes (44). Ion-exchange systems have been investigated for their ability to act as API delivery systems, wherein once the API reaches the gastrointestinal tract the body’s salts displace the API allowing it to pass through the cell membrane in a controlled manner (20). However, human physiology makes API delivery via gastrointestinal (GI) tract challenging (45). The short GI transit time (ca. 12 h) makes the delivery of macromolecular therapeutics such as proteins and nucleic acids difficult (45). The limitations of API delivery in the GI tract (44) have helped to shape the development of polymer-based API delivery systems to deliver macromolecules such as insulin orally or via inhalation (46, 47).

An ideal API delivery system would allow a patient to monitor and administer drugs (such as insulin) on demand with control over the dose and no need for invasive injections. Variations of these are currently being developed for the self-regulated treatment of diabetes (48).

Transdermal patches were amongst the first systems to be available to patients with APIs being attached to an adhesive patch before delivering a specific dose through the patient’s skin and into the bloodstream (49). Transdermal patches enable controlled release via a porous membrane slowly releasing an API from a reservoir within the patch. The first transdermal patch was FDA approved in 1979 for the delivery of prescription API scopolamine (Figure 2) for the treatment of motion sickness (50). Nowadays, many APIs are administered via transdermal patches (for example, Daytrana^®, EMSAM^®, Exelon^® and fentanyl, Figure 2) covering a wide range of medical conditions from Alzheimer’s to attention deficit hyperactivity disorder (ADHD) (51).

Whilst API delivery from transdermal patches is effective, the skin is a barrier to entry from external bodies which results in a high proportion of the API being prevented from entering the body and a reduced therapeutic efficiency (49). One solution to this problem is the utilisation of chemical enhancers (49) to alter the permeability of an API, for example, the skin permeability of oestradiol (Figure 2) can be increased twenty-fold via formulation with ethanol (52). A common side effect of the use of chemical enhancers is skin irritation at the site of the patch which may make the use of the enhancer non-viable. Another method is to chemically modify the structure of the API to improve its permeability; however, this can be difficult, expensive and time consuming (53). The use of arrays of microneedles for transdermal delivery is increasingly popular because of their broad applicability and minimal pain (54).

The use of microneedles in drug delivery began in the 1990s as a result of the emergence of microfabrication techniques that enable their manufacture (55). Microneedles are used in a variety of medical systems including skin pre-treatment for increased permeability, drug coated needles and drug encapsulated needles (55). Microneedles are now widespread in drug delivery having shown the ability to give controlled release of a wide range of low molecular weight drugs and vaccines (55). The delivery of the influenza vaccine using a microneedle is common in modern medicine (56). Microneedle delivery depends on a variety of factors including skin permeation, drug stability, drug storage and patient response (55). This emerging field of medicinal chemistry shows great promise in forwarding the field of drug delivery.

Injectable and implantable API delivery systems are particularly useful for conditions requiring the delivery of APIs to specific sites within the organism. Many APIs suffer from an inability to reach the required site of action due to a biological barrier (for example, the blood-brain barrier). Parkinson’s disease caused by dopamine deficiency cannot be treated by administration of dopamine because it does not cross the blood-brain barrier, however, the prodrug levodopa (Figure 2) is capable of crossing the blood-brain barrier after which it is metabolised to dopamine (Figure 2) (57).

Likewise, <2% of the administered dosage of naltrexone (Figure 2), an API used in the treatment of opioid dependence reaches the brain and naltrexone-polymer conjugates can increase the amount of API working at the site of action resulting in US Food and Drug Administration (FDA) approval for use for the treatment of alcohol dependence (2006) and opioid dependence (2010) (58).

Implanting API delivery systems at or near the desired site helps to maximise local delivery and minimise undesirable side effects. A polymer-based API delivery system known as Ocusert which controls the release of pilocarpine (Figure 2) and reduces pressure in the eyes (59); implantation of pilocarpine encapsulated between two polymer membranes controlled the release at a rate of 20 mg h^–1 for up to a week (59). Several polymeric versions of the Ocusert delivery system exist, all capable of delivering pilocarpine in a controlled manner with differing release profiles. Early uses of this system were limited by poor biodegradability, however, new formulations of biodegradable polymers have helped to improve degradation profiles (60).

Biodegradable polymers (such as poly(anhydrides) and polyesters) used for polymer-based API delivery systems can slowly degrade and release APIs (for example, carmustine (Figure 2) a chemotherapeutic treatment for brain cancer) and carmustine-loaded polyanhydride films directly at the tumour site were shown to significantly improve patient survival rates when treating glioblastoma multiforme (61).

PGLA has also been used in the controlled delivery of the API apomorphine (Figure 2) which is used in the treatment of Parkinson’s disease (62). Apomorphine has poor oral availability and a short half-life, resulting in multiple administrations being required which limits its widespread usage, therefore controlled release methods are used to overcome this shortcoming (62). The use of PGLA prevents the burst release of apomorphine and increases longevity of the API within the target tissues (62). This system demonstrated controlled release of the API over ten days, releasing 90% of the payload.

The investigation of smart devices in medicine has probed the use of API delivery systems that can control API release using an external stimulus or by interactions between the API delivery systems and changes in their environment. By implanting a biocompatible device within the patient and then triggering API release externally, the patient would be provided with the therapeutic benefit over an extended period of time. An ideal API delivery system would allow control of the dosage, timing, duration and site of API release, resulting in delivery of the therapeutic agent in a remote and non-invasive manner. A range of stimuli can be used to trigger API release including pH, infrared (IR) (63), ultraviolet (UV)-visible light (64, 65) magnetism (66), temperature (67), ultrasound (68), electric fields (69) and radiation (70). Many of these stimuli are already utilised in clinically translated API delivery systems (Table II). The development of API delivery systems that respond to these stimuli and provide the controlled release of loaded APIs potentially improves treatment efficiency and diminishes or prevents the onset of side effects. There are API delivery systems that respond to multiple stimuli to further improve selectivity for specific functions (71), see below for a fuller discussion.

Another emerging aspect of formulation science involves the use of shape memory materials (SMMs). SMMs demonstrate plastic deformation when stimulated by an external stimulus and return to their original shape upon removal of the stimuli (72). Shape memory polymers (SMPs) are stimuli responsive compounds which are able to demonstrate mechanical action in response to a range of stimuli depending on the material make up. SMPs offer a range of advantages including; wide glass transition states, tailored stiffness, high shape recovery, high elastic deformation, biodegradability, biocompatibility and low thermal conductivity (72). The ability of these materials to assume a specific shape upon triggering can be utilised for drug delivery. PCL and poly(lactide) (Figure 2) are often utilised in medical SMPs as they have distinctive glass transition states and are inherently biodegradable and biocompatible (73). The use of these polymers in SMPs can assist in drug delivery via two mechanisms: either the shape recovery of the polymer enhances drug release or the polymer facilitates delivery of the drug delivery device to the body in a minimally invasive manner (73). The incorporation of a drug into a SMP delivery system has been demonstrated to affect performance of the DDS however controlled release is still possible. The use of SMPs in urethral stents has been demonstrated using the SMP as a method of controlled release of anti-inflammatory drugs (74). This method demonstrated the ability of SMPs to show controlled release of a drug and upon completion degradation into non-toxic products (74). This example highlights the potential use of SMPs in drug delivery and wider medicinal applications (75).

Light triggered API delivery systems are very popular in the literature due to their ability to provide temporal and spatial control, functioning via various mechanisms including photochemical, photoisomerisation and photothermal (76). Photodynamic therapy (PDT) is one of the most well-established techniques and uses light in the UV-visible spectrum to treat skin and throat cancers (77). PDT is less effective when attempting to affect deeper set tumours such as prostate and liver cancers for which light in the IR spectrum is preferable as a result of its relatively low absorption by mammalian tissues (63).

Photochemical API delivery systems release a therapeutic payload upon covalent bond cleavage in response to light irradiation (76). An example of such chemistry is the cleavage of an o-nitrobenzyl ester derivative releasing a carboxylic acid (Figure 3). The carboxylic acid-displaying molecule was released over several hours at surface power of 1.3 mW cm^–2, however when increasing the power to 20 mW cm^–2 release was only observed over 5 min (78). This system demonstrates a high degree of control that shows promise in being utilised in API delivery studies.

Fig. 3.

A library of photo-responsive units have been explored for API delivery including coumarin, pyridylmethyl esters and porphyrins, all of which contain readily cleavable covalent bonds (79). Photo-responsive API delivery systems function on the requirement of light with a wavelength that possess sufficient energy per photon to affect the breakage of covalent bonds (80), making UV (81) and visible light (79) popular triggers. One of the most prevalent problems with light triggered API delivery systems is the relatively poor tissue penetration of UV and visible light, this has been addressed by the development of near-infrared (NIR) API delivery systems (82). NIR is only fractionally adsorbed by biological tissues thus allowing it to trigger API release in deeper areas of the body (82). Almutairi et al. report the use of a UV responsive nanoparticle DDS in which nintedanib (Figure 2), a drug used in the treatment of idiopathic pulmonary fibrosis, is released over ten weeks (83). The nanoparticles were shown to be biocompatible with no adverse effects observed despite the extended period of implantation (83).

Photo-responsive hydrogel-based API delivery systems (84) offer the opportunity to deliver sensitive bioactive macromolecules (84) and minimise the body’s immune response. A recent trend in the literature points towards the development of systems that do not require the use of UV as a result of the risk it poses to the skin and eyes. The use of NIR and visible light triggered systems are increasingly popular in photochemical API delivery due to reduced risk associated with these triggers (85).

Whilst a great deal of progress has been made in the field of photochemical API delivery many problems still persist and must be overcome before these systems are fully utilised in modern medicine. Early attempts at photochemical triggering often resulted in one effective dosage of the API before the system is empty, however new innovative systems have demonstrated pulsatile delivery with few adverse effects. Tissue penetration is still a problem in this field with visible light-based systems limited to the skin, throat and nose (63). As with all new systems being introduced to the body, biocompatibility is a huge stumbling block. Even the most biocompatible systems generate some form of immune response, sometimes in the form of inflammation but others can be more serious and so systems must be vetted fully before use. Despite these problems, photochemical API delivery remains a very popular research area with huge progress being made throughout this field.

Early attempts to develop stimuli responsive systems included the development of conducting polymers which were theorised to be able to release an API upon triggering with an electrical stimulus. Polypyrrole (PPy) in its conducting (oxidised) form allows oppositely charged ions to be doped into the polymer backbone which was pioneered by the Miller Group in 1984, who demonstrated their ability to release glutamate ions (Figure 2) via the reduction of PPy (Figure 4) films (86). The cationic PPy is doped with anionic or neutral API molecules. When an electric current is applied to the system the polymer changes redox state and the API is released in order to charge balance the system (87).

Fig. 4.

The sensitivity of electroactive species can be manipulated to create a range of API release profiles through redox switching. Despite the widespread usage of PPy as an API delivery agent it is difficult to process due to its poor solubility in most solvents. Many attempts have been made to improve the solubility of PPy with limited success (88). PPy is also non-biodegradable and therefore can be difficult to remove from the patient’s system once all the loaded API has been used (89). The success of utilising PPy films as API delivery agents prompted an investigation into other polymers such as polyaniline (PANi, Figure 2) (90) and poly(3,4-ethylenedioxythiophene) (PEDOT, Figure 2) (91) with varying degrees of success. The biocompatibility of the polymers and the amount and molecular weight of API that can be loaded onto these films are areas of current research (92, 93), as is the generation of biodegradable versions (94, 95).

Whilst single stimuli responsive systems are very useful, they are restricted to certain release profiles based on the stimuli in question. The complex nature of the human body and the conditions which affect it often require additional more complex solutions than single stimuli-responsive DDS. Multi-stimuli responsive DDS are being explored for their ability to create more varied release profiles, providing an improvement in tuneability and selectivity versus single responsive systems (96). In theory multi-responsive DDS allow for the treatment of a wider range of complex conditions by regulating release by one or more stimuli based on patient needs (96).

When constructing multi-responsive DDS separate units, each of which is responsive to a specific stimulus, are blended together without affecting each unit’s responsiveness. Several systems are currently in development based on the ability of one stimulus to act as a targeting moiety whilst the other stimuli are responsible for producing a response in the desired tissue.

pH is one of the most commonly used stimuli in dual responsive DDS. The ability of these systems to be selective based on the targeting of the lower pH of cancer cells makes them desirable in modern cancer treatments (97–100). pH is often combined with a variety of other stimuli including light, electricity and magnetism to create a desired response in cancerous tissues. pH and light responsive materials are popular dual responsive DDS. Nie et al. have demonstrated the ability of these systems to show controlled release of the chemotherapy agent doxorubicin hydrochloride (Figure 2) via photothermal drug release (101). The use of a pH responsive group ensured selectivity towards cancer cells over healthy cells with an NIR responsive group providing photothermal release of doxorubicin hydrochloride in a controlled manner (101).

Dual responsive DDS which incorporate multiple stimuli capable of creating the desired drug release response are less common, however several examples exist in the literature. Argouz et al. have developed such a system with the use of sodium alginate gel beads in a pH or magnetic drug release system (102). In this system pH sensitive sodium alginate is combined with methyl cellulose which has shown to be responsive to magnetic fields. Sodium alginate is a biodegradable, biocompatible, non-toxic polysaccharide and can be readily modified making it a useful tool in drug delivery (103). It has been combined with chitosan, pectin and gelatin for use in drug delivery with all systems displaying a high degree of biocompatibility (103). The resulting material has demonstrated the ability to show controlled release of the anticancer drug 5-fluorouracil (Figure 2) over extended periods of time (102). This system is comprised of both a targeting stimulus and two active delivery stimuli providing a high degree of impact when attempting to affect cancer tissues.

Kyriakides et al. took a different approach to multi-responsive DDS being able to generate constructs via simultaneous electrospinning and electrospraying, generating compartmentalised storage of multiple drugs (104). The use of this method provides a PCL fibre structure with a hyaluronic acid core, allowing drugs to be loaded in the polymer film (104). Further studies have shown the ability to trap other spheres of drug within an electrospun mat, allowed for delivery of multiple drugs with differing solubilities demonstrating various release profiles (104). A minimal immune response was found when using pirfenidone (Figure 2), an anti-fibrotic drug, in one of the release compartments (104).

Multi-responsive systems are becoming more prevalent in the literature with many systems demonstrating effectiveness in drug delivery, particularly when attempting to affect cancerous tissues. This field will continue to grow as scientists find more ways to incorporate more stimuli into existing systems, providing ample opportunity to treat a variety of conditions and improve patient care. Some examples for APIs displayed in Figure 5 and Figure 6 are highlighted in Table III.

Fig. 5.

Fig. 6.

Significant progress has been made in the field of API delivery over the past sixty years and the scope of controlled API delivery systems has greatly increased. Many challenges still remain in this field, such as delivering APIs to specific cells, targeting genes and designing systems to cross complex barriers such as the blood-brain barrier (42). New materials are being developed aimed at improving biocompatibility, generating new release profiles and improving patient care (142). Continued investment and effort in this field will lead to the development of API delivery systems capable of the delivery of APIs to specific tissues to the benefit of patients and the healthcare industry.

Advancements in the field of API delivery and controlled release have had a direct impact on other fields of chemistry such as synthetic and polymer chemistry, chemical engineering, materials science, chemical biology and bioengineering (33). Many API delivery systems exist generating a variety of release profiles and targeting different conditions. Conditions can now be treated at the required site of action leading to more effective treatments and broadening our understanding of biological mechanisms that affect diseases. Despite the increase of treatments and the deepening of our understanding of API release, clinical needs are still unmet and many challenges still remain. Administrative demand has forced new methods of API delivery to be formulated that protect sensitive molecules as well as targeting deep set regions of the body which are often unreachable by oral delivery systems. Advances in synthetic chemistry have allowed for the development of new classes of therapeutic agents that aim to address administrative demands and in tandem with materials science, have allowed release of APIs to occur over extended periods to treat chronic conditions.

The field of controlled delivery of APIs is broadening with new emerging concepts such as systems based on three-dimensional (3D) printed technologies and gene delivery systems becoming useable alternatives (143).

It is important that we continue to strive for a greater understanding of the human body and the DDS we are trying to input. We can begin to exploit expressions exhibited by specific diseases to improve targeting and tailor our systems to maximise therapeutic efficiency. The field of API delivery forms the intersection of chemistry, materials science, medicine and bioengineering. This has proved to be an extremely fruitful area with wide scope for exciting future work.

SQA is a colourless, odourless and tasteless hydrocarbon oil notably resistant to oxidation. It is known to act as an immune system stimulant as well as antioxidant agent. These properties are contributing to it increasingly being used in the pharmaceutical industry. For instance, emulsions of SQA with surfactants are added to vaccines to enhance the immune response (1). A number of studies have proved the properties of SQE and SQA as an antioxidant, drug carrier and emollient (2). The pharmaceutical industry uses SQA in vaccines and in anticancer treatment. SQA and SQE have been proven to have very little toxicity and improve the antibody responses to antigens in a number of primates (3).

Due to its capability to penetrate the human skin, SQA is best known for its applications in the cosmetic industry. The main interest is to transport and increase absorption of other active components (4). The main uses of SQA are shown in Figure 1.

Fig. 1.

The use of SQA is preferred for all of these uses to SQE as this molecule is stable and not susceptible to oxidation (5).

SQA is produced from hydrogenation of SQE, which is a non-saponifiable component of natural lipids. The reaction is shown in Scheme I.

Scheme I

SQE used to be obtained from shark liver oil, although this practice is generally not accepted by cosmetic manufacturers. Today most of the SQE is derived from vegetable sources, such as olive oil and sugarcane. The cost of SQE as well as the impurities present in the reactant depend on the source, as shown in Table I.

The purity of SQE will affect the selectivity towards SQA. For instance, a range of different purities of SQE were studied by Pandarus et al. (4). Results showed that under the same reaction conditions, high purity SQE (98 wt%) achieved 99% selectivity towards SQA, whilst in a lower purity sample (82 wt%) the selectivity was 39%.

This hydrogenation is traditionally carried out in a batch reactor with the use of a nickel based catalyst (6) or more recently palladium catalysts (7). The disadvantages of the traditional process are the fast and irreversible Ni catalyst deactivation and Ni leaching that requires product purification. The cost of the hydrogenation of SQE comprises ca. 40% of the total price of the entire process (8). Therefore, a new catalyst that would overcome disadvantages of the traditional process might bring considerable cost savings. Nevertheless, the other challenge of this reaction is its extreme exothermicity (ΔH_r = –604.26 kJ mol^–1). It is therefore important to consider that the intensified catalytic system will generate more reaction heat, which needs to be removed from the process.

This paper presents results on hydrogenation of SQE in a continuous reactor with use of newly developed catalyst on a ceramic 3D printed support to improve the process from the points of view of both catalyst activity and heat transfer.

The intensification of a catalytic reactor is mainly dependent on the operational limitations of the system and requirements related to the specific chemical reactions. For that purpose, working on the geometry of the catalytic support is one way to reach the desired improvements, which implies the definition of an adequate approach to design such 3D structures.

In a continuous flow reactor, the catalyst needs to be shaped in a form that suits both the process and chemical reaction itself. The shape must consider factors such as mass transfer limitations and reduction of selectivity due to side reactions (9). Packed beds of shaped supports tend to be used in continuous reactors. However, a common problem due to the nature of the shape packing is their associated pressure drop (10) as well as possible flow maldistribution. The latter can result in reactants not accessing the surface of the catalyst uniformly decreasing the overall performance of the process, for instance, by causing wide distribution in residence time that lowers selectivity (9).

Digital methods allow objects to be designed and produced for specific applications. Optimised structures offer the possibility of efficiently controlling the fluid dynamics and temperature uniformity (9). The work presented in this paper uses modelling of the flow phenomena involved in the hydrogenation of SQE to produce a 3D structure that optimises the process.

The catalyst support was based on a ceramic precursor that is commercially available in the form of cylindrical pellets and served as a benchmarking material for the developed 3D printed catalysts. The catalysts were prepared by 3D printing to provide a precursor that was consequently thermally converted to the ceramic of a desired structure. The structure was then impregnated with Pd. A novel preparation method was developed in Johnson Matthey, UK, to ensure high dispersion and high stability for Pd on ceramic. A summary of the physical characteristics of the samples is shown in Table II.

Table II

Catalysts Used for Hydrogenation of SQE

Catalyst	3 mm pellets	1 mm pellets	3D printed (crushed)	3D printed
Form	Extrudates (cylinders)	Extrudates (cylinders)	3D structure crushed	3D structure
Particle size	3 x 3 mm	1 x 3 mm	Irregular particles 1–3 mm	1 x 3 mm
Shape

All the catalysts were supplied by Johnson Matthey and consisted of 2 wt% Pd developmental samples. The testing was carried out using 3.5 mmol_Pd per 100 g of SQE.

Different particle sizes of conventional extrudates were used in an attempt to study the effect of diffusion limitation. Subsequently, the conventional and 3D printed ceramics were compared using crushed 3D printed supports sieved to the same size as the conventional pellets. In the case of crushed 3D printed catalyst, the crushing process preceded the catalyst impregnation to avoid its distribution inhomogeneity.

Tailored supports were designed for the hydrogenation of SQE in a continuous reactor. Because of the extreme exothermicity of this relatively fast reaction, it is desired to control the reaction rate in a way to avoid formation of a narrow reaction zone or hot spots especially at the first part of a reactor. One of the possible ways to achieve this would be to vary the metal loading on the support. However, increasing the loading could encourage the agglomeration of Pd particles and lead to rapid deactivation of the catalyst. An alternative way was to use 3D printing technology for the design of catalytic foams with different density. This, in turn, resulted in a way of keeping metal loading per gram of catalyst while varying the amount of metal in a certain part of the reactor.

Firstly, the structure topology of the support was chosen. Open-cell foams were selected because this type of structure can incorporate attractive benefits, such as suitable thermal conductivity, depending on the chosen material, appropriate mixing, low pressure drop and large surface area per unit volume to increase catalytic activity. Besides these benefits, open-cell foams also offer the advantage of being the repetition of a representative unit cell, which can significantly simplify its manufacture and the modelling of the physical phenomena involved. There is a set of information required to fully define the geometry of a system containing an open-cell foam support: (a) topology of the representative unit cell; (b) values of relevant geometric parameters of the cell; (c) total number of unit cells present in each direction (x, y and z ), i.e. dimensions of the system (overall or sections).

In the specific case of this work, twisted cubic cells were chosen as the base topology due to their simplicity. The geometric parameters to consider were the unit cell width, struts’ diameter, twist angle per unit length of foam and tilt angles of the cells relative to the main flow direction.

A set of foam geometries were tested based on different values of the referred parameters. For each foam, a sequence of steps was followed to assess the best options:

Fig. 2.

Fig. 3.

From the set of geometries investigated, the STL files of the two foams selected are shown in Figure 4. Based on these files, slices of 1 cm height were 3D printed according to both referred designs. The corresponding values of structure porosity (ɛ), specific (geometrical) surface area (A_S) and diameter of struts (d_S) are shown in Table III. For comparison, the properties of the beds consisting of 1 mm and 3 mm conventional pellets (extrudates) are also provided. The specific surface area of Design 2 foams is close to 3 mm pellets while keeping higher bed porosity, which is beneficial for lower pressure drop.

Fig. 4.

The use of two different types of structures, as shown in Figure 5, allowed the initial part of the reactor to achieve a slower rate of SQE hydrogenation, also yielding a lower temperature rise that prevents possible hot spot formation. In the second part of the reactor, more metal can be present per unit volume, which will allow the reaction to be completed to high conversion and thus high SQA yield without compromising the selectivity.

Fig. 5.

The catalyst was prepared by impregnation of a Pd precursor, aiming at 2 wt% Pd, on the different ceramic supports and subsequent reduction. The particle size of the Pd was derived from carbon monoxide (CO) metal area measurements. The sample was reduced under H₂/N₂ at 80°C for 10 min. The catalyst was then cooled under N₂ to 35°C. Subsequently, pulses of CO were introduced. A Pd:CO stoichiometry of 1:1 was assumed and the Pd atomic area was assumed to be 0.0800 nm². The Pd particle size calculated from the CO adsorption measurements was 3–8 nm (11).

The samples were examined in a JEM-2800 Transmission Electron Microscope (JEOL, Japan) using the following instrumental conditions: Voltage 200 kV; C2 aperture 70 μm (Z-contrast) imaging in scanning mode using an off-axis annular detector. The secondary electron (SE) signal was acquired simultaneously with the other scanning transmission electron microscopy (STEM) images providing topological information about the sample.

The STEM images showed the distribution of the metal in the support. Similar distribution was observed for different samples. Figure 6 shows the images corresponding to 1 mm pellets, representative of the catalysts used in the study. The bright areas indicate the location of Pd on the surface of the ceramic material. These images (Figure 6) illustrate a good dispersion and metal distribution of the metallic particles. The particle size was determined to be between 2 nm and 6 nm, similar to those observed via CO metal area.

Fig. 6.

The metal loading for the 3D printed catalysts was kept constant at 2 wt% Pd. The same Johnson Matthey developmental synthetic method was followed for the samples using the 3D printed support.

The metal concentration on the 3D printed foam was investigated via electron probe microanalysis (EPMA) and results for the areas illustrated in Figure 7 are presented in Figure 8. Although the centre of the foams presented lower Pd density, EPMA analysis showed metal present in the inner areas of the support, showing a gradient of concentration.

Fig. 7.

Fig. 8.

Batch Reactor Testing

In a batch reactor, a powder version of the catalysts prepared was tested for stability. For this series of experiments, vegetable SQE was used, with a purity of 82% as it is expected that this would be more representative of the behaviour of the catalyst in the pilot plant. The reaction was carried out for 4 h at 200°C and 20 bar of hydrogen.

The experimental results are shown in Figure 12. The partially hydrogenated SQE intermediates are presented as groups according to their GC-MS MW starting from SQE and its isomers with MW 410 to SQA (MW 422). L1 represents the analyses of the reaction products using a fresh catalyst. Subsequently, L2 to L4 represent runs 2 to 4 using the same catalyst and conditions with a fresh feed. L4a, however, corresponds to the product of the reaction after thoroughly washing the catalyst with xylene. Both activity and selectivity decreased after each cycle. A significant regeneration of both SQE conversion and selectivity towards production of SQA was observed in L4a.

Fig. 12.

This experiment followed a thorough washing of the catalyst attempting its regeneration. A likely explanation for recovering the activity and selectivity is that the washing had removed intermediates of the reaction deposited on the surface of the catalyst. Some waxes were observed in the final product, which tended to dissolve in the feed when hot.

Thermogravimetric experiments showed the presence of a product in the used catalyst after one use that decomposed at ca. 250°C, as shown in Figure 13.

Fig. 13.

In a continuous reactor, the temperature should be kept above 85°C to avoid solidification of the waxes that occur on the surface of the catalyst.

In conventional methods for production of SQE using Ni as the reducing agent, the metal tends to leach onto the final product. In order to avoid toxic metal in the SQA, a costly cleaning step is required after reduction. Although the process presented in this paper is Ni free, it was necessary to determine whether Pd leaches into the final product. Various Pd compounds have been proven to induce sensitisation in several animal species (11). It was thus necessary to test the levels of Pd present in the final SQA. The Pd content for L1 and L4a were analysed by inductively coupled plasma mass spectrometry (ICP-MS) to show levels of the metal below 5 ppb. This proved the safety of using this Pd catalyst for a cosmetic application. Both the absence of Pd in the final product and the possibility of regenerating the catalyst, suggested that the catalyst was stable under the reaction conditions.

A feasibility study was carried out to determine whether it would be possible to perform the hydrogenation of SQE in a continuous reactor. A trickle-bed milli-reactor of a bed volume about 10 ml was employed to perform hydrogenation tests with catalyst of conventional shape. The unit was operated for several days while each experiment at constant operating conditions took at least 6 h. At 180°C and 25 bar a stable SQE conversion 99.5% with SQE yield 69.8% was obtained (see Table IV).

Catalyst deactivation was not observed in the trickle-bed milli-reactor. However, the transmission electron microscopy (TEM) images of the spent catalyst showed a slight aggregation of Pd particles after ca. 40 h of reaction, as illustrated in Figures 14 and 15. Once the hydrogenation process in continuous mode was proven to be feasible in the trickle-bed milli-reactor, experiments in a pilot-plant of volume 324 ml were approached under the same operating conditions.

Fig. 14.

Fig. 15.

The pilot-plant for continuous hydrogenation of SQE is an automated, computer-controlled system (which was developed and built by Advanced Machinery & Technology Chemnitz GmbH within the project PRINTCR3DIT).

With the aim of transferring the SQE hydrogenation from conventional batch to a continuous method a tubular reactor was the most suitable choice. It can be run as a trickle-bed reactor. The reactor itself is made of stainless steel. Solid catalyst can be loaded from the top of the reactor. Figure 16 represents a schematic of the pilot plant, with the reactor highlighted by the red line.

Fig. 16.

For the purpose of a benchmark, a state-of-the-art catalyst supported on conventional 3 mm pellets form (see Table II) was loaded into the reactor. The catalyst was obtained following a proprietary synthetic method from Johnson Matthey. This was compared with a catalyst prepared using a 3D printed ceramic support. No significant pressure drop was observed between the top and the bottom of the reactor.

This conventional catalyst allowed an excellent conversion of SQE into SQA at temperatures from 180°C to 240°C with a pressure of 25–30 bar. However, adjusting the temperature to 240°C and the pressure to 30 bar resulted in high conversion with an acceptable level of isomerisation.

The 3D printed catalyst as a foam (see Table II) was loaded into the reactor. The total weight of the sample was 200 g, forming a catalyst bed of 75 cm high. As for the conventional catalyst, no significant pressure drop was observed during this trial between the top and the bottom of the bed.

Conversion and selectivity were high at low temperatures with the 3D printed catalyst, which differed from the low selectivity observed when using conventional supports. The novel 3D printed catalyst showed high selectivity towards SQA at a temperature of 180°C and a pressure of 25 bar (see Table IV). A possible explanation for this improvement could be the fact that the geometry of the foam enhanced the diffusion of SQE and hydrogen on the surface of the metal.

The effect of temperature on the selectivity to SQA was also investigated, as shown in Figure 17. Operation at temperatures below 180°C allowed good conversion of SQE. However, the selectivity towards SQA becomes insufficient. Nevertheless, at 240°C excellent conversion and selectivity were obtained even for the 100 wt% vegetable SQE feed.

Fig. 17.

The single load of 3D printed catalyst has produced more than 15 kg of good quality SQA up to the time of writing this article with no signs of deactivation. This suggests that the catalytically active phase does not change during the process and the metal does not significantly leach into the product. The use of 3D printed technology allowed optimisation of the reaction conditions and production of large amounts of SQA using a more compact reactor than current technologies.

In principle, it would be possible to carry out the hydrogenation of SQE in a batch reactor with the novel catalyst developed by Johnson Matthey. However, it would require catalyst reactivation between cycles. The use of a continuous reactor presents the advantages of reusability of the catalyst with no significant losses of activity or selectivity. On the other hand, for a typical SQA production plant, a batch reactor of ca. 1.5 m³ would be required, while a continuous system could carry on the process with a volume of about 0.05 m³, 30 times smaller. Being able to control the reaction rate along the reactor bed and preventing formation of undesired isomers is a potential advantage of the use of 3D printing technology. The use of CFD modelling for the efficient design of the support proved crucial in the success of the experiments. Being able to directly print the final catalytic support offered the advantage of an alternative way to change the process, which in turn, would allow higher levels of flexibility. In addition, because of the use of a Pd catalyst, the final product was Ni free, which removed the cleaning step that would normally occur in the manufacturing of SQA for the pharmaceutical and cosmetic industries.