Enzymatic Production of p-Methoxycinnamate Monoglyceride Under Solventless Conditions: Kinetic Analysis and Product Characterization

Abstract

With the increase in biodiesel production experienced in the last decades, biomass-derived glycerol is obtained at a high rate, so glycerol availability in the market has scaled up while this polyol price has been reduced, with the exception of high-quality glycerol. In this context, novel and sustainable products based on glycerol are actively looked for. Octyl-methoxycinnamate (OMC) is a common cosmetic ingredient and sunscreen with potential activity as an endocrine disruptor that is considered an emergent contaminant in aquatic environments. As possible substituents, glycerol-based methoxycinnamates such as monoglycerides can be obtained via lipase-driven esterification. In this work, we develop an enzymatic process under solventless conditions to obtain p-methoxycinnamate monoglyceride under mild conditions using Novozym 435—an immobilized industrial preparation of the lipase B of Candida antarctica—observing the effect of key process variables such as temperature and enzyme, water and acid concentrations. Furthermore, the obtained product was assessed for its activity as UVB-filter and for its stability under irradiation conditions, showing a similar SPF activity and a much higher stability toward photooxidation than OMC.

1. Introduction

In today’s world, even as fossil fuel reserves continue to grow due to increasingly efficient enhanced oil recovery (EOR) techniques [1] and the discovery of new fields [2], there is a perception that fossil fuels, which are non-renewable by nature, will begin to become scarce within a few decades, considering the theory of the Hubbert peak [3]. Even if this theory has been considered incomplete, accepting that there should be a temporal behavior of consumption in the form of multiple peaks, the non-renewable nature of fossil resources is accepted [4]. In this situation, biomass has emerged, again, as a plentiful, renewable resource for energy and products, including chemical products, materials, food and feed, and a critical resource for the development of circular economy [5]. While lignocellulosic biomass, the key raw material type for second generation biorefineries, is obtained each year at a rate of 181.5 billion tons [6], oil-based biomasses are important resources for first, second, third (algae) and fourth (synthetic biology) generation biorefineries [7]. Major edible oils such as palm, rapeseed and soybean oils production amounted to 177 Mtons in 2023–2024, with palm oil production reaching 76.26 Mtons during that period. As the potential oil production from microalgae per hectare is estimated at 136,000 L, compared to 5950 L of palm oil [8], it can be hypothesized that, despite the complexity of algae harvesting and dewatering, the total oil production potential from all sources can exceed 2 billion tons per year. With an increasing emission of CO₂ as a result of fossil fuel burning, part of the solution lies in biofuel consumption, with a potential higher than 3.5 billion TOE by 2050, with FAME biodiesel and biomass-based HVO diesel being the key players in this respect [9]. With a production of 44.7 million tons in 2023, FAME biodiesel is a classical biofuel, whose production involves glycerol as a byproduct. Concomitant glycerol production is increasing, with 2.9 million tons produced in 2023 and an 11.9% CAGR untill 2034, with an even higher potential production and relatively low prices: less than the US dollar (USD) 100 per ton crude glycerol [10].

In this scenario, new products and processes based on glycerol as a biorefinery platform chemical are being developed [11]. Likewise, plentiful lignocellulosic biomass residues rich in polyphenols allow for the development of processes based on classical and advanced liquid–liquid extractions, with diverse green solvents (ethanol, acetone, 2-MTHF, NADES, some ionic liquids) and several intensification approaches (microwaves, ultrasound, cold-plasma) [12]. Polyphenols include phenolic acid based on benzoic and cinnamic acid structures, which have antioxidant, antibacterial, anticancer and sun-protective properties [13]. Both flavonoids and cinnamates are well-known natural photoprotectors [14] and their derivatives, including esters, maintain this particular feature [15,16].

Octyl methoxycinnamate, also known as ethylhexyl methoxycinnamate (EHMC or EMC) or octinoxate, is a common ingredient used in sun care products, lotions, fragrances, and other cosmetics due to its antioxidant, photoprotective and preserving properties, as it can absorb radiation in the UVB range, from 280 to 320 nm. This is the second most produced UV protector worldwide, at a rate of 84.2 million tons per year [17]. As several synthetic UVB photoprotectors, its effect on aquatic environments, human and animal health is being constantly revised. Its hydrophobic nature explains its rapid skin absorption, being detected in human plasma and urine [18]. It is perceived as a potential endocrine disruptor, with effects in the cardiovascular, thyroid, reproductive and immunological systems, and its safety is being tested nowadays, as that of other common UVB protection agents (avobenzone (AVO), oxybenzone, octocrylene (OC), homosalate, octisalate), whose combined action on humans is also under study. In this situation, sun exposure is to be avoided, while protective clothing is recommended and natural derivatives of cinnamic acids [15], lignin and lignin derivatives [19,20], octinoxate analogs [21], etc., are being synthetized and tested. To this end, glycerol and p-methoxycinnamic acid (p-MCA) can be combined to obtain acid glycerides via heat-driven and catalytic processes, using both homogeneous and heterogeneous acid catalysts at moderate-to-high temperatures under non-oxidizing conditions [22,23,24]. In milder conditions, lipases can be used for esterification of p-MCA, and the lipase B of Candida antarctica is a very active biocatalyst for this reaction [25]. This enzyme in particular shows a high activity in hydrolysis of p-nitrophenylbutyrate (pNPB) and in the synthesis of benzoic acid monoglycerides in comparison to lipases from Candida rugosa AY30 and Thermomyces lanuginosus (TLL from Novozymes A/S) [26], suggesting a higher capacity of monoglyceride synthesis in comparison to other relevant industrial lipases. Although free lipases can be attractive from the process perspective in terms of economic feasibility, i.e., in the production of biodiesel with Eversa Transform 2.0 [27], immobilized lipases show higher stability and the possibility to be reused in flow conditions or recycled in batch-wise processes [28]. Novozym 435 has proven to be a successful immobilized biocatalyst for a large number of biocatalytic processes, involving phenolic acids of the cinnamic family in particular [29], even if more advanced immobilization procedures and supports can render more stable or more active biocatalysts [30].

The use of heat or acid catalysts does not allow for a completely selective production of monoglycerides, with kinetic models showing several reactions in series and potential-type kinetic equations [22,23,24]. However, the bulky nature of monoglycerides when using phenolic acids as substrates can lead to the production of monoglycerides without their further conversion into diglycerides [26].

This research aims to investigate p-MCA and its monoesters with glycerol, which are of special interest due to their properties as ultraviolet filters, which is why they could be used as ingredients in sun creams within the cosmetics industry [31]. Additionally, p-methoxycinnamic esters show antioxidant, antimicrobial and anticancer properties, alongside with their hepato-, cardio-, and neuroprotective ability [32,33,34].

The reaction is the esterification of a molecule of p-methoxycinnamic acid (p-MCA) with one of glycerol (G), to obtain a molecule of 1-mono-p-methoxycinnamate of glycerol (MG), as represented in Figure 1.

Unlike traditional synthesis of monoesters with chemical catalysts, acidic or basic, such as Amberlyst 15, 35 and 36 (strong acid resins from Dow Chemical) or more effective catalysts such as sulfonic acids on silica or Preyssler heteropolyacids [24,35], the reaction with lipases takes place under more moderate condition. This aspect is important when working with products that can deteriorate when subjected to high temperatures, as is the case with p-methoxycinnamic acid esters.

As, to the best of our knowledge, there are no studies on the selective synthesis of monoglycerides of the p-MCA in mild conditions using lipases, the aim of this work is to study the effect of diverse reaction conditions (temperature, acid concentration, added water concentration) using glycerol both as substrate and as green solvent A careful kinetic study is performed, determining the most adequate kinetic model for the studied conditions. Finally, key features of the p-methoxycinnamic monoglyceride as UV filter are tested—its UV filter activity and stability of the ester.

2. Results

2.1. Preliminary Experiments

2.1.1. Solubility of Methoxycinnamic Acid in Glycerol/Water Mixtures

The solubility of a chemical compound is determined by its structure, which is responsible for its interactions with the environment. P-MCA, as all cinnamic acids, has an aromatic ring and a vinyl group linked to a carboxylic group. This structure of conjugated double bonds, added to the methoxy group in the para position, gives the molecule greater resonance and stability. These characteristics are beyond its capacity as an ultraviolet filter, but at the same time, they make it more hydrophobic, so its dissolution in a polar medium such as glycerol will be limited.

The solubility of p-MCA has been studied experimentally both in glycerol and in glycerol–water mixtures in different proportions. Figure 2 shows how higher concentrations of water in the mixture with glycerol enhance the acid solubility at any temperature. Furthermore, an exponential increase in the solubility of this acid with rising temperature is observed; this trend, on the other hand, is milder as the polarity of the medium increases. The experimentally measured values can be fitted to a curve in which the solubility is a function of temperature:

$${\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}\hspace{0.33em}=\mathrm{n}{\mathrm{e}}^{\left({\displaystyle \frac{-\mathrm{T}}{\mathrm{m}}}\right)}$$

(1)

The results of these adjustments are collected in

Table 1. Exponential fit to experimental data on the solubility of methoxycinnamic acid in glycerol/water mixtures.

% w/w H2O	n	m	R2
0	1.84 × 10−3 ± 4.2 × 10−4	−22.5 ± 1.6	0.9876
15	1.30 × 10−3 ± 2.6 × 10−4	−23.0 ± 1.5	0.9897
30	1.1 × 10−3 ± 2.3 × 10−4	−23.6 ± 1.6	0.9874

[C_M] in mol/L and [T] in °C.

.

The solubility of p-MCA in glycerol and its mixtures with water is, a priori, one of the factors that can most significantly influence the solventless synthesis of its monoglyceride. Working with a poorly soluble compound necessarily implies operating with very diluted concentrations and working in homogeneous conditions, which is unfavorable in terms of productivity, but one must consider the positive aspects of working in milder conditions, typical of enzymatic synthesis, when working with a heat-sensitive product [26].

2.1.2. Internal Mass Transfer Assessment: Effect of Water Addition

The effect of adding water to the reaction medium on the esterification reaction of glycerol and p-MCA has been evaluated, as it has a relevant effect on the viscosity of the reaction liquid and, thus, on mass transfer phenomena via the molecular diffusivity of any compound involved in the reaction. In Figure 3, we can see that the presence of water in the medium favors the transformation to monoester. Contrary to what might be expected, water, a reaction product present in excess, does not shift the equilibrium toward the reactants when present in moderate concentrations. Its influence on the kinetics of the system seems to be more important, since it considerably modifies the hydrodynamics of the medium and facilitates the diffusion of reactants and products toward and from the active center of the lipase [36]. Figure 4 shows that there is an increase in the reaction rate as the initial quantity of water in the medium rises, reaching a saturation point at 20% w/w of H₂O, after which the initial rate of esterification does not increase.

Glycerol, on the other hand, given its high affinity for water, could reduce the accumulation of the acid near the active center of the catalyst, avoiding shifting the equilibrium of the esterification reaction toward the substrates. In addition, it has already been proven that the presence of water is essential for maintaining the structure of the lipase so that its activity and stability are enhanced [37]. All these phenomena can lead to the conclusion that for the production of glycerol mono-p-methoxycinnamate, MG, it is beneficial to use a medium in which glycerol and a moderate amount of water are mixed.

2.1.3. Internal Mass Transfer Assessment: Effect of Particle Size

Due to the major transport limitations detected inside the particle of the immobilized preparation of CALB, Novozym^®435, in this section, the importance of internal transport limitations in the production of glycerol mono-p-methoxycinnamate is studied and, for this purpose, two series of experiments have been carried out, both in glycerol medium and in a 30% w/w water mixture in glycerol, in which immobilized enzyme has been used after milling and sieving it to obtain fractions of different particle size. The results are presented in Figure 5 and Figure 6. The first figure corresponds to experiments carried out in pure glycerol; in these experiments, the particle size considerably influences the initial reaction rate, indicating the existence of slow mass transfer problems inside the pores of the Novozym^®435 particle. The effect is notable, such that equilibrium is reached more quickly when the particle size is smaller. If an effectiveness factor is calculated, values of around 0.6 are obtained for the native Novozym^®435 particle.

However, when working in a 30% w/w H₂O medium, the effect of reducing the particle size is imperceptible, so it is very likely that the operation is experiencing no transport limitations, with effectiveness factors of 0.96, and a practically 100% utilization of the available catalytic capacity, assuming η = 1 for the asymptotic value of the initial velocity that is reached at smaller particle diameters. This becomes evident when analyzing the initial reaction velocities (Figure 7), practically the same regardless of the catalyst particle size, except for the larger size particles (particle diameter equal to 650 μm), where η is approximately 0.92 (92% of the maximal reaction rate).

2.2. Kinetic Modeling

A series of experiments were carried out to study the kinetics of the p-methoxycinnamate monoglyceride (MG) production. The results of previous experiments were taken into account, alongside the practical limitations of the work in terms of the initial concentration of p-MCA, which is very limited due to the solubility of this acid in glycerol. In addition, the system was also studied when operating in a medium with 30% w/w water, conditions that further restrict the range of concentrations to be considered. Due to this, both for the pure glycerol medium and for the mixture of glycerol and water, concentration values of 20 g/L were used, very close to the limit value of solubility at 50 °C, the lower limit of the temperature range considered (50–70 °C).

Furthermore, also due to the limiting factor of solubility, a single experiment was carried out at a higher concentration of 40 g/L and at the central temperature of the range (60 °C), which is considered, in principle, sufficient to verify the trends and conclusions drawn. The conditions of the experiments performed are given in

Table 2. Operating conditions of experiments on esterification of p-MCA with glycerol catalyzed by Novozym^®435 (N = 250 rpm; C_E = 30 g/L).

Run	T (°C)	CB,0 (g/L)	CW,0 (% w/w)
M1	50	20	0
M2	60	20	0
M3	70	20	0
M4	60	40	0
M5	50	20	30
M6	60	20	30
M7	70	20	30
M8	60	40	30

.

The results obtained for experiments M1 to M8 were used to determine the kinetic model that appropriately fits them and is therefore capable of explaining the phenomenology of the esterification of p-MCA with glycerol using Novozym^®435 as catalyst. In this study, an analogous but separate treatment was given to the experiments where the medium is pure glycerol and to those where the medium is a mixture of glycerol and 30% w/w in water.

The generally accepted model for lipase-catalyzed esterification is based on the Bi-Bi Ping-Pong mechanism [38,39]. One of the reactants, glycerol, is present in great excess, so this expression simplifies to a hyperbolic Michaelis–Menten type equation, as shown in Equation (2):

$${\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{M}\mathrm{G}}}{\mathrm{d}\mathrm{t}}}=-{\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{\mathrm{k}·{\mathrm{C}}_{\mathrm{E}}·{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}{\mathrm{K}+{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}}$$

(2)

Typically, the second reactant, the acid (p-MCA), is under supersaturated conditions (above its solubility point) and the rate of dissolution of the acid (still in a solid state in the medium), as it is consumed because of the reaction, can be considered high enough for the chemical stage to be a limiting factor. The concentration of acid in solution (C_p-MCA) can be considered constant and equal to the solubility point under those conditions, C_p-MCA,S, so that Equation (2) turns into Equation (3). This model corresponds, in fact, to a zero-order model that evolves toward a simple hyperbolic model of the Michaelis–Menten type.

$${\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{M}\mathrm{G}}}{\mathrm{d}\mathrm{t}}}=-{\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{\mathrm{k}·{\mathrm{C}}_{\mathrm{E}}·{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A},\mathrm{S}}}{\mathrm{K}+{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A},\mathrm{S}}}}$$

(3)

Thus, if the reaction proceeds in batch and under supersaturation conditions, the production of glycerol mono-p-methoxycinnamate can be described as a process in two consecutive stages: (1) acid supersaturation conditions (Equation (3)) and (2) homogeneous acid dissolution conditions (Equation (2)). We have tested this kinetic model for the synthesis of MG in pure glycerol and in a mixture with 30% w/w water. Using the integral method, first the fitting of the model to the relevant data was tested, and the physical sense of the proposed parameters were verified; then, the simultaneous adjustment to all the experiments was carried out. One of the restrictions, which at the same time is used as a physical criterion for the discrimination of models, is the Arrhenius equation, which explains the exponential evolution of a kinetic constant with temperature. The non-linear adjustment is carried out by applying the least squares procedure.

The results of the kinetic parameters thus determined are collected in

Table 3. Proposed kinetic model for fitting the experimental data of the esterification reaction of methoxycinnamic acid and glycerol catalyzed by Novozym^®435 in pure glycerol medium and glycerol with 30% w/w added water.

Reaction media	${\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{M}\mathrm{G}}}{\mathrm{d}\mathrm{t}}}=-{\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{\mathrm{k}·{\mathrm{C}}_{\mathrm{E}}·{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A},\mathrm{S}}}{\mathrm{K}+{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A},\mathrm{S}}}}$; $k=k_0·e^{{\displaystyle \frac{-E_{a,k}}{RT}}}$ ${\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{M}\mathrm{G}}}{\mathrm{d}\mathrm{t}}}=-{\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}{\mathrm{d}\mathrm{t}}}={\displaystyle \frac{\mathrm{k}·{\mathrm{C}}_{\mathrm{E}}·{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}{\mathrm{K}+{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}}}\hspace{1em}\hspace{0.33em}\mathrm{i}\mathrm{f}\hspace{1em}{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A}}<{\mathrm{C}}_{\mathrm{p}\text{-}\mathrm{M}\mathrm{C}\mathrm{A},\mathrm{S}}$
Pure glycerol	Parameter	Value ± error	SQR	F95	AICc
	Eak (kJ/mol)	25.7 ± 3.67	1.13 × 10−3	1230	−470
	lnk0	1.50 ± 1.32
	K (mol/L)	0.28 ± 0.002
Glycerol + 30% w/w water	Eak (kJ/mol)	38.2 ± 3.21	7.70 × 10−4	4310	−520
	lnk0	6.80 ± 1.16
	K (mol/L)	0.28 ± 0.004

, for the production of the monoglyceride in pure glycerol and in a 30% w/w water mixture, with lnk₀ as the neperian logarithm of the preexponential factor of the Arrhenius equation for the kinetic constant k and E_ak as the activation energy of such constant, with K being the apparent Michaelis–Menten constant for the acid substrate. As can be seen, the fit is very good. The errors accompanying the different parameters are smaller than the value of the parameter itself, and the statistical parameters have very good values. We can observe a low value for the sum of square residues (SQR), a high value for Fisher F—much higher than the threshold value at 95% confidence for the number of parameters and experimental data used (which is 2.9)—and a low value of the Akaike’s information criterion (AIC), indicating that the model fits correctly to the experimental data. The goodness-of-fit is also evident in Figure 8 and Figure 9, where experimental data are shown as points and the multi-temperature kinetic model fitting is shown as lines (Figure 8 shows the fitting for the pure glycerol experiments, while Figure 9 collects the results for the runs performed in the presence of 30% w/w water). The residue analysis indicates no trends in the errors with the reaction time and low errors between experimental and calculated data (Figure 10 and Figure 11). In short, the magnitude of the error, which in general can be considered very acceptable, is slightly greater at low time value, due to the fact that MG concentration is still small so that the percentual error is considerable, but not in absolute value.

2.3. Product Characterization

2.3.1. UV Filter Activity (SPF)

The ability of p-MCA and, therefore, also of its monoglyceride, to absorb energy in the ultraviolet wavelength range, gives this compound the properties of a radiation filter. As a first step in evaluating this property, an absorption scan of an MG solution was performed with a spectrophotometer, obtaining the spectrum shown in Figure 12. It can be observed that both OMC and MG absorb both in the UV-B wavelength range (280–315 nm) and cover a part of the UV-A wavelength range (315–400 nm). From this, it can be concluded that these compounds are useful as UV filters.

For the quantification of the activity of a compound or a cosmetic/pharmaceutical preparation, a parameter called sun protection factor or SPF has been conventionally adopted, and is defined as the coefficient of the relationship between the minimum time of appearance of erythema—minimum erythematous dose (DEM) in skin protected with a sunscreen and in unprotected skin.

As explained in Section 4.3.4 of this manuscript and, in particular, in the subsection devoted to the measurement of sun protection factor (SPF), multiple factors are taken into consideration in the calculation of this parameter, including the intensity of the incident radiation and the wavelength of the radiation following the method of Diffey and Robson. An in vitro measurement method is applied using a spectrophotometer, very close to some standardized methods such as ISO 24443: 2012 and accepted by the European Cosmetics Association (COLIPA). The results of the acid (12.30 ± 0.26) and the monoglyceride (11.92 ± 0.59) are comparable to those of the reference compound –OMC-(11.16 ± 0.38), so that its use in sun creams replacing OMC seems quite viable.

2.3.2. In Vitro Photostability

Finally, the stability of the UV filter under simulated solar radiation is evaluated, since it is also important to know, apart from its activity, the degree of degradation that this compound may undergo with exposure to radiation, which is a key factor for its final application in sun creams [40,41]. For this study, a solar simulator was used [40]. Samples were prepared and collected over time to proceed to measure their SPF. As a measure of photostability, half-life was defined as indicated in the corresponding subsection (photostability test) of Section 4.3.4.

Figure 13A is a graphical representation of the effect of simulated solar radiation on the measured activity of the ultraviolet filter (SPF) over time, while Figure 13B shows the effect of accumulated solar radiation energy on the normalized SPF, relative to the SPF measured without subjecting the product to solar radiation (Figure 13B); a solar simulator was used in both cases. The greater stability of MG compared to OMC can be observed, since the speed at which it degrades is much lower. This translates into a longer useful life as a UV radiation filter, as can be seen by observing the calculated values of the half-life of the compounds studied, where the half-life or t₅₀ of MG is up to four times greater than that of OMC (1014 ± 85 min and 216 ± 45 min, respectively). When using thermal synthesis at 200 °C of the monoglycerides using a nitrogen stream to avoid oxidation [22], similar results regarding SPF values of the p-methoxycinnamic monoglyceride (12.5 ± 1.4) were obtained, while the half-life was also very similar (1016 ± 60 min). Therefore, the greater photostability of p-methoxycinnamic monoglyceride allows us to consider the viability of applying glycerol mono-p-methoxycinnamate as a UV filter in sun creams.

3. Discussion

The esterification of p-MCA with polyols and sugars can provide a sustainable strategy to obtain more photostable UV filters without the perceived, almost confirmed health issues posed by classical UV filters such as OMC [15,21,22,23,24,25]. However, controlled, sustainable processes toward these ingredients with in-depth kinetic modeling are relatively scarce [22,23,24]. Using glycerol both as a reagent or substrate and as a solvent avoids the use of additional solvents and increases reagent concentration, thereby improving productivity. A high glycerol concentration is desirable to target monoglycerides as products, although, even at glycerol excesses of 3:1 in thermal processes, selectivity to monoglycerides is higher than 80% at 100% acid conversion [22]. This high selectivity becomes absolute with higher excesses of glycerol and enzymatic processing, as observed in this work.

With thermal and acid catalyzed processes [22,23,24], kinetic models are not only more complex, as they should account for the formation of diglycerides, but kinetic equations are simpler, potential, with partial first order for each of the involved reagents (acid and alcohol) in each esterification reaction, as the main phenomena. In addition, there is a parallel, less relevant, concomitant double esterification of glycerol as an added reaction in thermal processes, or the first-order deactivation of the catalyst in catalytic processes. The enzymatic process seems to be more complex, considering the kinetic equation of the sole esterification reaction is taking place, followed by the formation of monoglycerides, although it depends on the concentration of the acid in the reaction media, as it is soluble in pure glycerol up to 2–12 g/L, depending on the temperature (40–80 °C) and, when adding 30% water, this solubility range is between 1–3 g/L. Therefore, for the acid concentration tested, most of the reaction happens under zero-order conditions until this solubility limit is reached. Complex hyperbolic kinetic equations can also be found when using solid Preyssler structure acid catalysts [31], with kinetic models of the LHHW type. In this work, Gallego-Villada et al. observed that these catalysts were active for the esterification of trans-cinnamic acid with n-butanol at 90 to 120 °C and relatively low catalyst concentration (4–5 mM).

When comparing the effect of temperature on the kinetic constants, the activation energy for the enzymatic catalytic constant (in the numerator of the kinetic equation) is 25.7 kJ/mol with pure glycerol and 38.2 kJ/mol when adding up to 30% water, a very mild effect if compared to the activation energy for the thermal reaction between glycerol and one acid molecule: 92.6 kJ/mol [22] and 64–87 kJ/mol for the acid-catalyzed reactions [23,24]. Even if catalytic reactions usually have a lower dependence on temperature (and thus a lower activation energy) than their thermal counterparts, in this case, as indicated in Section 2.1.3, some internal mass transfer hindrance can be expected when using the original immobilized lipase (the particles with an average diameter of 300–400 nm). This is more evident with pure glycerol and almost negligible when adding 30% water, so 38–40 kJ/mol could be the real activation energy for the enzymatic kinetic constant, notably lower than p-toluensulfonic acid (pTSA, homogeneous catalyst) [23] and supported sulfonic acids [24], suggesting a higher active site efficiency of the enzymatic process. In fact, the enzymatic process needs higher amounts of glycerol due to solubility issues and higher reaction time (70 h instead of 1 h) as it proceeds at lower temperatures (40–50 °C instead of 150–200 °C) and lower active site concentrations (30–90 mM for pTSA, 0.16 mM for the Candida antactica lipase B present in Novozym^®435). With similar reactions using 4–5 mM Preyssler acid catalysts [35], the activation energy was 51.9 kJ/mol [35].

When comparing the kinetic model retrieved for this enzymatic process to similar processes driven by lipases, we can observe a higher complexity of the kinetic model for the p-methoxycinnamic monoglyceride synthesis due to the low solubility of the acid substrate in comparison to benzoic acid [26] or ibuprofen [36]. However, the enzyme does not show evidence of deactivation under the conditions tested, similar to those used for benzoic acid and ibuprofen with the free enzyme [26,36], which suggest a higher stability of the immobilized lipase not only to heat but also to the product obtained, which could accumulate near the enzyme showing deleterious effects due to product surface accumulation and fouling [26,36,41,42]. When esterifying phenolic acids with glycerol or, in general, with monosaccharides, the presence of water can be highly relevant, with increasing initial reactions rates and yields as added water increases up to a certain value, from which the equilibrium is shifted toward the substrates, thus reducing the ester yield [36,43,44].

As with other sugars, the presence of glycerol in the ester could have low cytotoxicity [15], a feature of sunscreen ingredients of notable relevance that can determine its in-silico design before the chemical synthesis [21]. These compounds also showed a high-to-very-high SPF activity, similar to that of OMC ([15], this work). Finally, sunscreens with OMC are relatively unstable to sunlight or UV irradiation in comparison to other ones with inorganic UV filters, as shown by Gonzalez et al. [45], which means there is a need to add photostabilizers and/or carefully mix diverse UV filters to obtain a positive synergy in this regard [45,46]. Therefore, the synthesis of p-MCA derivatives with a higher photostability, as observed in this work, would be of great interest for use as a sole UV filter or as an ingredient in mixtures for sunscreens.

4. Materials and Methods

4.1. Materials

For heterogeneous esterification reaction of p-MCA and glycerol using enzymes as catalyst, the following reagents were used: p-methoxycinnamic acid or p-MCA (trans-4-methoxycinnamic acid) (98%) (Merck, Darmstadt, Germany), extra pure glycerol (>99.8%) (Fisher Scientific, Hampton, NH, USA), commercial enzyme Novozym^®435 and HPLC-grade methanol (Scharlab, Barcelona, Spain).

4.2. Equipment

4.2.1. Experimental Facility

For glycerol esterification experiments with p-MCA using an immobilized industrial derivative of Candida antarctica lipase B (Novozym^®435) as catalyst, a facility consisting of a magnetically stirred isothermal batch reactor was used. It consists of a glass beaker with a rounded bottom and Teflon disk agitation, a combination that prevents abrasion of the biocatalyst, thus avoiding erratic behavior due to deactivation phenomena, leaching, among others. The vessel is immersed in a glycerol bath (IKA Yellow Line, model MSC basic C). Its temperature is measured with a platinum probe that is coupled to a temperature controller (IKA, model TC3). The stirring speed can also be regulated and kept controlled at the setpoint value.

4.2.2. Analysis of Samples

The evolution of the compounds during the esterification reaction of p-MCA with glycerol using Novozym^®435 as biocatalyst, as well as some of the properties of the products obtained, were carried out by Reversed-Phase High Performance Liquid Chromatography (HPLC) (JASCO, LC-2000 series device equipped with a refraction index detector).

The solubility study of the p-MCA was carried out in a cylindrical stainless steel reactor with a capacity of 50 mL, heated by a hot plate. A Teflon lid was included where the thermocouple and the spectrophotometer sensor were inserted through two holes on it. A magnet with a speed of 250 rpm was used for stirring. The temperature was measured with a thermocouple and controlled with a PID controller. The reactor was connected by means of a sensor with fiber optic cable for online measurement in reactor (661662-UV Ultra-Mini-Probe, Metrohm Hispania, Madrid, Spain), to a UV/VIS spectrophotometer (Jasco, model V-630, Jasco Corporation, Tokyo, Japan) which measures the absorbance at a wavelength of 600 nm (turbidity) of the solution for each temperature studied.

The measurement of absorbance over a given wavelength range of values to determine the UV filter activity of the glycerides studied (Sun Protection Factor, SPF) were carried out using a Jasco UV/VIS spectrophotometer, model V-630. Its wavelength measurement range is from 190 to 1100 nm, and the spectral bandwidth is 2 nm.

In order to study the degradation of the UV filtering properties of p-methoxycinnamic acid glycerides, as a consequence of their exposure to the sun (simulating conditions of use), the samples were subjected to the radiation of a solar simulator lamp (model HQI-R-250W, Osram, Muchich, Germany) located in the Photovoltaic Solar Energy Unit of the Energy, Environmental and Technological Research Center (CIEMAT).

4.3. Methods

4.3.1. Development of an Esterification Experiment

Experiments were carried out in a round bottom beaker by adding the appropriate weight of acid to reach the desired concentration (20–100 g/L) to 20 mL of pure glycerol. It was then placed in a glycerol bath, which maintains a constant and controlled temperature (50–80 °C), and the contents of the beaker was stirred to facilitate the dissolution of the acid and thermostatized the reaction mixture. Once the acid was dissolved, the enzyme derivative was added.

The samples taken during the transformation were frozen and diluted in a 10 g/L solution of pure methanol. They were filtered with a 0.5 µm light Teflon filter (Advantec, model DISMIC-13 JP, Irvine, CA, USA) and diluted in isopropanol before HPLC analysis.

4.3.2. High Performance Liquid Chromatography (HPLC)

Reaction compounds separation, for the subsequent calculation of their concentration, was carried out by means of the HPLC chromatograph described in Section 4.2.2.

A Mediterranea Sea-18 column from Teknokroma was used as the nonpolar stationary phase. As a relatively polar mobile phase, a 55:45 v/v mixture of methanol–aqueous solution of sulfuric acid pH 2.2 was used, flowing isocratically with a flow rate of 0.8 L/min. The column temperature was kept constant at 50 °C.

The absorbance of the different compounds was measured at 254 nm, obtaining a chromatogram that records the peaks of the analytes, determining the retention time and peak area of each compound. The peak area of each compound is the value used to calculate its concentration through a previously performed calibration. The internal standard used for the calibration was methyl benzoate 10 g/L.

The progress of the reaction was followed by the disappearance of p-MCA, applying the calibration equation. Due to stoichiometry and in the absence of other products, the conversion to monoglyceride corresponds to that of acid in a one-to-one ratio. In this way, the concentration of the monoglyceride can be calculated.

4.3.3. Mathematical Analysis: Kinetic Model Fitting

Fitting of the proposed kinetic model with Aspen Custom Modeler v14 was performed with a Levenberg–Marquardt algorithm for non-linear regression of experimental data coupled to the numerical integration of the mass balance of the monoglyceride (results in a non-linear ODE) by means of a fourth-order Runge-Kutta method.

The model goodness-of-fit to experimental data was assessed on the basis of conventional goodness-of-fit statistical criteria, among which are Fischer’s F-value and Akaike’s information criterion (AIC), both used as information criteria. The value of the former is based upon a null hypothesis that considers that the model fits appropriately to the experimental data. It is defined as follows

$$F_{95}={\displaystyle \frac{{\displaystyle \sum _{n=1}^N\frac{{\left(y_{calc}\right)}^2}{K}}}{{\displaystyle \sum _{n=1}^N\frac{SQR}{N-K}}}}$$

(4)

where N is the total number of data, K is the number of parameters, SQR is the sum of quadratic residues defined as ${\left(y_{\exp }-y_{calc}\right)}^2$, with y_exp and y_calc referring to the experimental and calculated values of the dependent variables, respectively.

AIC also considers both the total amount of data and the number of parameters. AIC can be calculated from the following equation:

$$AIC=N·\ln \left({\displaystyle \frac{SQR}{N}}\right)+2K$$

(5)

4.3.4. Determination of Product Properties

• Activity test: Measurement of sun protection factor (SPF)

Compounds that have ultraviolet filter properties are capable of absorbing radiation at the wavelength corresponding to ultraviolet radiation (200–400 nm).

To measure the capacity of a compound as an ultraviolet filter, the sun protection factor (FPS) was used. In this work, an in vitro SPF measurement procedure developed by the company Jasco Europe was selected, which is significantly close to the SPF data obtained in vivo through other measurement tests. This method was designed to be carried out with a double beam UV-Vis spectrophotometer, specifically with the Jasco V-630/SPF model. The calculation of the SPF was carried out by applying the following equation [47]

$$\mathrm{S}\mathrm{P}\mathrm{F}={\displaystyle \frac{{\sum }_{(400-290)}{\mathrm{E}}_{\mathrm{\lambda }}.{\mathrm{B}}_{\mathrm{\lambda }}}{{\sum }_{(400-290)}{\mathrm{E}}_{\mathrm{\lambda }}.{\mathrm{B}}_{\mathrm{\lambda }}.{\mathrm{T}}_{\mathrm{\lambda }}}}$$

(6)

where E_λ corresponds to the spectrum of solar irradiation on the earth, B_λ corresponds to the effectiveness, for each wavelength, of causing damage to the skin (erythematogenic efficiency) and T_λ is the spectral transmittance of the sample.

The samples were applied, uniformly, on top of a particularly porous and ultraviolet-permeable membrane. To do this, a 3M brand tape (Transpore 3M -3M, Minnesota, USA-with dimensions 25 mm × 5 m) was used to simulate an application similar to that of the product on the skin according to the in vivo method. The amount applied should be 2 mg/cm². A tape size of 2.5 cm × 4 cm was used, that is, 10 cm² and, therefore, 20 mg of sample was applied to the tape.

Double beam UV-Vis spectrophotometer (Jasco V-630 model) was used to measure the spectral absorbance curve in the ultraviolet region (400–290 nm), as it allows reducing the light-scattering effect due to the opalescence of the membrane. To carry out the measurements, a piece of adhesive tape was placed in the reference window of the spectrophotometer and another with the sample (with the dimensions mentioned above) in the measurement window. A scan was made in the UV wavelength range. The measurements for each sample was generally taken six times, varying the position of the membrane with respect to the radiation; in this way the possible homogenization error was minimized. Subsequently, the average of the six measurements was calculated, obtaining the final SPF value of the said compound using a computer program developed by the Jasco company and following the method of Diffey and Robson [47].

• Photostability test in solar simulators

Samples belonging to the catalytic esterification of glycerol with methoxycinnamic acid were subjected to radiation from the HQI-R 250W lamp. This lamp reproduces, as accurately as possible, the spectrum of solar radiation. The samples were irradiated for 19 hours, with samples taken every hour. Subsequently, they were analyzed with the spectrophotometer (double beam UV-VIS, Jasco V-630/SPF) to measure their SPF, following the procedure explained in the previous section. When collecting the different absorption spectra of the monoglyceride of methoxycinnamic acid at different radiation times, it can be observed that the time of exposure to radiation influences the degradation of the sample, progressively losing the absorption capacity in the established wavelength range and, therefore, losing properties as an ultraviolet filter.

To carry out the photostability test, the samples were spread evenly on a particularly porous tape (Transpore 3M) with the help of a small rectangular spatula. Its measurement should be 2.5 cm × 4 cm, and the amount of sample applied should be 20 mg. The tape, with the sample applied, was placed on a transparent glass slide and placed under the HQI-R 250 W lamp to be irradiated. Several slides were prepared, all with the same type of sample, and they were removed from radiation exposure at a certain time (every hour and every hour and a half). Once each sample was irradiated, the absorbance was measured, and the SPF value was calculated.