SciELO - Scientific Electronic Library Online

vol.20 issue3A statistical approach to the experimental design of the sulfuric acid leaching of gold-copper oreAdsorption of aromatic amino acids in a fixed bed column author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Brazilian Journal of Chemical Engineering

Print version ISSN 0104-6632On-line version ISSN 1678-4383

Braz. J. Chem. Eng. vol.20 no.3 São Paulo July/Sept. 2003 

Extraction of lapachol from Tabebuia avellanedae wood with supercritical CO2: an alternative to soxhlet extraction?



L.M.VianaI*; M.R.FreitasI; S.V.RodriguesII; W.BaumannIII

IDepartamento de Química Orgânica, Universidade Federal Fluminense, 22.240-090 Niterói - RJ, Brazil, E-mail:
IIDepartamento de Química Analítica, Universidade Federal Fluminense, 22.240-090 Niterói - RJ, Brazil
IIIInstitut für Physikalische Chemie, Universität Mainz, 55099 Mainz, Germany




The solubility of lapachol in supercritical CO2 was determined at 40°C and pressures between 90 and 210 bar. Supercritical fluid extraction of lapachol and some related compounds by CO2 from Tabebuia avellanedae wood is compared to Soxhlet extraction with different solvents. A standard macroscale (100-200 g wood) and a microscale (»10 mg wood) experimental setup are described and their results are compared. The latter involved direct spectrophotometric quantification in a high-pressure autoclave with an integrated optical path and a magnetic stirrer, fitted directly into a commercial spectrophotometer. The relative amount of lapachol extracted by supercritical CO2 at 40°C and 200 bar was about 1.7%, which is similar to the results of Soxhlet extractions. Lower contents of a- and b-lapachone as well as dehydro-a-lapachone are also reported.

Keywords: solubility, supercritical, extractability, CO2, lapachol, lapachone, dehydro-a-lapachone.




Lapachol was first introduced to the scientific community by Paterno (1882), who cites and reviews an article on colouring agents from plants written by Max Siewert and published in a book on the occasion of the World Fair in Buenos Aires in 1876. Regrettably, that original article was not available to the authors of this paper. This compound and related naphthoquinones are of pharmaceutical interest due to their anticancer and antiviral effect, see, for example Lui et al.,1985 and references cited therein. The wood, bark and roots of trees belonging to the Bignoniaceae family contain up to a few percent of lapachol and often a considerable amount of related compounds. Thus, isolation of lapachol from extracts of the wood, bark or roots of trees such as those of the genus Tabebuia might provide an economic source, although its synthesis has already been achieved with a yield of 40% (Sun et al., 1998).

In this article classic Soxhlet wood extracts and wood extracts resulting from a supercritical CO2 extraction (SFE) procedure are compared with respect to content of lapachol and of some related compounds. SFE with supercritical CO2 is highly favourable in such applications, since CO2 is not toxic and even food quality CO2 is relatively inexpensive. Furthermore, SFE is in general a faster process than Soxhlet extraction.


Two wood samples, A and B were studied. Sample A was fine saw flour and sample B, shredded flitters; both were air-dried. Lapachol was bought from Sigma and a- and b-lapachone were kindly supplied by Rosaly Silveira Silva at the Institute of Organic Chemistry, Federal University, Niterói, Brazil for use as reference material for chromatographic and spectroscopic purposes. Dehydro-a-lapachone was supplied by Sigma from their rare chemicals stock. All reference compounds have been checked for purity by high performance liquid chromatography (HPLC). Fig. 1 displays the structure of the compounds.



SFE-quality CO2 was from White Martins, Rio de Janeiro, and all solvents used were HPLC grade.



Soxhlet extraction was carried out using a standard setup. Sample size was between 50 and 200 g. The extracts were shades of brown, which varied according to the type of extraction.

SFE was carried out on two different scales and therefore in two different experiments.

Experiment 1 is described in Fig. 2 and this setup was used for the SFE experiments with about 1 g of wood. Static CO2 was used for 10 min at a pressure of 150 bar and a temperature of 40°C. Continuous extraction into a 250 ml Erlenmeyer flask yielded a fine powdered extract. The colour of the extract was yellowish brown.



Experiment 2 used a laboratory-constructed high-pressure cell (autoclave) which represents a combined micro-extractor/spectral photometer cell (volume of about 1.7 ml) with a built-in cycling pump. It is an improved version of a cell described in earlier work (Vianna-Rodrigues et al., 1998), which was not equipped with a circulating pump. This cell design allows for extraction of mg to mg amounts of material. Wood samples of about 10 mg were used here. The length of the optical path of the cell can be varied from 0 to about 6 mm, and it is constructed so as to fit directly into the spectral photometer used (model Lambda 15, Perkin Elmer, Überlingen, Germany). Fig. 3 shows the cross section of this cell. The cell is equipped with flow-through channels that are temperature-controlled by a water bath thermostat. Thermal isolation is achieved by a polyurethane foam covering the whole cell, which is mounted on a 10mm acrylic glass base.



In experiment 2 the initial pressure, e.g.150 bar, and a temperature of 40°C were adjusted in the cell and absorption was observed, thereby monitoring the ongoing extraction of a known amount of sample. After extraction equilibrium was reached in the closed cell (usually after 10 to 30 min), indicated when a constant absorption value was asymptotically reached, the spectrum was saved and pressure was adjusted to the next higher value, e.g. 180 bar, and so on. The same procedure was followed for pure lapachol samples. If no compounds other than lapachol absorb, the lapachol content of the wood sample can be calculated from the ratio of the sample and the reference absorption, taking into account the known amounts of sample and reference material in the cell. The cell volume cancels out in this procedure!

HPLC was used for the direct separation and quantification of lapachol and related compounds in the wood extracts. The system was modular: a model 2200 pump and a model Lambda 1000 spectral photometer, both from Bischoff Analysentechnik, Leonberg, Germany; a model 8110 injection valve, from Rheodyne, Cotati, California, USA, equipped with a 5 ml injection loop; and a CR-6A integrator from Shimadzu, Kyoto, Japan, started by a reed contact attached to the injector valve. Isocratic separations were carried out in a 250mm x 4.6mm ID column packed by Bischoff with ODS II Hypersil C18 5mm spherical material. An acetic acid buffer (0.25% v/v) -- acetonitrile eluent (50:50, v/v) was used at 2 ml/min flow, as already reported in the literature (Awang et al., 1986) for the chromatographic isolation of lapachol and related compounds from wood extracts.

The reference compounds as well as the dry extracts were dissolved in acetonitrile at concentrations of 5 to 13 mg/ml before being injected into the HPLC. A Sartorius mechanical balance with 10 mg interpolating resolution was used for weighing.

More sophisticated chromatographic methods (Awang et al., 1986; Steinert et al., 1995; Steinert et al., 1996) such as eluent gradients were not used, since additional selectivity comes from appropriately adjusted wavelengths of the detecting photometer in the range of 370 to 530 nm.

Retention times were 3.4 min for a- and b-lapachone, which could only be separated by spectral discrimination under these chromatographic conditions. The retention time for dehydro-a-lapachone was 3.8 min and that for lapachol, 5.1 min. Retention times were stable and reproducible to lower than 0.04 min over two weeks. The ordering of the compounds was the same as that reported for these compounds in Jácome et al. (1999), where extracts from Zeyheria montana M. were studied by HPLC.

Quantification was performed through comparison of the respective peak areas resulting from the extract and reference solutions. In such cases where peaks were not completely baseline separated, the peak height, instead of the peak area was used.

The solubility of lapachol in supercritical CO2 was determined at 40°C and pressures between 90 and 210 bar following a procedure described in (Vianna-Rodrigues et al., 1998).



Solubility of Lapachol in CO2

In order to make sure that determination of the extractability of lapachol by SFE with CO2 is not limited by solubility, the solubility of lapachol was determined in supercritical CO2 at 40°C and pressures between 90 and 210 bar. Fig. 4 shows the results.



The pressure error is ±1.5 bar and the solubility error is ± 5%, mainly due to weighing errors and uncertainty in the determination of absolute cell volume. Solubility is already about 1 g/l at 150 bar and increases up to 1.4 g/l at 210 bar.

Results from Experiment 1

Four different Soxhlet extraction procedures, 1 to 4, (see Table 1), for wood samples A and B were used with different solvents, and one SFE procedure was carried out ,5 (see Table1) with supercritical CO2 at 150 bar and 40°C.



Fig. 5 shows the absorption spectra of the compounds studied in the chromatographic eluent. It reveals that a- and b-lapachone can be independently quantified from the results of two chromatographic runs carried out at two distinct adequate wavelengths.



As an example, Fig. 6 shows the HPLC chromatograms of a CO2 extract obtained at detection wavelengths of 360 nm and 470 nm, as well as that of a methylene chloride extract obtained at 360 nm.



Table 1 shows the results from Soxhlet extraction and SFE, which were achieved for lapachol in experiment 1. The mass of the wood samples ranged from 20 to 130 g. Soxhlet extraction was considered to be finished when the initial characteristic yellow color of the eluent had paled.

The relative standard error for lapachol content in the extracts was about 7%, including all weighing and peak integration errors. From the percentage of extract from the wood samples and using the lapachol content of the extracts, the lapachol content of the wood was estimated and is also shown in Table 1. The error is roughly estimated to be 20%, due to the poor homogeneity of the natural wood samples. For the comparative studies in this work an exact statistical evaluation is not necessary and would only be possible with much larger numbers, n, of Soxhlet extractions. But since the extraction error is much smaller than the error introduced by the poor homogeneity of the wood material, a larger n, although allowing for better statistical characterization, would not yield better results. These values are also shown in Table 1.

Table 2 shows the percentage of a- and b-lapachone and of dehydro-a-lapachone in wood samples A and B, as determined from the same extracts that were used for determination of lapachol, (Table 1) in these wood samples. The same error holds as that for the data in Table 1.




To the authors' knowledge the solubility of lapachol in supercritical CO2 has not been reported to date. Fig. 4 shows that it is higher than 1 g/l at pressures above 160 bar. Therefore, extraction of up to 2mg lapachol in a cell volume of about 2ml would not be solubility limited. This means that the extraction of wood samples of about 10 mg with a lapachol content of a few percent is not solubility limited. Solubility of the other compounds studied in this work was not determined, since the amount of them was very low, so solubility limitation was not expected.

Extraction of lapachol and related compounds by supercritical CO2 has not been described in the literature. Therefore, the results of extraction by supercritical CO2 (SFE) must be compared to those using classic Soxhlet extraction procedures with some eluents of different polarities.

Table 1 shows the results of Soxhlet extractions of lapachol from two wood samples, A and B, and the results using an SFE procedure at 40°C and 200 bar. Sample size was on the order of 100 g.

The first column in Table 1, %(w/w) extract from wood, shows that on average SFE extracts as much material as the Soxhlet extraction procedures do, with the exception of procedure 4 which shows the lowest extraction yield. It must be emphasized here that the extract is a raw material which contains a lot of different compounds in addition to lapachol.

The second column, %(w/w) lapachol in the extract, shows that procedure 4 results in the highest percentage of lapachol in the extract, which is the purest extract, referred to lapachol. This second column also shows that SFE gives a less pure extract, roughly comparable to that from an n-hexane Soxhlet extraction procedure, which is in agreement with what is to be expected from the similarly low polarity of n-hexane and CO2 under the given conditions. The last column in Table 1 is calculated as (% extract) x (% lapachol in the extract) of the results in Table 1, which is equivalent to the % lapachol that could be extracted from wood samples A and B under the cited conditions. It shows that the relative lapachol contents in the two studied wood samples are distinctly different, independent of the extraction procedure: wood sample A has a smaller relative content of lapachol in the wood. This is not what was expected from the different sample structures: wood sample A was fine sawed and wood sample B coarsely sawed, and therefore opposite results were expected. Hence the results are not related to the wood surface accessible during the extraction process. It is also interesting to note that although the purity of the extract resulting from the acid/base extraction, procedure 4, is very good, the total relative amount of lapachol extracted from the wood samples is small, and it is smaller with the wood sample A, apparently due to the low relative amount in the raw extract. SFE yields similarly small relative amount of raw extract and therefore low relative amount of lapachol in the wood.

The results presented in Table 1 for extractions of lapachol from wood samples on the order of 100 g may be compared to those in Table 3, where aliquots of wood sample B of about only 10 mg were extracted by the on-line SFE/quantification procedure using the cell/autoclave described previously (see Fig. 3). In wood sample B 1.7± 01% lapachol was obtained, which is in good agreement with the results from the larger scale SFE with subsequent quantification by HPLC shown in Table 1. In all experiments, extraction efficiency increased with increasing pressure. The values in brackets in Table 3 are given as a representative example which shows that the extraction efficiency increased by about 20% with a pressure increase from 150 to 210 bar. Quantification was not affected by the anomalous absorption band of lapachol, which is reported for low molecular weight alcohols in Portugal et al. (1997).



Table 3 also shows the lapachol content in two lapacho tea samples. The very low content is consistent with the fact that these teas are prepared from the inner bark of Tabebuia avellanedae, and this part of the tree is known to have a low lapachol content (Steinert et al.,1996).

In agreement with what is known from the literature content (Steinert et al.,1996), the content of a- and b- lapachone as well as that of dehydro-a-lapachone in the wood is low compared to lapachol. The results are shown in Table 2. Apparently, Soxhlet extraction with CH2Cl2 (procedure 2) is the most efficient procedure for all three compounds, consistent with the results reported for dehydro-a-lapachone in Steinert et al. (1995). Acid/base extraction (procedure 4) is not adequate, but SFE (procedure 5) extracts reasonable amounts of these compounds from wood sample B. Benzene was tested for wood sample A only and gave good results. In general, the content of these compounds in Tabebuia avellanedae wood is distinctly higher than that reported for Zeyheria montana M. root (Jácome et al., 1999). Although excessive light was avoided as much as possible, the conversion of a-lapachone to dehydro-a-lapachone (Hooker, 1936; Burnett and Thomson, 1967) could have influenced the ratio obtained for these two compounds, thus explaining the relatively large amounts of dehydro-a-lapachone obtained from the extraction with CH2Cl2 or by SFE.



Extraction of lapachol with CO2 by two different SFE procedures was compared to standard Soxhlet extraction procedures under different conditions. The relative amount of lapachol extracted by SFE at 40°C and 200 bar was found to be comparable to that obtained by Soxhlet extraction. The purity of the extracts with respect to lapachol was similar for SFE and n-hexane Soxhlet extraction, but in the same wood samples it was better with the very time-consuming acid/base Soxhlet extraction procedure.

In addition, the extracts were analyzed for content of a- and b- lapachone and of dehydro-a-lapachone. In most cases relative amounts of much less than 0.1% of these compounds were found in the two wood samples studied, which is consistent with what was expected based on the literature (Steinert et al., 1995). It was shown that the extraction yield for these compounds is much larger with the SFE procedure than with the acid/base Soxhlet extraction procedure, in contrast to the results for lapachol. The highest relative amount extracted was obtained with CH2Cl2 by a Soxhlet extraction procedure.

Thus, SFE with CO2 has shown to be a cheap and fast alternative for standard Soxhlet extraction of lapachol and some related compounds from wood samples unless the highest possible extract purity is required for lapachol, thereby necessitating acid/base Soxhlet extraction .



The authors are highly indebted to Luiz Gonsaga and Dr. Frank Neitzel for their technical assistance in some of the extraction experiments. LMV and SVR gratefully acknowledge the support from CNPq (Reg. No. 1738608896966715). WB is grateful to the International Bureau, DLR, Germany, for having covered the travel expenses through another project (BRA 58/98 ENV), and to the Federal University in Niterói for accommodation.



Awang, D.V.C., Kindack, D. and Dawson, B.A., Journal of Chromatography, 368, 439-443 (1986).        [ Links ]

Burnett, A.R. and Thomson, R.H., J. Chem. Soc. C, 1261 (1967).        [ Links ]

Burnett, A.R. and Thomson, R.H., J. Chem. Soc. C, 2100-2104 (1967).        [ Links ]

Girard, M., Kindack, D., Dawson, B.A., Ethier, J.C. and Awang, D.V.C., J. Nat. Prod., 51, 1023-1024 (1988).        [ Links ]

Hooker, S.C., J. Am. Chem. Soc., 58, 1190 (1936).        [ Links ]

Jácome, R.L.R.P., Oliveira, A.B., Raslan, D.S., Müller, A. and Wagner, H., Química Nova, 22, 175-177 (1999).        [ Links ]

Kingston, D.G.I. and Rao, M.M., Plantar Med., 39, 230-231 (1980).         [ Links ]

Lui, C.Y., Arena, A.A., Gyllenhaal, C. and Groves, M.J., Drug Development and Industrial Pharmacy, 11 (9&10), 1763-1779 (1985).        [ Links ]

Patermo, J.E., Gazz. Chim. Italiana, 12, 337 (1882).        [ Links ]

Portugal, S.G.M., Herrera, J.O.M. and Brinn, I.M., Bull. Chem. Soc. Jpn., 70, 2071-2076 (1997).        [ Links ]

Rao, M.M. and Kingston, D.G.I., Planta Med., 45, 600-604 (1980).         [ Links ]Rodrigues, S.V., Nepomuceno, D., Viana, L.M. and Baumann, W., Fresenius J. Anal. Chem., 360, 58 - 61 (1998).         [ Links ]

Steinert, J., Khalaf, H. and Rimpler, M., J. Chromatogr. A, 693, 281-287 (1995).        [ Links ]

Steinert, J., Khalaf, H. and Rimpler, M., J. Chromatogr. A, 723, 206-209 (1996).        [ Links ]

Sun, J.S., Geiser, A.H. and Frydman, B., Tetrahedron Letters, 39, 8221-8224 (1998).         [ Links ]

Wagner, H., Kreher, B. and Lotter, H., Helv. Chim. Acta, 72, 659-667 (1989).        [ Links ]



Received: June 10, 2001
Accepted: February 13, 2003



*To whom correspondece should be addressed

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License