ADH-catalysed hexanol oxidation with fully integrated NADH regeneration performed in microreactors connected in series

Abstract

Hexanal is produced by the oxidation of hexanol using NADH dependent alcohol dehydrogenase (ADH). Coenzyme, NADH regeneration is needed in order to make the ADH-catalysed process more sustainable. In this investigation, the coenzyme regeneration was catalysed by the same enzyme, ADH, which catalysed the main reaction, i.e., the oxidation of hexanol. Different sources of ADH were studied (suspended and immobilized enzyme ADH and permeabilized baker's yeast cells) to find the optimal catalyst. The best results were obtained by using suspended enzyme, where 100% conversion of the coenzyme was achieved with a very short residence time (τ = 0.8 s). The final phase of investigation was development of an integrated system with two microreactor chips connected in series. The first chip was used for hexanol oxidation and the second for the simultaneous coenzyme regeneration. Regenerated coenzyme was reused by recirculation in to the first chip, where the oxidation step was continuously performed for three days without the need for the addition of fresh coenzyme.

---

1. Introduction

During the last decade, there has been an increase in interest for natural products. “Green note” aldehydes and alcohols are high-value molecules widely used in the aroma industry to impart the green character associated with freshness. One of the most widely used is n-hexanal. n-Hexanal has a pleasant grassy odor^1,2 and no known adverse health effects and because of this latter quality has been approved as a food additive by the United States Food and Drug Administration,³ namely, as a synthetic flavoring substance and adjuvant if it is used in the minimum quantity required to produce its intended effect, and otherwise in accordance with all the principles of good manufacturing practice. Several processes for the production of natural chemicals based on fermentation, extraction from plants, and enzyme-catalyzed reactions have been described.^4–6 Santiago-Gómez et al.⁷ reported hexanal production by cells of a recombinant Yarrowia lipolytica yeast, using reducing dithiotreitol and oxidizing potassium ferricyanide. They obtained 6 mmol dm⁻³ of hexanal, which was an improvement over some procedures previously described in literature.^4,8,9 Márczy et al.⁵ reported hexanol production from hydrolyzed sunflower oil in a two-step reaction, resulting in 8.2 mmol dm⁻³ of hexanal. Karra-Chaabouni et al.¹⁰ investigated the use of alcohol oxidase from Pichia pastoris together with catalase from bovine liver to oxidase n-hexanol to n-hexanal. In this case, the necessary cofactor, flavine adenine dinucleotide (FAD), is incorporated into the molecular structure of the biocatalyst and regenerated during the process with the addition of oxygen. Catalase was used to convert the regeneration by-product, H₂O₂, to water and oxygen. Gandolfi et al.¹¹ proposed the production of aldehydes or carboxylic acids by oxidation of the corresponding primary alcohols with acetic acid bacteria (Acetobacter sp. or Gluconobacter asai). They obtained >97% yield (t = 3 h) of n-hexanal. In addition to the enzyme extracts, hexanal can also be produced from linoleic acid, but then enzymes from linoleic acid cycle (LOX) isolated form different plants (such as mint, strawberry, and tomato) are used.^12,13

The general opinion is that the demand for green note chemicals would increase if it became possible to produce them in a more economical way.¹⁴ In addition, the use of enzymes for catalysis compared to classical chemical catalysts is highly desired in food industrial processes, because the resultant products would be classified as “natural” by food regulatory agencies, a feature that increases their public acceptance as ingredients for foods.¹⁵

By combining enzymatic biocatalysis and microreactor technology, a new way for hexanal production was proposed.^16,17 A high surface to volume ratio, faster diffusion dominated transport, enhanced heat transfer and thus reduced energy demands, good process control, high throughput, usage of minimal (microlitres) of reagent volumes, etc., are some of a microreactor advantages that are usually stressed.¹⁸ In our previous work, we demonstrated that microreactors could be a better solution for hexanal production in comparison to traditional biotransformation in a batch system (macroreactor).^16,17 On the other hand, the developed system had one significant drawback, whereby constant usage of the new enzyme and coenzyme was necessary to feed in to the system when long term continuous biocatalysis was performed. Therefore, the next step in the system development is investigation of an efficient coenzyme regeneration system. Herein, in order to make the process even more sustainable, the same enzyme used in the hexanal production reaction, ADH, was used for coenzyme regeneration. Preliminary investigations of coenzyme regeneration were performed and different regeneration and reactor systems were investigated and compared.^19–22

In this study, the final step, an integrated system composed of biocatalytic hexanol oxidation with simultaneous coenzyme regeneration, was investigated. Enzyme alcohol dehydrogenase (ADH) isolated from baker's yeast was used for both hexanol oxidation and simultaneous coenzyme regeneration. Acetaldehyde was used as a substrate for the coenzyme regeneration because of its low price and high specificity of the ADH towards it.^23–25 Three different microreactor systems for the biocatalytic hexanol oxidation with simultaneous coenzyme regeneration (namely, single microreactor chip, two microreactor chips connected in series, and three microreactor chips connected in series) were investigated and compared.

2. Materials and methods

2.1. Materials

2.1.1. Chemicals. NAD⁺ and NADH were purchased from Jülich Fine Chemicals (Germany). Acetaldehyde and hexanal were purchased from Fluka A.G. (Switzerland). Ethanol, glycine, and Na₂P₂O₇·10H₂O were purchased from Kemika (Croatia), and acetonitrile was purchased from Fisher Scientific (Great Britain). Hexanol was purchased from Merck (Germany) and cyclohexanone was purchased from Carlo Erba (Italy), whereas hexane was purchased from Sigma (Germany). Alcohol dehydrogenase from baker's yeast (1.1.1.1), with an activity of V.A. = 451 U mg⁻¹ (where 1 U is the amount of enzyme that catalyzes the conversion of 1 μmol of substrate per minute under standard conditions (pH = 7, T = 25 °C)), was purchased from Sigma (Germany).

2.1.2. Microreactors. One microreactor system that was used consisted of two smooth glass microchannels with relative roughness around 1% (length : width : depth = 330 mm : 250 μm : 50 μm, and an internal volume of 6 mm³, one with two “Y” shaped inlets and outlets and one with one inlet and one “Y” shaped outlet; Micronit Microfluidics B.V., Netherlands) and another microreactor consisted of micromixers with a relative roughness around 10% (length : width : depth = 53.3 mm : 200 μm : 150 μm and the internal volume of 2 mm³, equipped with two “Y” shaped inlets and one outlet; Micronit Microfluidics B.V., Netherlands). All the microreactors were fixed in a stainless steel housing (Micronit Microfluidics B.V., Enschede, Netherlands). Two syringe pumps (PHD 4400 Syringe Pump Series, Harvard Apparatus, USA) equipped with high pressure stainless steel syringe (8 cm³, Harvard Apparatus, USA) were used to feed the substrate and coenzyme in to the microreactors. A syringe pump (PHD 33 Syringe Pump Series, Harvard Apparatus, USA) equipped with a valve box and with a plastic syringe (10 cm³, BD Plastipak™, Spain) was used for the recirculation.

The flow profiles in the microchannel were monitored under a microscope (Motic B1-220A, binocular, Weltzar, Germany) at magnifications of 40× and 100× (eyepiece magnification = 10×; objective magnification = 4× and 10×).

2.2. Methods

2.2.1. Microreactor experiments. Hexanol oxidation (Fig. 1c) was carried out in three different systems.

Fig. 1 First microreactor system experimental setup (a and b) and reaction scheme (c) for ADH-catalyzed hexanol oxidation with NADH coenzyme regeneration based on acetaldehyde reduction.

The first system (Fig. 1a and 2a) consisted of two glass microreactor chips (both with two “Y” shaped inlets and outlets) connected in series. The first microreactor was used for hexanol oxidation, while a second one was used for simultaneous coenzyme regeneration. Hexanol dissolved in hexane (c_i,hexanol = 5.5 mmol dm⁻³, Φ = 30 mm³ min⁻¹) was fed in the first microreactor using the syringe pump. A mixture of suspended enzyme (γ_i,enzyme = 0.1 mg cm⁻³, t = 0) and coenzyme (c_i,NAD⁺ = 5.5 mmol dm⁻³, t = 0) dissolved in aqueous buffer (75 mmol dm⁻³ glycine-pyrophosphate buffer, pH = 9, T = 25 °C) were fed (Φ = 10 mm³ min⁻¹) as the second inlet stream in to the first microreactor, using the syringe pump for recirculation. At the outlet of the first microreactor, an aqueous phase containing the enzyme and coenzyme was directed in to the second microreactor, where coenzyme regeneration takes place. Acetaldehyde (c_{i,acetaldehyde} = 12.5 mmol dm⁻³) dissolved in buffer was used as a substrate and fed (Φ = 10 mm³ min⁻¹) as the second inlet stream in to the second microreactor, using an additional syringe pump. One outlet of the second microreactor containing the regenerated coenzyme was reused as the mixture of enzyme and coenzyme and fed as the second inlet stream into the first microreactor, using the above-mentioned syringe pump for recirculation.

Fig. 2 Process scheme for hexanol oxidation: (a) first microreactor system: two microchips connected in series, with hexanol oxidation in the first chip, coenzyme regeneration in the second chip, and with enzyme and coenzyme recirculation; (b) second microreactor system: three microreactors connected in series, with enhanced mass transfer in the first chip (micromixers), phase separation in the second chip, and coenzyme regeneration in the third chip, with enzyme and coenzyme recirculation; and (c) third microreactor system: integrated hexanol oxidation and coenzyme regeneration in one microreactor, with enzyme and coenzyme recirculation.

The second system (Fig. 2b) was similar to the first one, apart from two glass microreactors chips connected in to the series, but a microreactor with micromixers was integrated in front of the microreactor for hexanol oxidation (equipped with one inlet and one “Y” shaped outlet), in order to enhance the mass transfer. The initial concentrations of compounds in all the inlet streams were equal to those in the first system.

The third system (Fig. 2c) consisted of a one glass microreactor chip (with two “Y” shaped inlets and outlets). A mixture of hexanol (c_i,hexanol = 5.5 mmol dm⁻³) and acetaldehyde (c_{i,acetaldehyde} = 12.5 mmol dm⁻³) was dissolved in hexane and fed in to a microreactor (Φ = 30 mm³ min⁻¹) using a syringe pump. A mixture of suspended enzyme (γ_i,enzyme = 0.1 mg cm⁻³, t = 0) and coenzyme (c_i,NAD⁺ = 5.5 mmol dm⁻³, t = 0) dissolved in aqueous buffer (75 mmol dm⁻³ glycine-pyrophosphate buffer, pH = 9, T = 25 °C) was constantly recirculated, using the pump for recirculation.

Outflows from the microreactors containing the substrate and the product were collected in vials placed on ice, to stop the reaction via enzyme inactivation. Measurements of all the experiments presented in this paper were performed in triplicate, and with a 95% confidence range, the results showed no significant difference.

2.2.2. Analytics.
2.2.2.1. Alcohol dehydrogenase assay. The alcohol dehydrogenase activity was measured before every experiment using an enzymatic assay,²⁶ where the NADH concentration was determined by measuring the absorbance at λ = 340 nm with a spectrophotometer (Shimadzu UV-1601, Kyoto, Japan).
2.2.2.2. Determination of hexanol and hexanal concentration. Hexanol and hexanal concentrations in the organic phase were determinated using GC (Shimadzu GC-2014, Kyoto, Japan) with a flame ionization detector, according to the procedure described elsewhere.²⁷
2.2.2.3. Determination of ethanol concentration. Ethanol was analysed by GC (Shimadzu GC-2014, Kyoto, Japan) according to the procedure described elsewhere.²¹
2.2.2.4. Determination of coenzyme concentration. Spectrophotometric determination (Shimadzu UV-1601, Kyoto, Japan) of the NADH coenzyme concentration was performed before and during the experiments at λ = 340 nm. The calibration curve was prepared using a standard solution of NADH (0–0.25 mmol dm⁻³, R² = 0.9988). The concentration of NAD⁺ was calculated according to the stoichiometric ratio and the measured concentrations of ethanol and NADH.

3. Results and discussion

3.1. Hexanol oxidation

As mentioned in the introduction, in our previous work, we developed a hexanol oxidation process using enzyme alcohol dehydrogenase as a biocatalyst.^16,17,27 Different types of microreactors (tubular microreactor and microreactor equipped with micromixers), and different inlet concentrations of biocatalyst and substrates, as well as different flow rate ratios of the aqueous and the organic phase were investigated. The results obtained indicated that microreactors could be a better solution for hexanal production compared to traditional biotransformation in a batch reactor system (X = 5.3%, t = 3 min). Taking into consideration that the reaction occurs on the interphase area between two phases, the number and volume of segments (i.e., the typical flow observed in a microreactor with rough walls) were measured and also recalculated for all types of reactor. It was observed that smaller segments (therefore with a higher surface to volume ratio) appear in a microreactor equipped with micromixers. Analysing and comparing the obtained results for different types of microreactors¹⁵ revealed that the highest conversions (X = 10%) of hexanol to hexanal were achieved in a microreactor equipped with micromixers in comparison to conversions obtained for the same residence time in a tubular microreactor (X = 0.8%). These results indicate that the mixing had a great effect on mass transfer, and that with the introduction of mixing smaller slugs were obtained, ensuring a higher area between the phases, and ultimately resulting in higher conversions.

On the other hand, in order to develop a fully integrated system for hexanol oxidation, complete phase separation at the outlet of a microreactor is essential. As mentioned in the methods section, the idea is to direct an aqueous phase containing the enzyme and coenzyme at the outlet of the first microreactor in to a second microreactor chip, where the coenzyme is regenerated. According to Žnidaršič-Plazl and Plazl,²⁸ by using a microreactor system consisting of a smooth glass microchannel with a relative roughness of around 1% and with “Y” shaped inlet and outlet channels, a parallel fluid flow could be developed. Additionally, they noticed that when the aqueous and the organic phase with a relative viscosity of 1 : 3 at 25 °C entered the microchannel system at the same flow rate, the less viscous n-hexane (organic phase) occupied a much smaller part of the channel. In order to position the interphase area in the middle of the channel and to enable separation at the “Y” shaped exit of the system, the flow rate of the less viscous phase was elevated in terms of its relative viscosity – in the case of the aqueous : hexane system 3-fold at 25 °C. Before the hexanol oxidation experiments in the smooth wall microreactor chips were carried out, the flow profile and flow stability along the microreactor length were investigated. Here, the solutions were fed in to the microchannel at a 3 : 1 volumetric flow ratio, and the obtained results are presented in Fig. 3. It can be seen that the flow instability for all flow ratios was lower than 30 : 10 mm³ min⁻¹ (Fig. 3a and b). By operating at the stated flow ratio or higher, a parallel and stable fluid flow could be developed from the entrance to the end of the microchannel (Fig. 3c and d).

Fig. 3 Microscopic observation of the flow pattern of an aqueous and organic phase in the middle section of microreactor for a flow ratio of (a) 3 : 1 mm³ min⁻¹, (b) 15 : 5 mm³ min⁻¹, (c) 30 : 10 mm³ min⁻¹ and (d) at the exit of the microreactor (30 : 10 mm³ min⁻¹).

Since a stable fluid flow, with the interphase positioned exactly in the middle of the microchannel, was necessary for development of the integrated system, all the experiments were performed at the flow ratio of 30 : 10 mm³ min⁻¹.

Moreover, another effect that was important to investigate prior to development of the integrated system is component diffusion from one phase to the other. Here, it was important to find the optimal residence time, i.e., where the diffusion of components (especially NAD⁺ and the enzyme, as they need to be recirculated in the system) will be negligible. In order to estimate the diffusion of molecules from one phase to the other, coenzyme (c_NADH = 8.5 mmol dm⁻³) and enzyme (V.A. = 35 U cm⁻³) were dissolved in the aqueous phase, while the organic phase was kept component free. The results are presented in Fig. 4.

Fig. 4 Distribution of (a) NADH and (b) enzyme in the aqueous (○) and organic phase (●), as a function of the residence time for a flow ratio of 3 : 1.

As can be observed, for the lowest tested total flow rate (Φ = 40 mm³ min⁻¹, τ = 9 s), a low molecular diffusion of coenzyme (Fig. 4a) and enzyme (Fig. 4b) can be seen. Because of the flow instability, the measurements were not performed for the higher residence times. Additionally, in experiments performed with ADH, a significant deactivation of the enzyme in the organic phase was noticed.

Diffusion measurements were not performed for hexanol and hexanal, due to the low solubility of these compounds in water. Measurements for NAD⁺ were also not performed, due to the lack of an analytical method, but it could be assumed that a similar behavior to that of NADH would be present.

Ethanol diffusion measurements performed for the total flow rate of Φ = 40 mm³ min⁻¹, τ = 9 s, demonstrate that ethanol is almost completely diffused from one phase to the other.¹⁹ Ethanol and acetaldehyde are very similar compounds with regards to their properties. Therefore one could expect a similar behavior of acetaldehyde with regards to its phase distribution during microreactor experiments.

Before development of the integrated system, hexanol oxidation was performed in a microreactor with a smooth microchannel wall, and the corresponding results are presented in Fig. 5.

Fig. 5 Hexanol oxidation in a microreactor with (●) smooth and rough (○) channel walls for a phase flow ratio of: (a) 1 : 1, and (b) 3 : 1 and (c) size of the interface area for both types of reactors.

In order to compare hexanol oxidation with the previous obtained results,¹⁶ oxidation was first performed with a 1 : 1 phase flow ratio (c_i,hexanol = 5.5 mmol dm⁻³, c_i,NAD⁺ = 5.5 mmol dm⁻³, γ_i,ADH = 0.1 mg cm⁻³). The conversion obtained in a microreactor with smooth walls was approximately twice as high (τ = 36 s; X = 60%) than the one obtained in a microreactor with rough walls (τ = 36 s; X = 30%) for the same residence time. According to Harries et al.,²⁹ mass transfer occurs at the interface area between phases. The size of the interface area for a microreactor with smooth walls was calculated according to the microchannel length and height and based on microscopic observation of the stable fluid flow. The interface area was calculated to be 1.66 × 10⁻⁵ m². Because the size of segments changes with the change of residence time in a microreactor with rough walls, the interface area was calculated for different residence times, under the assumption that the segments have a rectangular microchannel shape. Comparisons of the interface areas are presented in Fig. 5c. As can be observed, a microreactor with smooth walls has a larger interface area for the same residence time compared to a microreactor with rough walls, and so a more efficient mass transfer, meaning a higher conversion, can be achieved in that type of microreactor.

As mentioned earlier, only for the volumetric ratio of organic and aqueous phase of 3 : 1 is the interphase area positioned exactly in the middle of the microchannel. Unfortunately, the conversion (Fig. 5b) obtained for this ratio was 3-fold (τ = 36 s; X = 20%) smaller compared to the conversion obtained for the same residence time where the flow ratio was 1 : 1.

3.2. Coenzyme regeneration

In order to make the hexanol oxidation system sustainable, it was necessary to develop an efficient coenzyme regeneration system. To solve this problem, coenzyme regeneration was performed catalysed by the same enzyme (ADH) that catalysed the main reaction, with acetaldehyde used as the substrate. Disadvantages of the selected reaction system include the possibility of enzyme deactivation, both by the substrate, acetaldehyde, and also by the product, ethanol; instability (i.e., the possibility of self-condensation in the solution); and the volatility of acetaldehyde.³⁰ These problems can be solved by using a continuously operated microchannel system at different flow rates, which provide for only a short contact time between the enzyme and the components with inhibition effects.

Different forms of biocatalyst (suspended and immobilized ADH, and permeabilized baker's yeast cells as the source of ADH), in different reactors (polytetrafluoroethylene (PTFE) and glass microreactor, microreactor with an electromagnet and oscillating magnetic field, batch reactor) were studied as possible solutions for the integrated coenzyme regeneration. A summary of the results is presented in Table 1.

Table 1 Comparison of NAD⁺ coenzyme regeneration within different types of microreactor and a batch reactor using immobilized and suspended biocatalyst (enzyme and permeabilized baker yeast cells)

	Type of biocatalyst			Type of reactor	τ (s)	X (%)	Reference
Suspended biocatalyst	Enzyme			Batch reactor (t = 40 s)	—	92.73	31
				Glass microreactor	2	95.89	19
				PTFE microreactor	47.10	94.36 (±1.62)	21
	Permeabilized baker's yeast cells			Batch reactor (t = 20 s)	—	91.36	31
				Glass microreactor	7.2	69.86 (±1.12)	21
				PTFE microreactor	47.10	86.67 (±1.83)	21
Immobilized biocatalyst	On the wall	Enzyme		Glass microreactor	3.6	11.99 (±2.29)	21
				PTFE microreactor	94.3	11.91 (±1.33)	21
		Permeabilized baker's yeast cells		Glass microreactor	3.6	7.80 (±3.55)	21
				PTFE microreactor	94.3	6.58 (±1.56)	21
	On the nanoparticles	Enzyme	Permanent magnet	PTFE microreactor	6000	100 (±1.21)	21
			Electromagnet and oscillating magnetic field	PTFE microreactor	180	100 (±1.85)	23
			Mechanical stirring	Batch reactor	240	100	23

---

The best results were obtained using a suspended enzyme in a microreactor with smooth microchannel walls, where a 100% conversion of NADH was achieved (τ = 0.8 s). Therefore, this reaction system was used for the development of ADH-catalysed hexanol oxidation with fully integrated NADH regeneration.

3.3. Integrated system

The final phase of the investigation was development of the integrated systems. Three different microreactor systems (Fig. 2) were investigated and compared. In the first system, two microreactor chips were connected in series. As mentioned earlier, the first chip was used for hexanol oxidation and the second one for the simultaneous coenzyme regeneration (Fig. 2a). Hexanol dissolved in hexane was fed in to a microreactor at a flow rate of Φ = 30 mm³ min⁻¹. Using a second syringe pump, a mixture of suspended enzyme and coenzyme dissolved in aqueous buffer was fed in to the microchannel at a flow rate of Φ = 10 mm³ min⁻¹ to obtain a 3 : 1 flow ratio of the two phases. At the outlet of the microreactor, an aqueous phase containing the enzyme and coenzyme was directed in to a second microreactor, where coenzyme is regenerated. As mentioned earlier, the coenzyme regeneration was catalysed by the same enzyme (ADH) that catalysed the main reaction, in order to make the process more sustainable. Acetaldehyde dissolved in an excess of buffer was used as the substrate for the coenzyme regeneration and fed in to the second microreactor at a flow rate of Φ = 10 mm³ min⁻¹, to obtain a stable and parallel flow and separation of the phases at the exit of the second microreactor. After regeneration, the regenerated coenzyme was reused by recirculating it in to the first microreactor where the oxidation step was continuously performed. The results obtained are presented in Fig. 6. As can be seen, a maximal conversion of 19.5% was achieved. During the first 4 hours, the system was stable with slight decrease in conversion. After that time, the conversion decreased significantly, and after three days it was only 1%. The main reasons why the reaction stopped after three days are the buffer and the presence of acetaldehyde in the reaction mixture. Lee and Whitesides³² determined that for oxidation reactions, such as the one performed in this paper, higher pH values (pH ≈ 9) are required for the maximal enzyme activity. In addition, higher pH values are needed to shift the reaction equilibrium to favor hexanal production. Therefore, 75 mmol dm⁻³ glycine-pyrophosphate buffer (pH = 9; T = 25 °C) was used in all experiments. Before performing the experiments in a microreactor, the stability of the suspended enzyme was monitored dynamically in buffer (data not shown) at different concentrations of acetaldehyde – a known inhibitor of ADH. The initial tested concentrations of acetaldehyde were 100 mmol dm⁻³, 50 mmol dm⁻³, 25 mmol dm⁻³, 12.5 mmol dm⁻³, 6.25 mmol dm⁻³, 3.75 mmol dm⁻³, and 0 mmol dm⁻³. The results showed that the enzyme activity decreases by approximately 90% of the initial value during the three days, with a deactivation constant of k_d = 0.12029 ± 0.0001 h⁻¹, and that an initial acetaldehyde concentration higher than 12.5 mmol dm⁻³ had a high inhibitory effect on ADH activity. The initial acetaldehyde concentration that was used in the integrated systems (c_{i,acetaldehyde} = 12.5 mmol dm⁻³) also had a long-term negative effect on the enzyme. A combination of the buffer and the acetaldehyde negative effect caused the reaction to stop after three days.

Fig. 6 Hexanol oxidation in the integrated system with two microreactors connected in series (○, first system), and in a microreactor system with one microreactor (●, third system).

However, in the proposed system, the regenerated coenzyme was reused by recirculation in to the chip, where the oxidation step was continuously performed for three days without the need to add fresh coenzyme.

The previously obtained results revealed that higher conversions¹⁶ of hexanol to hexanal were achieved in a microreactor equipped with micromixers. Also, when the microreactor chip with the micromixers and the microreactor with the smooth microchannel were connected in series it was possible to enhance the mass transfer and generate separate flows at the exit of microreactor.³³ Based on this, a second microreactor system was developed (Fig. 2b). The system was very similar to the first one. The only modification concerned the microreactor where the hexanol oxidation takes place. Namely, one microreactor was replaced by two microreactors connected in series. As mentioned in the methods section, the first chip was a microreactor equipped with micromixers with two “Y-shaped” inlets and one outlet. The outlet of the first chip was then fed as the inlet stream of the second microreactor with one inlet and two “Y-shaped” outlets. The outlet of the second microreactor, e.g. the aqueous phase containing enzyme and coenzyme, was directed in to a third microreactor chip, where the coenzyme regeneration takes place. The flow rates and substrate concentrations were the same as in the first system. Unfortunately, despite our best efforts, it was not possible to achieve a stable flow with the interphase positioned exactly in the middle of the second microchannel, proving that this developed system was not sustainable for the development of an integrated system for hexanol oxidation.

A third microreactor system was developed in order to make the integrated process as simple as possible (Fig. 2c). Only one microreactor chip with smooth walls and two “Y-shaped” inlets and outlets was used. Hexanol and acetaldehyde dissolved in hexane were fed in to a microreactor as one inlet stream, while the mixture composed of suspended enzyme and coenzyme dissolved in aqueous buffer was fed in to a microreactor as the second inlet stream. At the outlet of the microreactor, the aqueous phase containing the enzyme and coenzyme was recirculated and fed in to a microreactor as the second inlet stream. The obtained results are presented in Fig. 6. The maximal conversion obtained at the beginning of the experiment was 17.1%. Compared to the first microreactor system, the conversion decreased rapidly. The reason for this is that enzyme deactivation caused by its contact with the organic phase was intensified, compared with the first microreactor system.

Comparing the results obtained with the results obtained from the batch system^26,31,34 (V = 20 cm³) for the same initial concentrations of substrate, an improvement in hexanol conversion (hexanal production) was achieved in all proposed systems. Vrsalović Presečki et al.²⁶ reported that the maximal conversion obtained in a batch reactor, where the coenzyme was regenerated by acetaldehyde reduction, was 11.1% (t = 20 min). Additionally, they described a significant inhibition with acetaldehyde in their system. They concluded that acetaldehyde had an extremely strong inhibitory effect on the rate of the main reaction, causing the equilibrium in this system to not be significantly shifted in the direction of hexanol oxidation.

In one of our previous experiments,¹⁹ acetaldehyde inhibition was also monitored and an approximately 20-fold lower inhibition effect of acetaldehyde was noticed in the microreactor experiments, compared to the results obtained in a batch reactor.

Taking all the results in to consideration, we believe that with some additional optimization, and further production cost projections, microreactors could serve as the next-generation production process for not only hexanal production but also for the fast and efficient production of different compounds.

4. Conclusion

In this study, microreactor technology has been demonstrated to be a good alternative to classical production. This is the first time that the production of hexanal has been performed in an integrated multiple microreactor system. Despite the fact that system was not stable for more than the days, two out of the three tested systems demonstrated promising results. Additional system optimizations, e.g. increasing the enzyme operational stability and better product separation and purification, are necessary for the possible application of hexanal as a “green note” component safe as a food additive.

List of symbols and abbreviations

c	Concentration (mmol dm^−3)
T	Temperature (°C)
t	Time (h)
V	Reactor volume (mm^3)
X	Conversion (%)
V.A.	Volumetric activity (U mg^−1or mg^−1)
kd	Deactivation constant (h^−1)
γ	Mass concentration (g cm^−3)
λ	Wavelength (nm)
τ	Residence time (s)
Φ	Flow rate (mm^3 min^−1)

Abbreviations

ADH	Alcohol dehydrogenase
i	Inlet
NAD^+	Nicotinamide adenine dinucleotide
NADH	Nicotinamide adenine dinucleotide hydrate

Acknowledgements

This study was financially supported by the University of Zagreb short-term financial scientific research support under the title: “Application of Chemical Engineering Methodology in Biocatalytic Process Development”.

References

1. B. L. Muller, C. Dean and I. M. Whitehead, Bioflavour, 1995, 95, 14–17.
2. C. Fabrellas, L. Matia and F. Ventura, Water Sci. Technol., 2004, 49, 267–272.
3. G. A. Burdock, Fenaroli's Handbook of Flavor Ingredients, CRC Press LLC, Boca Raton, Florida, 1995.
4. I. M. Whitehead, B. L. Muller and L. Dean, Cereal Foods World, 1995, 40, 193–197.
5. J. S. Márczy, A. S. Németh, Z. Samu, Á. Háger-Veress and B. Szajáni, Biotechnol. Lett., 2002, 24, 1673–1675.
6. P. Brunerie and Y. Koziet, US Patent, 5, 620,879, 1997.
7. M. P. Santiago-Gómez, H. T. Thanh, J. De Coninck, R. Cachon, S. Kermasha, J. M. Belin, P. Gervais and F. Husson, Process Biochem., 2009, 44, 1013–1018.
8. R. G. Berger, A. Kler and F. Drawer, Plant Cell, Tissue Organ Cult., 1987, 8, 147–151.
9. S. R. Chou and C. K. Chin, Lipids in food flavors, American Chemical Society, Washington, 1994, pp. 282–291.
10. M. Karra-Chaabouni, S. Pulvin, A. Meziani, D. Thomas, D. Touraud and W. Kunz, Biotechnol. Bioeng., 2003, 81, 27–32.
11. R. Gandolfi, N. Ferrarra and F. Molinari, Tetrahedron Lett., 2001, 42, 513–514.
12. M. Gargouri, B. N. Akacha and M. D. Legory, Appl. Biochem. Biotechnol., 2004, 119, 171–180.
13. F. Schade, J. E. Thompson and R. L. Lagge, Biotechnol. Bioeng., 2003, 84, 265–273.
14. W. R. Kenealy, Y. Cao and P. J. Weimer, Appl. Biochem. Biotechnol., 1995, 44, 507–513.
15. S. M. Mugo and K. Ayton, J. Mol. Catal. B: Enzym., 2010, 67, 202–207.
16. A. Šalić, A. Tušek, Ž. Kurtanjek and B. Zelić, Biotechnology and Bioprocess Engineering, 2011, 16, 495–504.
17. A. Tušek, A. Šalić, Ž. Kurtanjek and B. Zelić, Eng. Life Sci., 2012, 12, 49–56.
18. W. Ehrfeld, V. Hessel and H. Löwe, Microreactors: New Technology for Modern Chemistry, Wiley-VCH, Weinheim, 2000.
19. A. Šalić, M. Ivanković, E. Ferk and B. Zelić, J. Chem. Technol. Biotechnol., 2012, 88, 1721–1729.
20. A. Šalić, M. Ivanković, E. Ferk, M. Stuparić, M. Tišma and B. Zelić, Chemical Engineering Transactions, 2012, 27, 193–198.
21. A. Šalić, P. Faletar and B. Zelić, Biochem. Eng. J., 2013, 77, 88–96.
22. A. Šalić, K. Pindrić, G. Hojnik Podrepšek, M. Leitgeb and B. Zelić, Green Process. Synth., 2013, 2, 569–578.
23. M. D. Legoy, H. S. Kim and D. Thomas, Pheromone Biochemistry, 1985, 20, 145–148.
24. S. S. Godbole, S. F. D'Souza and G. B. Nadkarni, Enzyme Microb. Technol., 1983, 5, 125–128.
25. S. Suye, K. Kamiya, T. Kawamoto and A. Tanaka, Biocatal. Biotransform., 2002, 20, 23–28.
26. A. Vrsalović Presečki, K. Makovšek and Đ. Vasić-Rački, Appl. Biochem. Biotechnol., 2012, 167, 595–611.
27. A. Šalić, K. Pindrić and B. Zelić, Appl. Biochem. Biotechnol., 2013, 171, 2273–2284.
28. P. Žnidaršič-Plazl and I. Plazl, Pheromone Biochemistry, 2009, 44, 1115–1121.
29. N. Harries, J. R. Burns, D. A. Barrow and C. A. Ramshaw, Int. J. Heat Mass Transfer, 2003, 46, 3313–3322.
30. H. K. Chenault and G. M. Whitesides, Appl. Biochem. Biotechnol., 1987, 14, 147–197.
31. A. Vrsalović Presečki, Study of fumarase and alcohol dehydrogenase in the biotransformations, PhD Thesis, University of Zagreb, Faculty of Chemical Engineering and Technology, Croatia, 2006.
32. L. G. Lee and G. M. Whitesides, J. Am. Chem. Soc., 1985, 107, 6999–7008.
33. A. Šalić, A. Tušek, D. Fabek, I. Rukavina and B. Zelić, Food Technol. Biotechnol., 2011, 49, 495–501.
34. A. Vrsalović Presečki and Đ. Vasić-Rački, Process Biochem., 2009, 44, 54–61.

---