Industrial & Engineering Chemistry Research - ACS Publications


Industrial & Engineering Chemistry Research - ACS Publicationshttps://pubs.acs.org/doi/abs/10.1021/ie970199xSimilarb...

0 downloads 135 Views 240KB Size

4940

Ind. Eng. Chem. Res. 1997, 36, 4940-4948

Fractionation of a Peel Oil Key Mixture by Supercritical CO2 in a Continuous Tower Ernesto Reverchon,* Achille Marciano, and Massimo Poletto Dipartimento di Ingegneria Chimica e Alimentare, Universita` di Salerno, Via Ponte Don Melillo, I-84084 Fisciano (SA), Italy

The fractionation of peel oils with supercritical carbon dioxide was studied using a mixture of four key compounds. Experiments were performed on a 2-m column operated counter currently. The operating pressure and temperature ranged between 75 and 90 bar and between 40 and 80 °C, respectively. Solvent-to-feed ratios of 60, 80, and 120 were used. The total and partial extract refluxes were tested in order to evaluate the process feasibility. The effects of different feed insertion points and column packings were also tested. Experimental results indicate that fractionation can be successfully obtained between 75 and 80 bar and between 50 and 80 °C. In general, an increase of solubility corresponds to a decrease of selectivity, and thus, optimization of the separation is required. Experiments also indicate that temperature helps in separation and, furthermore, increases the recovery of oxygenated compounds. The upper limit to the operating temperature is given, however, by the thermal stability of the product. The total and partial refluxes of the extract at the column top show a definitely positive effect on the separation. The theoretical number of mass-transfer units of the apparatus is between 2 and 3, and this finding seems to be confirmed by experimental results. Higher fractionation efficiencies should require a larger number of these units to be obtainable, for example, with taller columns or with a column series. Introduction Citrus peel oils are usually produced by cold pressing of the citrus fruits. They are formed by hundreds of compounds. Among these, the volatile ones can be roughly classified in two main families: hydrocarbon terpenes and oxygenated terpenes, characterized by different chemical structures and molecular weights. Hydrocarbon terpenes contribute to citrus peel oil compositions from about 60% up to 99%, but they give little or no contribution to their fragrance. Moreover, they tend to decompose under the action of oxygen, light, and heat, producing undesired odor changes in these essential oils. For these reasons, some processes have been developed to perform the deterpenation of peel oils, among these vacuum distillation, extraction with ethyl alcohol, and partition between solvents. Some processes based on the use of supercritical CO2 have also been recently developed to provide citrus peel oil fractionation. The selective supercritical desorption from silica gel of the different compound families forming citrus peel oils has been proposed by Cully et al. (1990), Barth et al. (1994), Chouchi et al. (1995, 1996), and Dugo et al. (1996). Particularly, the processes proposed by Barth et al. (1994) and Chouchi et al. (1995, 1996) were performed by desorbing hydrocarbon terpenes and oxygenated terpenes at two different pressures. It also allowed the elimination of allergenic compounds contained in citrus peel oils. Mathematical modeling of this process based on two key compounds has also been proposed (Reverchon, 1997) and extended to citrus peel oils (Reverchon and Iacuzio, 1997). The only limit of the desorption process is that a semibatch operation is performed that could not be the most effective to process very large quantities of citrus peel oil as required by the industry. Fractionation of citrus peel oils was also attempted by means of a batch high-

pressure tower using supercritical CO2 as a processing medium (Sato et al., 1995, 1996). However, this process does not overcome the limitation of noncontinuous operation. Continuous countercurrent processing by supercritical CO2 has also been attempted (Stahl et al., 1987, Perre et al., 1994; Reverchon et al., 1996). The continuous processing allows the treatment of large quantities of citrus peel oil, but has two drawbacks. The first problem is that it is not able to eliminate coumarins and psoralens, which are among the most allergenic compounds in citrus peel oil. They are solid compounds and can precipitate inside the column. This limitation, however, can be overcome by the use of a supercritical desorption step carried out on the bottom product of the oil treated in the continuous tower. A second problem is represented by the difficulty in obtaining the required separation owing to the similarity between the equilibrium solubility of some hydrocarbons and some oxygenated terpenes (Bu¨dich et al., 1996). In any case, a detailed study of all the process conditions required for citrus peel oil deterpenation by a continuous tower has not been performed yet. The scope of this work is, therefore, to analyze the fundamental aspects of the continuous countercurrent processing of citrus peel oils by supercritical CO2. Most of the compounds present in citrus peel oils are limited to trace quantities, and thus, it is not possible to follow their evolution during the fractionation process. Several components, however, have very similar structures, and it is possible to suppose that they behave similarly. We carried out some experiments on a test mixture made of four key components. Experiments were performed by changing the feed insertion point, the operating pressure and temperature, the reflux ratio, the solventto-solute feed ratio, and the column packing. Experimental Section

* Phone: +[39](89)964116. Fax: +[39](89)964057. E-mail: [email protected]. S0888-5885(97)00199-1 CCC: $14.00

Apparatus. A scheme of the experimental apparatus is shown in Figure 1. The column, C, is 1920 mm long © 1997 American Chemical Society

Ind. Eng. Chem. Res., Vol. 36, No. 11, 1997 4941

Figure 1. Experimental apparatus. B, CO2 cylinder; BS, bottom sample; BV, back-pressure valve; C, column; CB, cooling bath; DPI, differential pressure indicator; FM, rotameter; MB, bottom product regulating valve; MT, top product regulating valve; OR, essential oil reservoir; PF, feed pump; PR, extract reflux pump; PS, solvent pump; PIC, column pressure indicator; PIST, top separator pressure indicator; RC, column heater; RF, feed heater; RR, extract reflux heater; RS, solvent heater; RST, top separator heater; SA, atmospheric separator; ST, top separator; TC, temperature controller; TS, top sample; VA, atmospheric separator valve, VF, feed valve; VS, solvent valve; VT, top product valve; VM volumetric flow meter.

and has a i.d. of 17.5 mm. It consists of 5 cylindrical sections (Autoclave Engineers), 305 mm long and with a o.d. of 25.4 mm, connected to each other and to the process lines by 6 similar 4-port elements indicated by Roman numbers in the figure. All parts are made in AISI 316 stainless steel and are designed to withstand up to a maximum internal pressure of 700 bar at 30 °C. The temperature along the column is controlled by five PID controllers (Watlow Model 965), TC in the figure. Each of them reads the temperature from a type J thermocouple inserted in the column through the side of a four-port connection and regulates two electric heaters, RC, one placed on a face of the connection element itself and the other wound around the cylindrical section just below it. The whole column is thermally insulated by ceramic cloths (insulation is not shown in the figure). Two different column packings were tested: (1) glass Raschig rings of 5-mm nominal size, 1050 m-1 specific surface, and 0.66 voidage degree (); (2) steel packings of 5-mm nominal size 1600 m-1 specific surface, and 0.9 voidage degree. The solvent is fed to the column by a high-pressure diaphragm pump (Milton Roy Model Milroyal B), PS,

that can deliver CO2 flow rates up to 12 kg h-1 and that is provided with a cooled head. The CO2 temperature at the column inlet is controlled by a PID controller, similar to those used for the column, that actuates the electric heater, RS wound around the inlet line. Similar temperature-controlling techniques are also adopted for the liquid feed (heater RF) and the raffinate reflux (RR). The oil mixture is withdrawn directly from a reservoir, OR, and fed to the column by a piston pump (Milton Roy Model Minipump), PF. The use of the four-port connections between the cylindrical sections of the column allows us to place the feed to the extraction column at different levels, which can be identified by the Roman numbering of the connections. The stream exiting from the top of the column is heated to 80-90 °C before being depressurized to 20 bar by a micrometering valve, MT. In this valve, it is cooled by a strong Joule-Thomson effect, and then it is fed to a separator kept at a temperature around 0 °C by another controller acting on RST. The pressure in the separator is regulated by a backpressure valve (Tescom Model 26-172524), BV, that depressurizes CO2 to atmospheric pressure where a second separator, SA, collects a minor part of the most volatile extract. Before the vent, a rotameter, FM, and a dry test meter, VM, measure the CO2 flow rate and the total quantity of solvent used, respectively. The extract withdrawn from the bottom of the separator (ST) is partially collected and partially recycled to the top of the fractionation column by a piston pump, PR, similar to that used for the liquid feed (Milton Roy Model Minipump). Extract and raffinate samples were weighted, and most of them were analyzed by a gas chromatograph (Varian Model 340) connected to a mass spectrometric ion trap detector (GC-MS, Finnigan, ITS Magnum). The product flow rates were calculated by weighing the samples collected at the top and bottom of the column at fixed time intervals. Materials. In Table 1, we reported the composition of the mixture used in this work with some properties of the key compounds and their typical concentrations in some peel oils. The composition of the mixture has been determined by the following considerations. Limonene is a hydrocarbon terpene, and it is the predominant compound in all peel oils, with concentrations between 30% and 80%. Therefore, it was chosen as the most abundant key compound with a concentration of 60% by weight. Linalool is one of the most representative among the oxygenated compounds, and therefore, its concentration was set at 20% by weight, representing the second most abundant component in the key mixture. γ-Terpinene has a molecular weight similar to limonene but is a hydrocarbon terpene which has a volatility very near to oxygenated compounds. Therefore, its presence in the mixture allows us to indicate the effectiveness of separation for those compounds that are more difficult to separate than limonene. Linalyl acetate is representative of oxygenated compounds with a molecular weight higher than linalool. In particular, it is a valuable compound whose complete recovery in

Table 1. Compositions of the Test Mixture Related to Those of the Main Citrus Peel Oils compd

class

MW

density, kg m-3

model mix, wt %

bergamot, wt %

mandarin, wt %

bigarade, wt %

lime, wt %

limonene γ-terpinene linalool linalyl acetate

more volatile non-oxygenated terpenes non-oxygenated terpenes oxygenated terpenes heavy oxygenated terpenes

136.23 136.23 154.24 196.29

840 849 868 901

60 10 20 10

32 7.5 12 30

66 19 0.1