Supercritical carbon dioxide separation of fish oil ethyl esters by means of a continuous countercurrent process with an internal reflux

https://doi.org/10.1016/j.supflu.2013.12.003Get rights and content

Highlights

  • A continuous countercurrent process for the production of omega-3 oil is presented.

  • A methodology for the design of a process with internal reflux is developed.

  • The internal reflux is generated by a thermal gradient at the top stage.

  • The relationships between N, S/F, and the top stage temperature T1 are determined.

  • The internal reflux process provides comparable or better results than external.

Abstract

The continuous countercurrent fractionation of fish oil ethyl esters using supercritical carbon dioxide is studied, modelling a process with internal reflux generated by a thermal gradient at the top stage. A methodology for process design is proposed and applied to determine the relationships between the temperature at the top stage (T1), the number of theoretical stages (N), and the solvent to feed ratio (S/F), with the aim of providing a quantitative comparison with the external reflux process. The internal reflux process is viable and, for stated process specifications (mass fraction and recovery of C20 + C22 ethyl esters of 95%), provides comparable or better results than the external reflux process. For example, operating at 13.3 MPa and 50 °C, and keeping T1 in the range (66–70) °C, the specifications are attained with N and S/F in the range 16–30 and 88–120, respectively.

Introduction

In the last two decades, the production of oils rich in polyunsaturated fatty acids (PUFA) of the omega-3 series has gained increasing attention, due to valuable applications of these compounds in the nutraceutical and pharmaceutical industry. Many anthropological, epidemiological, clinical and biochemical studies emphasize the nutritional value of omega-3 PUFA in the prevention of several diseases and indicate that typical Western diets are well below adequate daily intakes of these compounds [1], [2], [3]. In particular, the nutritional and pharmacological value of eicosapentaenoic acid (EPA, C20:5) and docosahexaenoic acid (DHA, C22:6) in the prevention of cardiovascular diseases [4], [5] and the reduction of hypertriglyceridemia [5] has been well ascertained. As a result of these issues, many nutraceutical products based on oils with EPA + DHA mass fraction from 30% to 80% are currently marketed as dietary supplements. Furthermore, mixtures with EPA and DHA (in the form of ethyl esters) over 80%, with stated values of EPA/DHA ratio (typically in the range 1–1.6) are recognized by several Pharmacopoeias as Active Pharmaceutical Ingredient (API) against hypertriglyceridemia and myocardial infarction.

The raw material for the production of omega-3 oils is typically represented by fish oils. Most of the fatty acid chains in fish oil triglycerides have a number of carbon atoms ranging from 14 to 22, being PUFA more abundant in the longest chains. By simply selecting fish naturally rich in omega-3 (e.g., sardine, anchovies, mackerel, herring, menhaden, etc.), crude oil containing EPA + DHA from 10% to 25% can be obtained [6]. Since a considerable enrichment in a specific fatty acid cannot be achieved on a triglyceride feedstock, fractionation processes are usually performed on fish oil ethyl esters (FOEE), preliminary obtained by transesterification of fish oils.

In order to increase EPA and DHA concentration starting from FOEE, the shortest fatty acid chains (C14–C16–C18) must be separated from the longest (C20–C22), as well as saturated and monounsaturated esters from polyunsaturated ones. Separation on the basis of the degree of unsaturation is typically attained by means of crystallization processes, which may be performed in conventional solvents (e.g., methanol, acetone) or in the urea-adduction process [7]. With regard to the separation on the basis of chain length, short alkyl esters are classically separated by distillation processes. However, when dealing with the long polyunsaturated fatty acid chains of fish oils, the conventional vacuum distillation is not feasible, because the high temperatures required (well above 200 °C) unacceptably degrade PUFA [7], [8].

Until now, only two processes have proven to be feasible for FOEE fractionation on chain length basis: short path distillation and supercritical fluid extraction. Short path distillation is a particular kind of distillation, performed at a pressure much lower (0.1–100 Pa) than conventional vacuum distillation, thus with lower operating temperatures and with very short residence time. This process is currently applied in some industries for producing omega-3 oils with EPA + DHA mass fractions typically around 50–60%. However, even operating at extremely low pressures (0.1 Pa), required temperatures are moderately high (in the range 140–170 °C) and process selectivity is relatively low. In addition, to achieve further increase of EPA + DHA mass fractions with this technology more passages are required, causing a substantial reduction in recovery and an increased residence time of EPA and DHA at the process temperatures [9].

A potential alternative to the short path distillation process is the fractionation using supercritical carbon dioxide (SCCO2). A number of seminal studies dealing with semicontinuous processes demonstrated that this solvent is capable of separating FOEE at temperatures below 100 °C and pressures below 20 MPa, showing a selectivity that mainly depends on the length of the carbon chain, being the short chain esters preferentially solubilized. In these processes, a batch of oil is charged in an extraction vessel and SCCO2 is then passed through the charge. Since the solvent extracts preferentially the short chain esters, the EPA and DHA esters are concentrated in the raffinate oil remaining in the extraction vessel [7], [10], [11], [12], [13]. In order to improve recovery and concentration of EPA and DHA esters, the use of a rectification column combined to the extraction vessel was proposed. In these studies, the reflux in the rectification column was generated by means of a temperature increase at the top of the column (or along the column), which causes re-condensation of part of the extracted oil (retrograde condensation) [11], [12], [13]. Although these original processes allow the production of fractions highly concentrated (mass fractions up to 95%) in specific chain length classes, they require very high solvent to feed ratios (from 200 to 400) [13]. More recently, Riha and Brunner [14] showed that it is possible to obtain C20 and C22 ethyl esters (EE-C20 and EE-C22) at mass fraction above 95%, together with 95% recovery, operating a continuous countercurrent process in a column provided with external reflux of extract and using reasonable solvent to feed ratios (in the range 60–130). This process was modelled and simulated by Gironi and Maschietti [15], who obtained a good agreement with experimental data using a thermodynamic model based on the Peng–Robinson Equation of State (PREOS). In the same work, relationships between solvent to feed ratio (S/F), number of theoretical stages (N), and external reflux ratio were investigated in order to find out optimal operating conditions.

An alternative to external reflux in countercurrent fractionation processes using SCCO2 is the application of a thermal gradient along the column, leading to an internal reflux [16]. The internal reflux can be generated if the solubility of the mixture to be fractionated in SCCO2 exhibits a substantial variation with temperature. Focusing on continuous processes, experimental data on the fractionation of liquid mixtures using SCCO2 and exploiting the internal reflux are reported in a few studies, such as in the case of the refining of lampante olive oil [17], the recovery of squalene from vegetable oil deodorizer distillates [18], [19], and the deterpenation of citrus oils [20]. In these studies, a positive effect of the internal reflux on the separation process efficiency is reported but no process modelling is provided. In addition, to our knowledge, there are no modelling studies in the literature where the relationships between the thermal gradient, generating the internal reflux, and the other process parameters (e.g., N, S/F) are studied. In particular, there is a lack of research works where the internal and external reflux processes are quantitatively compared with respect to stated process specifications, thus allowing the potential for application of the internal reflux process to be assessed.

In the present work, the continuous countercurrent separation of FOEE using SCCO2 is studied modelling a process which exploits retrograde condensation to generate an internal reflux in the column. In particular, the reflux for the enriching section is generated operating the top stage at a temperature (T1) higher than the temperature of the body of the column (T), which is isothermal. The relationship between the main process parameters, such as S/F, N, T and T1, is investigated, allowing a quantitative comparison with the external reflux process.

Section snippets

Thermodynamics of the system CO2–FOEE

Experimental data on phase equilibria of binary systems composed of carbon dioxide (CO2) and long chain ethyl esters at high pressures are reported in several literature works [21], [22], [23], [24], [25], [26]. Some experimental data are also available on synthetic multicomponent mixtures of relevant FOEE [22], [27] or natural mixtures of FOEE [7], [15], [22], [28], [29], [30], [31], [32]. When thermodynamic modelling is available, models based on cubic equations of state are selected in most

Simulations of the internal reflux process

A flow scheme describing the continuous process investigated in this work by means of process simulations is shown in Fig. 2. The feed oil (F) is introduced at an intermediate location of a multistage separation column, whereas the solvent (S) is introduced at the bottom. Because of density differences between the solvent rich and the ester rich streams, the former flows upward inside the column while the latter flows downward. Since the solvent preferentially extracts lighter ethyl esters, the

Simulations of columns with 5 theoretical stages

A first set of simulations on the internal reflux process was carried out for a column with 4 isothermal stages (the isothermal section), operating at temperature T, and the top stage operating at T1, being T1 > T. The temperature of the isothermal section was fixed at 42 °C, 50 °C, and 60 °C, whereas the pressure was fixed at 10.1 MPa, 12.2 MPa, and 14.2 MPa, respectively. The process was studied for three different values of the solvent to feed ratio: 60, 100, and 140. The results of the process

Conclusions

A process design methodology is presented for the supercritical fluid fractionation of liquid mixtures in countercurrent columns, where internal reflux is generated by a thermal gradient at the top stage. In the proposed methodology, the process design is carried out using the temperature at the top stage (T1) in the same fashion as the external reflux ratio (L0/D), with respect to the number of theoretical stages (N) and the solvent to feed ratio (S/F).

The study carried out on the

References (41)

  • C. Crampon et al.

    High pressure apparatus for phase equilibria studies: solubility of fatty acid esters in supercritical CO2

    J. Supercritical Fluids

    (1999)
  • J.N. Jaubert et al.

    Phase equilibria measurements and modelling of EPA and DHA ethyl esters in supercritical carbon dioxide

    J. Supercritical Fluids

    (2001)
  • C.J. Chang et al.

    Vapor–liquid equilibria and densities of CO2 and four unsaturated fatty acid esters at elevated pressures

    Fluid Phase Equilibria

    (2005)
  • A. Staby et al.

    Phase equilibria of fish oil fatty acid ethyl esters and sub- and supercritical CO2

    Fluid Phase Equilibria

    (1993)
  • C. Borch-Jensen et al.

    Phase equilibria of long-chain polyunsaturated fish oil fatty acid ethyl esters and carbon dioxide, ethane, or ethylene at reduced gas temperatures of 1.03 and 1.13

    Fluid Phase Equilibria

    (1999)
  • V. Riha et al.

    Phase equilibrium of fish oil ethyl esters with supercritical carbon dioxide

    J. Supercritical Fluids

    (1999)
  • F. Gironi et al.

    High-pressure gas–liquid equilibrium measurements by means of a double-chamber recirculation apparatus. Data on the system carbon dioxide–limonene at 50 and 70 °C

    J. Supercritical Fluids

    (2010)
  • F. Gironi et al.

    High-pressure gas–liquid equilibrium of the system carbon dioxide–citral at 50 and 70 °C

    J. Supercritical Fluids

    (2011)
  • M. Maschietti

    High-pressure gas–liquid equilibrium of the system carbon dioxide–β-caryophyllene at 50 and 70 °C

    J. Supercritical Fluids

    (2011)
  • F. Gironi et al.

    Phase equilibrium of the system supercritical carbon dioxide–lemon essential oil: new experimental data and thermodynamic modelling

    J. Supercritical Fluids

    (2012)
  • Cited by (13)

    • Supercritical fluid adsorption of natural extracts: Technical, practical, and theoretical aspects

      2022, Journal of CO2 Utilization
      Citation Excerpt :

      Based on these drawbacks, further operations are required to concentrate the target compounds in the extract or to remove the undesirable ones, thus providing products with well-defined composition and functionalities. To achieve this goal, techniques such as fractionation by pressure reduction [12,13], separation in countercurrent column [14,15], adsorption processes [16,17], and fractionation by antisolvent extraction [18,19] have all been performed using supercritical fluids. Additionally, the application of crude natural extracts often finds inconveniences due to the bioactive compounds’ susceptibility to degradation.

    • Supercritical CO<inf>2</inf> processing of omega-3 polyunsaturated fatty acids – Towards a biorefinery for fish waste valorization

      2021, Journal of Supercritical Fluids
      Citation Excerpt :

      In 2010, Rubio-Rodríguez et al. [11] discussed the prospects of SFF, aiming at production at large scale and the need for phase equilibria data and models to describe the process. These models have been provided in recent years: Maschietti and Pedacchia [138] developed a predictive model for CC-SFF of fish oil FAEEs, taking into account the main operating parameters (p, T, number of theoretical stages, reflux ratio and solvent-to-feed ratio); similarly, Fiori et al. [139] developed an Aspen Plus™ model for the CC-SFF of omega-3 FAEEs from trout by-products, finding optimal operation conditions (T =80 °C, p =19.5 MPa, external reflux ratio = 0.92, and solvent to feed ratio = 63) and estimating operating and investment cost as 2.30–2.50 €/kg concentrate and 2–15 €/kg concentrate, respectively; Pieck et al. [140] have recently developed a simplified equilibrium-stage model for CC-SFF of fish oil FAEEs that successfully correlates the effect of varying solvent-to-feed ratio on the yield and composition of the feed and raffinate, which in turn define the economic viability of the process [140]. Development of SFF modelling and favorable results of economic analysis should motivate the implementation of SFF plants at industrial scale.

    • Separation of eicosapentaenoic acid and docosahexaenoic acid by three-zone simulated moving bed chromatography

      2020, Journal of Chromatography A
      Citation Excerpt :

      Because of their similar chemical structures, as shown in Fig. 1, it is difficult to get pure component from their mixture. At present, the main methods including urea inclusion [19], low temperature crystallization [20], molecular distillation [21], silver ion complexation [22], supercritical fluid extraction [23,24], and lipase [25] can effectively enrich EPA and DHA. However, these measures are difficult to obtain pure component.

    • Supercritical CO<inf>2</inf>-mediated countercurrent separation of essential oil and seed oil

      2015, Journal of Supercritical Fluids
      Citation Excerpt :

      Gironi et al. proposed a SCCO2-countercurrent extraction as a novel method for concentrating the oxygenated compounds by removing the terpenes from lemon oil [17]. This method has been applied to purification of olive oil [18], recovery of squalene from vegetable oil [19], extraction of bergamot essential oil [20], refining of palm oil [21], fractionation of fish oil ester [22]. In this work, in order to develop a process to separate essential oil from the seed oil present in citrus juice processing residue such as yuzu, SCCO2 fractionation of d-limonene and canola oil mixture as a model of valuable components (i.e. essential oil and seed oil, respectively) from the residues was carried out in a continuous countercurrent column.

    • Experimental study on fluid phase equilibria of hydroxyl-terminated perfluoropolyether oligomers and supercritical carbon dioxide

      2015, Journal of Supercritical Fluids
      Citation Excerpt :

      Considering also the direct experimental observation of partial condensation under pressure reduction at constant temperature, it can be concluded that this system exhibits retrograde behaviour of the first kind, with critical points occurring at both pressures and temperatures lower than the cricondenbar [17]. This kind of retrograde behaviour is also implied in other phase equilibrium literature data on SCCO2 + PFPEs [9,11,12], it is in line with the outcomes of experiments of rapid expansion of solutions of perfluoropolyether diamide in supercritical carbon dioxide [18], and it is often encountered in other systems (SCCO2 + solute) for temperatures and pressures not far from the critical point of CO2 [15,19]. The data related to the dew point region are also depicted in Figs. 7–10, showing the solubility of the four ZDOL oligomers in the gas phase (as gram of oligomer per kg of CO2) at 50 °C, 60 °C, 70 °C and 80 °C.

    • Fractionation of aqueous isopropanol mixtures in a countercurrent packed column using supercritical CO<inf>2</inf>

      2015, Journal of Supercritical Fluids
      Citation Excerpt :

      Note that type II diagram of Fig. 4 exhibits a very narrow right side of the biphasic curve that implies very accurate calculations that are out of the scope of this work. Such work was proposed for instance by Maschietti and Pedacchia [41], who compared, thanks to simulation, internal and external reflux configurations in the case of fish oil esters fractionation. This study has confirmed the feasibility of the continuous operation of purification of an aqueous phase containing isopropanol, using supercritical CO2 as the separation agent.

    View all citing articles on Scopus
    View full text