Analytical and Bioanalytical Chemistry

, Volume 377, Issue 7, pp 1190–1195

Superheated liquids for extraction of solid residues from winemaking processes


  • J. González-Rodríguez
    • R&D DepartmentPérez Barquero
  • P. Pérez-Juan
    • Laboratory and Enologic Research of CastillaLIEC
    • Department of Analytical ChemistryUniversity of Córdoba
Original Paper

DOI: 10.1007/s00216-003-2194-5

Cite this article as:
González-Rodríguez, J., Pérez-Juan, P. & de Castro, M.D.L. Anal Bioanal Chem (2003) 377: 1190. doi:10.1007/s00216-003-2194-5


Solid residues from winemaking processes have been subjected to extraction with superheated water–ethanol mixtures. Identification and characterization of the extracted compounds were achieved by spectrophotometry, gas chromatography with either flame-ionization or mass detectors, and-high performance liquid chromatography with UV detection. Extraction was performed statically with single or repeated cycles. All variables affecting the extraction process have been studied and optimised. The extraction time and temperature were 65 min and 210 °C, respectively. Extracts comprised two phases—an aqueous phase, rich in phenolic compounds, and an oily phase, comprising mainly fatty acids. The method allows manipulation of extract composition by changing the applied pressure, temperature, water-to-ethanol ratio, and pH. The method is faster than traditional extraction procedures for obtaining valuable compounds from these residues.


Solid-liquid extractionLiquid chromatographyGas chromatographyMass spectrometrySuperheated liquidsWinemaking residues


After grape crushing to obtain the must, later subjected to the fermentation process, a residue comprising seeds, peel, and stems from the bunches of grapes is obtained. This residue is used to obtain a low sugar-content liquid when mixed with water, and after fermentation can either be exploited as a raw material to obtain a low-quality vinegar or sold to alcohol factories to obtain ethanol. The residue can also be used as fertilizer, because of its high nutrient content. The composition of this solid residue varies depending on the variety of grapes, but its main constituents are tannins, organic acids, reducing sugars, nitrogen compounds, anthocyanins, wax, inorganic salts, and lipids [1].

The extraction of valuable compounds from this raw material is performed in several ways, depending on the target compound to be extracted. Anthocyanins are extracted by using sulfur dioxide as extractant or by use of sorbent resins [2]; seeds are previously separated and smashed and the oil is extracted by using hexane as solvent [3]; tannins are extracted with hot water and then precipitated with NaCl [4]; simple phenolics (hydroxylated and methoxylated benzoic acids, cinnamic acid and derivatives and catechins) have been extracted by soaking with organic solvents [5, 6].

When extraction is performed in the absence of both light and air, degradation and oxidation processes are significantly reduced in comparison with other extraction techniques, and the extracts obtained are richer in valuable compounds. This is the main advantage of using subcritical or supercritical fluids and one of the reasons for developing extraction processes in this way.

Supercritical and subcritical fluid extraction with carbon dioxide has been proposed and organic modifiers are added to increase the polarity of the fluid for extraction of phenolic compounds [7, 8], but the costs of the methods on the industrial scale are high. The stability of phenolic compounds under subcritical conditions up to 150 °C using methanol as extractant has been studied by Palma et al. [9] to prove the feasibility of extracting phenolic compounds from solid residues from the winemaking process with pressurized liquids. Although the results were satisfactory, the use of this solvent has two main shortcomings—the high cost of methanol and its high toxicity, which involves a subsequent step for removal of the extractant before human use.

Subcritical water extraction of essential oils from aromatic plants is very promising [10, 11]. It has prompted the development of an alternative to conventional methods based on organic solvents. The use of superheated ethanol–water mixtures has previously been checked for the extraction of phenolic compounds from oak wood [12, 13]. The extracts obtained were compared with commercial wood extracts, with excellent agreement concerning composition, and with drastic reduction of both time and costs.

Because of the nature of the compounds that can be found in solid residues from the winemaking process, the use of superheated water or mixtures of water and ethanol could be suitable for the extraction of such compounds. The aim of this work was to study the leaching of these residues, by use of either superheated water or ethanol–water mixtures, on the basis of separation of the extracted species by liquid or gas chromatography or the monitoring of the whole extract by photometric detection.


Apparatus and instruments

The laboratory-made extractor used, shown in Fig. 1, consists of a Knauer (Berlin, Germany) 64 high-pressure pump, used to propel the extractant liquid through the system, a stainless steel cylindrical extraction chamber (150 mm×11 mm i.d., 14 mL internal volume), which was closed with screw caps at both ends to enable its filling with extractant and emptying of the extract. The screw caps also contained stainless steel filters plates (2 mm thick, 6.35 mm i.d.) to ensure the solid residues remained in the extraction chamber. This chamber, together with a stainless steel preheater, was located in a gas chromatograph oven (HP 5720A; Hewlett–Packard, Wilmington, DE, USA) used as heating source capable of working up to 400 °C. A loop made from a 1-m length of stainless-steel tubing and cooled with water at room temperature was used to cool the extract from the oven to a temperature close to 25 °C. A pressure needle valve coupled to the outlet of the cooler and a selecting valve located between the high-pressure pump and the oven allowed flushing of the extract with N2 after extraction.
Fig. 1.

Extraction system. hpp, high pressure pump; er, extractant reservoir; ph, pre-heater; ec, extraction cell; o, oven; c, cooler; v1, selection valve; v2, restriction valve

Liquid chromatographic analysis of the extracts was performed using a modular 1100 Hewlett–Packard (Pittsburgh, PA, USA) liquid chromatograph consisting of a G1311A high-pressure quaternary pump, a G1322A vacuum degasser, a 7725 Rheodyne high-pressure manual injection valve (HPIV), and a G1315A diode-array detector. Separation of the components both of the extracts and of standard solutions was performed on a 250 mm×4.6 mm i.d., 5 μm particle, Hypersil ODS column (Supelco, Bellefonte, PA, USA) protected by a precolumn of the same material.

Gas chromatographic analysis of the extracts was performed using a Varian 3900 gas chromatograph equipped with a Chrompack CP-Wax 57 CB fused-silica capillary column (50 m×0.25 mm i.d., 0.2 μm film thickness) and a flame ionisation detector (FID). Finally, a Saturn 2200 mass spectrometer (Varian) equipped with a Chrompack CP-Sil 8 CB fused-silica capillary column (50 m×0.25 mm i.d.) was used to characterise the compounds. A Cary 50 Conc spectrophotometer from Varian (Mulgrave, Australia), connected to a computer with Cary WinUV v.2.0 (Varian) software for data collection and treatment, was also used for photometric monitoring of the extracts.

All the extracts were centrifuged using a Selecta (Barcelona, Spain) Mixtasel centrifuge and filtered through 0.45 μm Minisart filters from Sartorius (Göttingen, Germany). A vacuum pump (Vac Elut SPS 24, Varian, P.S. Analytical, UK) was used for filtering HPLC solvents. Statistical treatment was performed with Statgraphics Plus 2.1. for Windows.

Reagents and solutions

Ethanol 96% (v/v) PA from Panreac (Barcelona, Spain) and distilled water were used for preparing water–ethanol mixtures, used as extractants, with ethanol contents from 0 to 100%. Vanillin and gallic, protocatechuic, and vanillic acids, used as chromatographic standards, were from Sigma–Aldrich (St Louis, USA). 4-Methyl-2-pentanol (243 mg L−1 in 60:40 water–ethanol) and 2-picolinic acid ethyl ester (150 mg L−1 in 60:40 water–ethanol), used as internal standards, were from Merck (Darmstadt, Germany). Solutions of methanol, acetic acid (both HPLC grade and supplied by Merck), and ultrapure water at pH 3 were used as mobile phases. Methanol and ultrapure water at pH 7.0 were used for conditioning and regeneration of the chromatographic column. Ultrapure water obtained from a Millipore (Bedford, MA, USA) Milli-Q plus system was also used. The nitrogen used to displace the extract from the extraction cell was supplied by Air-Liquide (Paris, France).

Solid residues from the winemaking process (SRWP) were provided by Pérez Barquero Winery (Montilla, Córdoba) from the 2002 vintage.


Subcritical extraction

The sample cell in Fig. 1 was filled with 1.5 g SRWP and two pieces of Albet 235 filter paper were inserted at both ends of the cell to prevent the frit from clogging. After assembly of the extraction cell in the oven, the oven was brought to the working temperature (200 °C) and the cell was pressurised to 40 atm with water–ethanol solution, by maintaining valve v2 closed. Once the system was pressurized valve v1 was closed and static extraction was performed for 65 min. After this, v1 was turned to the open position to allow nitrogen to flow through the cell by opening valve v2; the extract was collected in a vial at room temperature.

Chromatographic separation and detection

For the optimization study 2-μL aliquots of the extract were injected into the chromatograph in the (1:15) split mode. The flow-rate of the carrier gas (helium) was 1.1 mL min−1, the injector and detector temperatures were set at 270 °C, and the oven-temperature program was: 40 °C (5 min), 40 °C to 200 °C (4 ° min−1), 200 °C (40 min).

For characterization of the volatile compounds in the extract an aliquot of 1 μL of extract was injected into the chromatograph–mass spectrometer in splitless mode. The carrier gas (helium) flow-rate was 1 mL min−1, injector and detector temperatures were set at 270 and 300 °C, respectively, and the oven temperature program was: 40 °C (5 min), 40 °C to 200 °C (6 ° min−1), 200 °C (20 min), 200 °C to 275 °C (10 ° min−1).

For characterization of the polyphenols present in the extract, 20 μL of this was injected into the liquid chromatograph. Elution was performed in isocratic mode, at a flow rate of 0.3 mL min−1, with a mobile phase consisting of 10:90 methanol–water adjusted to pH 3 by addition of the appropriate amount of acetic acid. The absorption wavelengths were set at 280 and 310 nm and UV spectra in the range 220–360 nm were also recorded. The chromatographic peaks were identified by comparing their retention times and UV spectra with those from the reference compounds.

Results and discussion

A static approach was tested for the extraction of compounds from SRWP in order to increase sample–extractant contact, thus favouring attainment of the partition equilibrium.

Optimization of variables

The variables affecting subcritical extraction were studied in order to maximize the yield of compounds extracted from SRWP in a time as short as possible. With this aim a multivariate approach was used for optimizing the physical variables affecting the extraction, and a univariate approach was used to study the influence on the extraction of the ethanol percent in the extraction mixture. The amount of sample used (1.5 g) was that necessary to fill the extraction cell. The range over which the variables were studied and the optimum values found are given in Table 1.
Table 1.

Results of the optimization study


Tested range

Optimum value


EtOH (% v/v)







Extraction time (min)



Extraction temperature (°C)



Sample weight (g)





Pressure (atm)


A multivariate approach was used for optimization of extraction time and temperature within the ranges 10–70 min and 80–250 °C, respectively. The pressure was that required to maintain the extractant in liquid state. As the vapour pressures of ethanol and water at 200 °C are 30 and 20 atm, respectively [12], a pressure of 40 atm was chosen in order to guarantee the liquid state of the extractant in all the experiments. The chromatographic areas of three abundant compounds characteristic of this kind of residue (vanillin and gallic, and acetic acids) were the dependent variables. The results from ANOVA studies and the response surfaces for the experimental design obtained were statistically equal in all cases. The signal increased when the time and temperature were increased but with a higher effect from the former. A time of 65 min and a temperature of 210 °C were selected as optimum. Temperatures higher than this were not used, in order to avoid cleavage and oxidation reactions of some of the extracted compounds (Ref. [13] and unpublished work).

Since the P-value in the ANOVA table is lower than 0.05%, there is a statistically significant relationship between the variables at a 95% confidence level. The equation of the model chosen in the optimization is: Area=1658−325T+160t+0.751T2+1.4Tt−1.55t2, where T=temperature and t=time.

The amount of ethanol in the extractant was studied by the univariate method in the range 0–100% (v/v). It was found that maximum extraction of some compounds, e.g. protocatechuic acid, was achieved by increasing the ethanol concentration, as expected, but the concentration of some other compounds such as vanillin and gallic and vanillic acids decreased. Other compounds, such as acetic acid, were not affected by changes in the extractant composition. This behaviour was checked by both liquid chromatography (Fig. 2) and gas chromatography (Fig. 3). On the other hand, photometric monitoring of the extract (Table 2) showed an increase of the absorbance at 280 nm when the ethanol content was increased; this can be attributed to an increase of the polyphenol content of the extract. A value of 40% ethanol was selected as a compromise between the extraction efficiency of all the compounds of interest.
Fig. 2.

HPLC chromatograms of the extracts obtained with different ethanol content in the extractant: (a) 20%, (b) 40%, (c) 80%, (d) 100%. 1, gallic acid; 2, protocatechuic acid; 3, catechin

Fig. 3.

Gas chromatograms of the extracts obtained with different ethanol content in the extractant: (a) 20%, (b) 40%, (c) 80%, (d) 100%. 1, gallic acid; 2, acetic acid; 3, vanillic acid; 4, vanillin; i.s., internal standard

Table 2.

Absorbance at 280 nm for extractants with different ethanol content and of different pH











EtOH (% v/v)












A pH study was also conducted by varying the pH of the extractant from 2 to 10. Figure 4 shows the HPLC chromatograms of two extracts obtained by using water at pH 10 (Fig. 4a) and 2 (Fig. 4b). As expected low pH favours extraction of the target compounds in general and that of polyphenols in particular (Table 2). This was also the behaviour of some of the volatile compounds of the extract as shown by the gas chromatograms (Fig. 5). For example, gallic acids were more efficiently extracted whereas extraction of vanillin and acetic and vanillic acids was not affected for this variable. The different behaviour could depend on the pK of each compound.
Fig. 4.

HPLC chromatograms of the extracts obtained at extractant pH 10 (a) and 2 (b). 1, gallic acid; 2, protocatechuic acid; 3, catechin; 4, vanillic acid; 5, vanillin. Note the difference in the absorbance scales

Fig. 5.

Gas chromatograms of the extracts obtained at extractant pH 3 (a) and 10 (b). 1, gallic acid; 2, acetic acid; 3, vanillic acid; 4, vanillin, i.s., internal standard

Ground and non-ground residues were subjected to extraction. Polyphenol indices and amounts of oil were higher for extracts of the former.

Composition of the extract

The subcritical extracts are composed by two phases—aqueous and oily. From 1.5 g SRWP, 0.195 g oil was obtained under the optimum working conditions; this contained mainly linoleic, estearic, oleic, and palmitic acids and their esters. The aqueous phase was rich in polyphenols, fibre, and proteins. Proteins and fibre precipitate over a period of time after extraction and were removed by filtration.

The volatile compounds in the extract obtained using the optimum working conditions were characterized by GC–MS in TIC (total ion count) mode. The compounds were identified by use of the NIST library in the Varian Saturn format. As can be observed in Fig. 6, most of the compounds in SRWP are volatile phenols and fatty acids. Few differences were found in the volatile fraction when extracts from the seeds or from the overall SRWP were subjected to GC–MS. This is indicative of a low contribution of peel and stems to this fraction.
Fig. 6.

GC–MS chromatograms obtained in TIC mode of the subcritical extract obtained at T=210 °C, P=40 atm, 40% ethanol content, pH 3, and t=65 min. Compounds: 1, 1-hydroxy-2-propanone; 2, 2-hydroxy-2-methylbutanoic acid; 3, hydroxybenzoic acid; 4, furanmethanol; 5, diethyl carbonate; 6, cyclohexanone; 7, γ-ketopimelic acid; 8, di-n-propyl succinate; 9, methylcyclopentenolone; 10, 3-furancarboxilic acid; 11, corilon; 12, butyleneglycol diacetate; 13, aletone; 14, 2-hydroxymethylfuran; 15, glycerol diacetate; 16, maltol; 17, 3,5-dihydroxy-2,5-dihydropyran-4-one.; 18, dihydro-4-hydroxy-2(3H-furanone); 19, 5-hydroxymaltol; 20, pyrocatechol; 21, homocatechol; 22, o-acetyl-para-cresol; 23, levoglucosane; 24, pyrogallol; 25, vanillin; 26, gallic acid; 27, p-tert-butylcatechol; 28, vanillic acid; 29, diethyl phthalate; 30, monosaccharides; 31, homovanillic acid; 32, methyl homovanillate; 33, coniferol; 34, α-resorcylic acid; 35, myristic acid; 36, 1-dodecen-3-ol; 37, 2-hydroxy-5-methylphthaldehyde; 38, ethyl vanillate; 39, palmitic acid; 40, ethyl palmitate; 41, linoleic acid; 42, oleic acid; 43, estearic acid; 44, ethyl linoleate; 45, ethyl oleate; 46, ethyl estearate

Exhaustive extraction

A study of exhaustive extraction of the compounds under study in the SRWP sample was performed. Table 3 shows the values for the total polyphenols index and the peak areas from the HPLC chromatograms, referred to the internal standard (picolinic acid ethyl ester), for two of the most abundant compounds (gallic acid and vanillin). The extractions were performed in three cycles using ethanol–water mixtures with an ethanol content of 10, 30, or 50%. The extraction percentages for the two compounds ranged between 55–65% for the first cycle, 15–20% for the second, and 5–15% for the third.
Table 3.

Results from HPLC analysis with UV detection and from total polyphenols index monitoring, as a function of the ethanol percentage in water, after exhaustive extraction of polyphenols from solid residues from the winemaking process by repeated cycles

Water/ethanol (% v/v)

Cycle 1a

Cycle 2a

Cycle 3a








































aReferred to picolinic acid ethyl ester used as internal standard (mg L)−1; G=gallic acid, V=vanillin, TPI=total polyphenol index

Changes of extract composition

As previously reported by the authors, the proposed extraction method allows alteration of the composition of the extract by changing extraction conditions such as temperature, extraction time, and, mainly, ethanol percentage and pH of the extractant. These changes allowed the extracts to be enriched in some of the compounds.


The objective of this work was to check the capacity of subcritical ethanol–water for extraction of valuable compounds from the solid residue from winemaking processes and to characterize the extraction process for use as a clean, cheap, and fast alternative to current extraction techniques. Subcritical extraction has a number of advantages compared with conventional alternatives—significant saving of organic solvents, possibility of manipulating the composition of the extracts by changing the extraction conditions, lower toxicity, lower costs, safer extraction processes, and simpler technology.

Results from analysis of the extracts by GC–MS, HPLC with UV detection, and GC–FID lead to the conclusion that some valuable compounds can be obtained from the extracts in high yield. The studies performed have shown that most of the abundant volatile and non-volatile compounds can be extracted with only two extraction cycles with efficiencies ranging from 75 to 95%, which makes industrial exploitation feasible.


The Spanish Comisión Interministerial de Ciencia y Tecnología (CICyT) is thanked for financial support (Project AGL2000–0321-P4–03).

Copyright information

© Springer-Verlag 2003