ARTICLE IN PRESS

JOURNAL OF
FOOD COMPOSITION
AND ANALYSIS
Journal of Food Composition and Analysis 16 (2003) 563–573
www.elsevier.com/locate/jfca

Original Article

HPLC–DAD analysis of ketones as their

2,4-dinitrophenylhydrazones in Brazilian
sugar-cane spirits and rum
Daniel R. Cardoso, Sandra M. Bettin, Roni V. Reche,
Benedito S. Lima-Neto, Douglas W. Franco*
* Carlos, Universidade de Sao
Instituto de Qu!ımica de Sao * Paulo, Caixa Postal 780, CEP 13560-970,
* Carlos, SP, Brazil
Sao
Received 21 June 2002; received in revised form 12 February 2003; accepted 17 March 2003

Abstract

Using high-performance liquid chromatography (HPLC) and 2,4-dinitrophenylhydrazine as a derivatiz-

ing reagent, an analytical method has been developed for the analysis of ketones in hydro-alcoholic
matrices, colored or not. The determination is carried out at 365 nm using a UV–Vis diode-array detector
(DAD). For the ketones as their 2,4-dinitrophenylhydrazones, a good separation was achieved with a
Supelco C18 column (25 cm
4.6 mm
5 mm) using a (methanol/acetonitrile)-water elution gradient. The
methodology is simple, rapid, exhibits good reproducibility, and allows the analysis of both ketones and
aldehydes in the same run. The application of this method to 34 sugar-cane spirits and 13 rums allowed the
identification and quantification of acetone, acetophenone, cyclopentanone and 2,3-butanedione, at
detection limits of 10 mg/l (2,3-butanedione) to 20 mg/l (cyclopentanone).
r 2003 Elsevier Ltd. All rights reserved.

Keywords: Liquid chromatography; Ketones; Brazilian cane-sugar spirit; 2,4-dinitrophenylhydrazones

1. Introduction

Although production of the traditional Brazilian sugar-cane spirit called ‘‘cacha@a’’ or

‘‘caninha’’ is around 109 l/year (Drinks International, 1994; Nascimento, Cardoso, Lima-Neto, &
Franco, 1998; Nascimento, Cardoso, De Keukeleire, Lima-Neto & Franco, 2000), very little is
known about its chemical composition.

*Corresponding author. Tel.: +55-16-2739970; fax: +55-16-2739976.

E-mail address: douglas@iqsc.usp.br (D.W. Franco).

0889-1575/03/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0889-1575(03)00061-9
 ARTICLE IN PRESS

564 D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573

Ketones are present in alcoholic beverages in small concentrations, but they are known to be
important for the flavor and aroma of many beverages (Heath & Reineccius, 1986). They are
found in wines, distilled alcoholic beverages and aged spirits (Liebich & Koening, 1970; Piggott,
Sharp, & Duncan, 1989; Puputti & Lethonen, 1986). The origin of the ketones in the sugar-cane
spirit is related to secondary fermentation processes and possible contamination during
production (Coutrim, Nakamura, & Collins, 1993; Fung & Grosjean, 1981; Nascimento et al.,
1998; Rehm & Reed, 1984). In general, ketones are not considered harmful, and the International
Health Organization has not classified ketones as carcinogenic (Campbell, Pfefferkon, &
Rounsaville, 1985; Jenner, Hagan, & Jean, 1994). However, their prolonged inhalation can not
only cause irritation of the mucous membranes, headaches, confusion, and narcotic effects, but
also lead to coma as well (Campbell et al., 1985; Jenner et al., 1994).
As part of our studies (Boscolo, Cardoso, Lima-Neto, & Franco, 2000; Boscolo, Andrade-
Sobrino, Lima-Neto, & Franco, 2002; Furuya et al., 2002; Nascimento, Marques, Lima-Neto,
De Keukeleire, & Franco, 1997; Nascimento et al., 1998; Nascimento et al., 2000) on ‘‘cacha@a’’
chemical composition, aiming to improve its quality control, a methodology has been developed
for ketone analysis in cacha@as. The outlined procedure uses an HPLC technique and 2,4-
dinitrophenylhydrazine as a derivatizing reagent. This analytical procedure has been applied to
ketone analysis in Brazilian cane-sugar spirit and rum samples.

2. Materials and methods

2.1. Reagents

The solvents (methanol, acetonitrile, pentane, dichloromethane and ethanol) were HPLC grade
(Malinckrodt). Based on other findings (Piggott et al., 1989; Puputti et al., 1986) for alcoholic
beverages, the following ketones were selected for this study: acetone, acetophenone,
acetylacetone, 2,3-butanedione, 2,3-butanedione monoxime, 2-butanone, cyclohexanone,
cycloheptanone, cyclohexylethanone, cyclopentanone, 5,5-dimethyl-1,3-cyclohexadione, 2,6-dimethyl-
cyclohexanone, 3-hexanone, 2-hexanone, 4-heptanone, 4-methyl-2-pentanone, 5-methyl-3-hepta-
none, 3-heptanone, 2-heptanone, 2-methylcyclohexanone, 5-methyl-2-hexanone, 5-nonanone,
3-nonanone, 2-nonanone, 2-octanone, 2-pentanone and 6-undecanone. For all the ketone
standards (Merck, Aldrich, L. Light & Co. Ltd. and K&K Laboratories), a purity of 99.5% or
more was certified. The 2,4-dinitrophenylhydrazine (Aldrich) reagent was purified twice by
recrystallization from butanol (Shriner, 1980). Milli-Q water was used in this work (Millipore
Corp., Bedford, MA, USA). The aldehydes used in the experiment for ketone and aldehyde
analysis were previously reported to occur in cacha@as (Nascimento et al., 1997).

2.2. Sampling

The samples were chosen taking into account both the producer’s tradition and consumer
preferences. At least three samples of each brand were collected and assayed.
The rum samples were selected according to their commercial relevance. Also in this case, at
least three bottles of each brand were collected and analyzed. The analytical data reported in this
 ARTICLE IN PRESS

D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573 565

paper for each case are the average values obtained from the analysis of three samples of the same
beverage.
Thirty-four Brazilian cane sugar spirits and 13 rum samples were analyzed:
Brazilian cane-sugar spirits: 51, Terra Brazilis, Capelinha, Capelinha Aged, Cigana, Cigana
Aged, Colonial, Colonial Prata, Delicate, Delicias da Ro@a, Havana, Jacutinga, Jacutinga Aged,
Jamel, Jequety, Jequety Aged, Melissa Aged, Milagre de Minas, Pitu, Pitu Gold, S*ao Saru#e Aged,
Setembrina, Santa In#es, Velha Prov!ıncia, Vila Velha, Vila Velha Aged, Volupia, ! !
Volupia
Aged, Tiquara, Tiquara Aged, Triumpho, Triumpho Aged, Ypioca Prata, Ypioca Gold Aged.
Rums: Aniversa! rio (Venezuela), Appleton Dark (Jamaica), Appleton Estate (Jamaica), Bacardi
(Mexico), Casino (Canada), Captain Morgan Dark (Canada), Havana Aged (Cuba), Montila
(Brazil), Mount Gay (Barbados), Myers’s Original Dark (Jamaica), Negrita (Caribbean), Selecto
(Venezuela), XK (Mexico).

2.3. GC analysis

The original samples were extracted using liquid–liquid extraction. For a typical experiment, a
300 ml aliquot of each beverage was mixed with 50.0 ml of dichloromethane and 100 ml of
pentane. The mixture was kept in a sealed PYREX flask at room temperature (25 C72 C) and
stirred for 48 h. After this time, the phases were separated and the organic phase volume reduced
to around 5.00 ml by evaporation. This liquid was adjusted to a final volume (10.0 ml) by adding a
dichloromethane–pentane (1:2) solution (Moret, 1992).
An HP 5890-A gas chromatograph equipped with a flame ionization detector (FID) and
two capillary columns were tested: HP-FFAP (cross-linked modified polyethylene glycol
esterified phase) (50 m
0.2 mm
0.30 mm) and a HP-5 (5% diphenyl and 95% dimethylpoly-
siloxine) (50 m
0.2 mm
0.33 mm) were used for the direct analysis. The analytical operating
conditions were: a (1:15) split, with hydrogen as carrier gas (flow rate of 2.8 ml/min). The gas
flows for the FID detector were 30 ml/min for hydrogen and 300 ml/min for synthetic air.
The temperatures of the injector and of the detector were set at 250 C. An injection volume
of 1.0 ml was used for the beverage extract. The standard of ketones in a dichloromethane–
pentane (1:2) solution was injected under similar conditions (see Fig. 1). Ketones were
identified by comparing their retention times with those of the standards and by standard
addition.

2.4. Ketone labeling and HPLC analysis

The analysis of ketones using liquid chromatography was performed after labeling with 2,4-
dinitrophenylhydrazine. The 2,4-dinitrophenylhydrazone standards were prepared mixing
A and B solutions. (A): 0.40 g of 2,4-dinitrophenylhydrazine dissolved in 2.00 ml of
H2SO4+3.00 ml of H2O+10.0 ml ethanol; and (B): 0.50 g or 1.00 ml of the ketone standard
dissolved in 20.0 ml ethanol. After this mixing, a precipitate was formed in each case,
isolated through filtration, and dried in vacuum (Nascimento et al., 1997; Papa & Turner,
1972). The elemental analysis (C, H, N) data for these 2,4-dinitrophenylhydrazones are in good
agreement with the calculated values. According to these data, their purity was always better
than 98%.
 ARTICLE IN PRESS

566 D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573

Fig. 1. Gas chromatographic separations of ketone standards (0.1 mg/l) on HP-FFAP column; 1—(acetone);
2—(ethanol, 2-butanona, 5,5-dimethyl-1,3-cyclohexadione, acetylacetone); 3—(2-pentanone); 4—(2,3-butanedione);
5—(4-methyl-2-pentanone); 6—(3-hexanone); 7—(2-hexanone); 8—(4-heptanone); 9—(5-methyl-3-heptanone); 10—
(3-heptanone); 11—(2-heptanone); 12—(5-methyl-2-hexanone); 13—(cyclopentanone); 14—(2-octanone); 15—(cyclo-
hexanone); 16—(2,6-dimethyl-ciclohexanone); 17—(2-methyl-cyclohexanone); 18—(5-nonanone); 19—(3-nonanone);
20—(cyclopentanone); 21–(2-nonanone); 22—(cyclohexyl-ethanone); 23—(acetophenone); 24—(6-undecanone);
25—(2,3-butanedione-monoxime); for chromatographic conditions, see Section 2.

The absorption maximum (lmax ) for the ketones as their 2,4-DNPH derivatives changes from
310 to 380 nm, since their molar absorptivities do not change very much (less than 7%) from their
respective lmax to 365 nm. This wavelength was chosen for their analytical detection and
quantification, except for acetylacetone. Since 1,3-dicarbonyls do not form 2,4-dinitrophenylhy-
drazone derivatives, acetylacetone may react with 2,4-DNPH to form the corresponding pyrazole
derivative (Grosjean, Green, & Grosjean, 1999). Therefore, for acetylacetone, the measurement
must be carried out at lmax ¼ 310 nm.
The ketones and aldehydes in the beverages were transformed into their 2,4-DNPH derivatives
by mixing 1.00 ml of a solution containing 200 mg/100 ml of 2,4-dinitrophenylhydrazine with
1.0 ml of H3PO4, 4.00 ml of the beverage. After 2 h, a 40.0 ml aliquot was withdrawn and analyzed
by the HPLC technique (Coutrim et al., 1993; Fung & Grosjean, 1981; Nascimento et al.,
1997). The HPLC experiments were carried out using Shimadzu model LC-10ADVP
chromatograph, a UV–Vis diode array detector, model SPD-M10A, and a Supelco C18 column
(25 cm
4.6 mm
5 mm). The injection volume was 20.0 ml and the detection was performed at
365 nm. The following (methanol/acetonitrile)-water gradient was used: (methanol/acetoni-
trile)(8:2)-water 60:40 (v/v) isocratic for 9 min (1.00 ml/min), from 60:40 to 95:5 in 16 min
(1.1 ml/min), from 95:5 to 60:40 in 9 min (1.0 ml/min), 60:40 isocratic for 15 min (1.00 ml/min).
For identification and quantification of the components in high concentration, the samples were
appropriately diluted for analysis. The identifications were carried out through relative retention
time and standard addition. These results were confirmed after a comparison between the DAD
absorption spectra of the peak and that of the standard.
A calibration curve was obtained for each detected ketone using standards of 0.5, 1.0, 2.5, 5.0
and 10 mg/l in acetonitrile. The plot of the peak area versus its respective concentration was
analyzed by linear regression (y ¼ a þ b  x). Correlation coefficients were always very close to
unity, see Table 1.
 ARTICLE IN PRESS

D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573 567

Table 1
Calibration curves (y ¼ A þ Bx) for HPLC determination of ketones as their 2,4-dinitrophenylhydrazones
2,4-DNPHos A B r
Acetone 4452.56 53620.83 0.999
Acetophenone 1247.96 15061.85 0.999
2,3-Butanedione 5218.93 36707.93 0.999
Cyclopentanone 3523.59 42108.79 0.999
r2=correlation coefficient.

Fig. 2. Partial HRGC chromatogram of a cacha@a sample on HP-FFAP column; 1—(acetone); 2—(ethanol); 3—
(cyclopentanone); 4—(acetophenone); for chromatographic conditions: see Section 2.

3. Results and discussion

Gas chromatography was the first choice as the technique to develop a rapid method for ketone
analysis, without a labeling procedure. Two types of columns were compared: HP-FFAP and
HP-5. The better resolution was obtained with the HP-FFAP column (see Figs. 1 and 2).
Nevertheless, there were co-elution problems between ethanol and 2-butanone and between
acetylacetone and 5,5-dimethyl-1,3-cyclohexanedione, which persisted despite various efforts to
solve them. These results discouraged further studies to improve the ketone extraction procedure
and recovery coefficients. These previous experiments using known amounts of ketone standards
show that less than 70% of the original compounds were recovered by liquid–liquid extraction.
Thus, studies using liquid–liquid extraction and GC analysis were only carried out at the semi-
quantitative level.
Co-elution and recovery problems were solved using HPLC with 2,4-dinitrophenylhydrazine as
a derivatizing reagent. The most efficient separation of the ketones as their 2,4-DNPH was
obtained with the Supelco C-18 column (25 cm
4.6 mm; 5 mm) using a methanol/acetonitrile-
water elution gradient. No co-elution problem among the ketones and aldehydes as their
 ARTICLE IN PRESS

568 D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573

Fig. 3. Standard liquid chromatographic separations of aldehydes and ketones as their 2,4-DNPH derivatives. a—(2,4-
dinitrophenylhydrazine); 2,4-DNPH derivatives: 1—(acetylacetone); 2—(formaldehyde); 3—(2,3-butanedione);
4—(acetaldehyde); 5—(furfuraldehyde); 6—(acrolein); 7—(acetone); 8—(propionaldehyde); 9—(butyraldehyde); 10—
(valeraldehyde); 11—(cyclopentanone); 12—(heptanaldehyde); 13—(acetophenone); 14—(nonaldehyde); 15—(decan-
aldehyde); For chromatographic conditions, see Section 2.

Fig. 4. Chromatogram for the ketones and aldehydes analysis in a cacha@a sample; a—(2,4-dinitrophenylhidrazine);
2,4-DNPH derivatives: 2—(formaldehyde); 3—(2,3-butanedione); 4—(acetaldehyde); 5—(furfuraldehyde); 6—(acro-
lein); 7—(acetone); 8—(propionaldehyde); 9—(butyraldehyde); 10—(valeraldehyde); 11—(cyclopentanone); 12—
(acetophenone). For chromatographic conditions, see Section 2.

2,4-dinitrophenylhydrazones was observed. Indeed, due to their different retention times, ketones
and aldehydes could be analyzed together in the same run, see Figs. 3 and 4. The spirits aldehyde
profile has been reported previously (Nascimento et al., 1997).
 ARTICLE IN PRESS

D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573 569

Table 2
Quantitative analysis data of ketones in cacha@as (mg/l)
Sample Acetone Acetophenone Cyclopentanone 2,3-Butanedione
1 0.20 ND 1.40 ND
2 0.39 ND 0.12 0.52
3 1.87 ND 1.11 ND
4 0.68 ND 0.11 0.25
5 1.25 ND 0.34 3.74
6 0.35 ND 1.31 7.35
7 ND ND 1.25 2.08
8 0.12 ND 2.87 1.90
9 4.91 ND 0.84 ND
10 1.40 ND 0.39 8.84
11 23.0 ND 1.95 2.81
12 1.28 ND 0.83 0.24
13 ND ND 1.81 1.85
14 1.18 ND ND 9.14
15 ND 1.34 ND 2.59
16 0.51 0.92 0.15 4.42
17 2.27 ND ND 6.64
18 11.1 ND 0.19 4.35
19 13.2 ND ND 2.62
20 14.8 2.74 0.28 4.81
21 3.86 ND ND 2.57
22 2.66 0.19 0.23 8.82
23 0.59 ND ND 5.90
24 0.50 ND 0.20 3.54
25 4.34 ND ND 9.77
26 2.86 0.92 0.15 4.42
27 0.39 ND ND 6.32
28 1.73 1.36 0.83 7.49
29 0.22 ND 0.14 8.07
30 4.28 ND 0.31 0.23
31 0.34 ND ND 1.03
32 0.19 ND 0.15 5.48
33 1.48 ND ND 4.23
34 0.59 ND ND 2.57
Mean 3.31 1.24 1.15 4.34
ND=not detected.

By applying the HPLC methodology to the beverages, the following ketones were identified and
quantified: acetone, acetophenone, cyclopentanone, and 2,3-butanedione. Of course, others
ketones can be present at a concentration below the detection limit. However, the present method
avoids extraction procedures as described in the literature (Liebich & Koening, 1970; Piggott et al.,
1989; Puputti & Lethonen, 1986), and can also be more sensitive with the use of C18 solid-phase
extraction as a pre-concentration technique. The analytical results for the 34 cacha@as and 13
rums are summarized in Tables 2 and 3.
 ARTICLE IN PRESS

570 D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573

Table 3
Quantitative analysis data of ketones in rum (mg/l)
Sample Acetone Acetophenone Cyclopentanone 2,3-Butanedione
1 0.68 ND 0.11 0.77
2 6.07 ND 0.31 1.54
3 2.78 ND 0.11 6.34
4 2.03 ND 0.76 1.79
5 2.54 ND 0.31 0.59
6 0.53 ND 0.77 2.14
7 0.18 ND 0.18 2.85
8 2.15 ND 0.26 1.00
9 0.92 ND 0.16 2.52
10 1.83 ND 0.17 0.35
11 0.70 ND 0.31 0.82
12 2.14 ND 0.80 ND
13 2.35 ND 0.27 0.29
Mean 2.15 0 0.35 1.75
ND=not detected.

Table 4
Reproducibility data for the HPLC analysis of ketones as their 2,4-dinitrophenylhydrazone derivatives (mg/l)
2,4-DNPH derivatives Mean Coefficient of variation (%)
Acetone 1.64 0.60
Acetophenone 2.05 0.82
Cyclopentanone 0.10 2.43
2,3-Butanedione 2.94 6.39
n ¼ 5; injections in duplicate.

The reproducibility of this HPLC method was determined by carrying out five assays with the
same sample over 2 days. Standard curve construction was carried out every day with each
solution being injected twice. The calculated standard deviations were 0.60%, 0.82%, 2.43% and
6.93%, for acetone, acetophenone, cyclopentanone and 2,3-butanedione, respectively. The
recoveries of ketones were 99%, 94%, 96% and 92% for acetone, acetophenone, cyclopentanone
and 2,3-butanedione, respectively, as shown in Tables 4 and 5. The detection limit for each
ketone was obtained by successive dilution of a 50.0 mg/l sample of the respective ketone
derivative in acetonitrile. By considering a signal-to-noise ratio of 3:1 and the ketones as their 2,4-
DNPH, detection limits were estimated to be in the range of 10 (2,3-butanedione) to 20 mg/l
(cyclopentanone).
No noticeable interference of the original beverage sample color in the ketone or aldehyde
analysis was observed.
Ketone quantification in cacha@as yields the following average levels: 3.31, 1.24, 1.15 and
4.34 mg/l for acetone, acetophenone, cyclopentanone and 2,3-butanedione, respectively. For rum
 ARTICLE IN PRESS

D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573 571

Table 5
Recoveries of ketone as their 2,4-dinitrophenylhydrazone derivatives, added to cacha@a sample 2 (mg/l)
2,4-DNPH derivatives Concentration Concentration Concentration Recovery (%)
present added found (mean)
Acetone 0.387 0.100 0.482 99
Acetophenone 2.050 0.120 2.040 94
Cyclopentanone 0.116 0.100 0.207 96
2,3-Butanedione 0.517 0.100 0.568 92
n ¼ 5; injections in duplicate.

cachaça
350 rum
1 = acetone
2 = acetophenone
300 3 = cyclopentanone
4 = 2,3-butanedione
mg/L

250
e

0
1 2 3 4

Fig. 5. Mean content of ketones in cacha@a and rum.

samples, these values are 2.15, 0.35 and 1.75 mg/l for acetone, cyclopentanone and 2,3-
butanedione, respectively. The average levels of ketones in cacha@a are similar to those found in
wines. It is important to note that the 2,3-butanedione content above 4 mg/l could be undesirable
for flavor and aroma and also harmful (Maarse, 1991).
It is clear that cacha@a and rum exhibit the same qualitative profile of ketones except for the
acetophenone. The cacha@a samples exhibit the higher content of ketones, see Fig. 5. Analysis of
the ketones profile data through Principal Component Analysis (PCA) did not allow a clear
differentiation between rum and cacha@a data not shown. Therefore, ketones are not a good
chemical discriminator to distinguish rum from cacha@a.

4. Conclusion

The HPLC method, based on the labeling of ketones and aldehydes with 2,4-dinitrophenylhy-
drazine, seems to offer a rapid, sensitive, selective and reproducible methodology for
quantification of the ketones and aldehydes at the mg/L level in alcoholic beverages. Therefore,
 ARTICLE IN PRESS

572 D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573

the procedure outlined here has the desirable characteristics to be used for simultaneous
determinations of ketones and aldehydes in routine quality control of alcoholic beverage.

Acknowledgements

This research was supported by CAPES, CNPq, and FAPESP. Stimulating discussions with
Prof. Denis De Keukeleire (Ghent University) and Prof. Michele Aresta (Bari University) are very
much acknowledged.

References

Boscolo, M., Andrade-Sobrino, L. G., Lima-Neto, B. S., & Franco, D. W. (2002). Spectrophotometric determination
of caramel content in spirits aged in oak casks. Journal of the Association of Official Analytical Chemists, 85(3),
744–750.
Boscolo, M., Cardoso, D. R., Lima-Neto, B. S., & Franco, D. W. (2000). Identification and dosage by HRGC
of minor alcohols and esters in Brazilian sugar cane spirit. Journal of the Brazilian Chemical Society, 11(1),
86–93.
Campbell, T. F., Pfefferkon, R., & Rounsaville, J. F. (1985). Ullmann’s encyclopedia of industrial chemistry. Federal
Republic of Germany: VCH.
Coutrim, M. X., Nakamura, L. A., & Collins, C. H. (1993). Quantification of 2,4-dinitrophenylhydrazones of low
molecular mass aldehydes and ketones using HPLC. Chromatographia, 37, 185–190.
Drinks International (1994). July/August, 40.
Fung, K., & Grosjean, D. (1981). Determination of nanogram of carbonyl as 2,4-dinitrophenylhydrazones by high-
performance liquid chromatography. Analytical Chemistry, 53, 169–171.
Furuya, S. M., Isique, W. D., Lima-Neto, B. S., Andersen, L. M., Skibsted, L. H., & Franco, D. W. (2002). Phenols and
metals in sugar-cane spirits. Quantitative analysis and effect on radical formation and radical scavenging. European
Food Research Technology, 215(2), 169–175.
Grosjean, E., Green, G. D., & Grosjean, D. (1999). Liquid chromatography analysis of carbonyl (2,4-dinitrophenyl)
hydrazones with detection by atmospheric pressure negative chemical ionization mass spectrometry. Analytical
Chemistry, 71, 1851–1861.
Heath, H. B., & Reineccius, G. (1986). Flavor chemistry and technology. Westport, CT, NY: AVI Pub. Co.
Jenner, P. M., Hagan, E. C., & Jean, M. (1994). Food cosmet. Toxicology, 2, 327.
Liebich, H. M., & Koening, W. A. (1970). Analysis of the flavor of rum by gas–liquid-chromatography and mass
spectrometry. Journal of Chromatographic Science, 8, 527–533.
Maarse, H. (1991). Volatile compounds in food and beverages. New York: Marcel-Dekker.
Moret, I. (1992). Determination of aroma components in wine by continuous liquid–liquid extraction: A study of the
accuracy and repeatability. La Rivista della Societa" Italiana di Scienza dell’Allimentazione, 3, 223–230.
Nascimento, R. F., Cardoso, D. R., De Keukeleire, D., Lima-Neto, B. S., & Franco, D. W. (2000). Quantitative HPLC
analysis of acids in Brazilian cacha@as and various spirits using fluorescence detection of their 9-anthrylmethyl
esters. Journal of Agriculture and Food Chemistry, 48(12), 6070–6073.
Nascimento, R. F., Cardoso, D. R., Lima-Neto, B. S., & Franco, D. W. (1998). Determination of acids in
Brazilian sugar cane spirits and other alcoholic beverages by HRGC-SPE. Chromatographia, 48(11–12),
751–757.
Nascimento, R. F., Marques, J. C., Lima-Neto, B. S., De Keukeleire, D., & Franco, D. W. (1997). Qualitative and
quantitative HPLC analysis of aldehydes in Brazilian cane sugar spirits and other alcoholic beverages. Journal of
Chromatography A, 782, 13–23.
Papa, L. J., & Turner, L. P. (1972). Chromatographic determination of carbonyl compounds as their 2,4-
dinitrophenylhydrazones I. Gas chromatography. Journal of Chromatographic Science, 10, 744.
 ARTICLE IN PRESS

D.R. Cardoso et al. / Journal of Food Composition and Analysis 16 (2003) 563–573 573

Piggott, J. R., Sharp, R., & Duncan, R. E. B. (1989). The science and technology of whiskies. New York: Wiley.
Puputti, E., & Lethonen, P. (1986). High-performance liquid chromatographic separation and diode-array
spectroscopic identification of dinitrophenylhydrazone derivatives of carbonyl compounds from whiskies. Journal
of Chromatography A, 353, 164.
Rehm, H. J., & Reed, G. (1984). Biotechnology: Biotransformations. Weinheim: VCH.
Shriner, L. R. (1980). The systematic identification of organic compounds: Laboratory manual (6th edition). New York:
Wiley.