Simultaneous supercritical fluid derivatization and extraction of formaldehyde by the Hantzsch reaction.
ABSTRACT A study where the Hantzsch reaction is used to produce the chemical derivatization of formaldehyde in a supercritical medium is presented in this paper. Pressure, temperature and other parameters such as static and dynamic extraction time must be optimized to increase the yield of this kinetically controlled reaction. A 2(5-1) (resolution V) factorial design was used to study the significant parameters affecting the supercritical process in terms of resolution and sensitivity. A subsequent central composite design was employed to find the conditions of maximum response. Ultraviolet-visible spectrophotometry was used as the detection technique. The optimum conditions were used for the determination of formaldehyde in real finger-paints by means of the previous addition of known quantities of this analyte to the paint. Results were compared with those obtained with supercritical fluid extraction and subsequent chemical derivatization and an improvement of sensitivity as well as a reduction of time of analysis, solvent waste and reagents consumption were observed.
Journal of Chromatography A, 896 (2000) 51–59
Simultaneous supercritical fluid derivatization and extraction of
formaldehyde by the Hantzsch reaction
F. Reche, M.C. Garrigos , A. Sanchez , A. Jimenez
aAnalytical Chemistry Department, University of Alicante, P.O. Box 99, 03080 Alicante, Spain
bInstitute of Toys Technology (AIJU), Av. Industria 23, 03440 Ibi (Alicante), Spain
A study where the Hantzsch reaction is used to produce the chemical derivatization of formaldehyde in a supercritical
medium is presented in this paper. Pressure, temperature and other parameters such as static and dynamic extraction time
must be optimized to increase the yield of this kinetically controlled reaction. A 2
used to study the significant parameters affecting the supercritical process in terms of resolution and sensitivity. A
subsequent central composite design was employed to find the conditions of maximum response. Ultraviolet–visible
spectrophotometry was used as the detection technique. The optimum conditions were used for the determination of
formaldehyde in real finger-paints by means of the previous addition of known quantities of this analyte to the paint. Results
were compared with those obtained with supercritical fluid extraction and subsequent chemical derivatization and an
improvement of sensitivity as well as a reduction of time of analysis, solvent waste and reagents consumption were observed.
2000 Elsevier Science B.V. All rights reserved.
(resolution V) factorial design was
Keywords: Hantzsch reaction; Factorial design; Derivatization, SFE; Supercritical fluid derivatization and extraction;
Extraction methods; Formaldehyde
of test methods. In the case of finger-paints, the
complexity of the matrix makes the development of
reliable methods for the determination of formalde-
hyde really difficult. Formaldehyde is broadly used
as a preservative in finger-paints and it is expected
that an upper limit (|0.1%) could be imposed for
these samples in the near future . Therefore, the
proposal of a normalized analytical method to get a
reliable and reproducible detection and determination
of this analyte makes necessary the elaboration of
specific methodology. Only a few works can be
found where formaldehyde is directly determined,
without any previous derivatization reaction [10,11].
In general, derivatization reactions are normally
used, such as those of 2,4-dinitrophenylhydrazine
zothiazolone hydrazone , pararosaniline [7,8],
Formaldehyde is a colorless gas at room tempera-
ture; high volatility and reactivity its two main
characteristics. Because of this high reactivity and
ability to be a chemical intermediate, formaldehyde
is considered a toxic and potentially carcinogenic
substance, although this fact has not been confirmed
in humans . Some recent papers dealing with the
determination of formaldehyde in a broad variety of
matrices, such as water [2–6], air [7–9], food,
biological samples and plastics , coatings ,
and cosmetics [12,13], have permitted the knowledge
of the interactions between formaldehyde and the
corresponding matrices, as well as the development
0021-9673/00/$ – see front matter
2000 Elsevier Science B.V. All rights reserved.
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–59
2-diphenylacetyl-1,3-indandione-1-hydrazone , di-
medone  and lutidine [12,13,16].
One of the most important methods for the
determination of formaldehyde is the lutidine meth-
od, which uses the Hantzsch reaction to derivatize
formaldehyde to 3,5-diacetyl-1,4-dihydro-2,6-di-
(DADHL)]. This method is relatively simple, quick,
sensitive and requires soft conditions being useful to
be employed with supercritical fluids. An important
application of this method is the determination of
formaldehyde in cosmetics . As these products
are chemically similar to finger-paints, it should be
expected that the Hantzsch reaction could be equally
employed in the determination of formaldehyde in
such materials. But the extraction of formaldehyde
from matrices is one of the important points to be
considered. As conventional extraction techniques,
such as Soxhlet, yield low recoveries, the application
of newer techniques for extraction is advisable.
The interest in supercritical fluid extraction (SFE)
has been growing rapidly during the last years. SFE
minimizes sample handling, provides fairly clean
extracts, expedites sample preparation and reduces
the use of environmentally toxic solvents .
Examples of SFE applications include N-nitro-
samines in food , semivolatile compounds ,
polychlorinated biphenyls and polycyclic aromatic
hydrocarbons from environmental samples ,
phthalates in poly(vinyl chloride)  or aromatic
amines in finger-paints . The application of SFE
to the extraction of formaldehyde, however, has been
However, one of the problems of SFE is the use of
supercritical CO as the extraction fluid. Because of
its low polarity the extraction of polar analytes, such
as formaldehyde, is difficult and recoveries are poor.
An alternative way to work with supercritical fluids
is through chemical derivatization, which permits the
decrease of the polarity of polar analytes, the in-
crease of their volatility and solubility in the super-
critical fluid and their easy separation from aqueous
and solid samples. The simultaneous supercritical
fluid derivatization and extraction (SFDE) is not
employed in a high number of applications, but some
work on organometallics from sediments and soils
, caffeine in coffee beans , phenol in wood
soot leachate  or alkylbenzensulfonates in waste-
water sludge  can be found in the literature.
A formal approach is convenient to study new
systems where several factors can be interacting and
more information is obtained with few runs by
varying several factors at once [27,28]. For this
reason, factorial designs are interesting in SFE
applications for different samples [22,29–34]. In this
case, a 2 (resolution V) factorial design was used
to study the influence of several parameters on
derivatization and extraction in terms of resolution
and sensitivity. An extra central composite design
was also developed to define the response surface as
a function of the significant parameters obtained
from the previous design.
The aim of the present work is the study of the
Hantzsch reaction as a method of derivatization of
formaldehyde under supercritical conditions. More-
over, a comparison with SFE and subsequent
derivatization is included.
2.1. Materials and chemicals
All reagents were analytical grade and obtained
from Panreac (Barcelona, Spain) and Normapur
(Prolabo, Barcelona, Spain). A formaldehyde stan-
dard solution (38%, m/v) was used to prepare all
solutions that were standardized iodimetrically.
2.2. Supercritical fluid derivatization and
SFDE was performed (off-line mode) using an
ISCO Model SFX-220 extraction system (ISCO,
Lincoln, NE, USA) consisting of an SFX-220 ex-
tractor, a SFX-200 controller and a 100DX-syringe
pump. Supercritical grade CO
Abello Linde (Valencia, Spain). A 0.2060.01-g
amount of sample was introduced into a stainless
steel cartridge (internal volume, 2.5 ml). All the
reagents were added directly to the cartridge and a
small amount of quartz wool, which helps to mini-
mize the dead volume of the cartridge, was added.
The capillary restrictor was coaxially heated and the
temperature was programmed to 2158C. In order to
trap the extracted derivatization product, the outlet of
the restrictor was introduced into a double vial
tandem  with 5 ml of distilled water in each one.
was obtained from
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–5953
All extractions were carried out in the static/dy-
namic mode, with the use of the selected static and
dynamic extraction times. The final extract was then
diluted to 15 ml.
15 min of dynamic extraction time and 80 ml of
modifier (methanol). In order to trap the extracted
formaldehyde, the outlet of the restrictor was intro-
duced in a double vial tandem with 5 ml of metha-
nol–water (10%, v/v) in each one and coaxially
heated at 858C. The final extract was then diluted to
15 ml. A 1-ml volume of the extract was then mixed
with 5 ml of the acetic acid–ammonium acetate
buffer (pH 6.4), diluted to 25 ml and heated at 608C
during 10 min . The obtained DADHL was
finally extracted with 10 ml of 1-butanol and the
absorbance was measured using 1-butanol as a
2.3. Design of experiments
A 2 fractional factorial design for a commer-
cial paint fortified with 1.5?10
hyde was carried out to distinguish the significant
parameters affecting the supercritical process. The
initial parameters to be included in the design were
pressure, temperature, static and dynamic extraction
times and acetylacetone volume. A 100-ml volume of
an acetic acid–ammonium acetate buffer (pH 6.4)
 was also added for each experiment. A
0.6060.01-g amount of wet paint was fortified with
0.25 ml of a formaldehyde aqueous solution (0.06
M) before the extraction. The results of the initial
design were used to plan a subsequent higher order
design (central composite), which was performed
with the same procedure.
mol of formalde-
3. Results and discussion
The effect of the different variables affecting the
supercritical derivatization and extraction of form-
aldehyde in finger-paints was studied by a 2V
fractional factorial design with two levels (low and
high) for five factors. This half-fraction of a 2
design was obtained by substitution of a fifth factor
with the highest order interaction between four
factors in a complete 2 factorial design. Then, the
design generator could be described as E51ABCD;
A, B, C, D and E being the five factors. This design
requires 16 experiments, performed randomized. It is
assumed that only the main factors and second-order
interactions between factors are significant for the
process [27,28]. An extra experiment was included to
have a rough estimation of the responses in the
center of the design. Attending to the unknown
behavior of the Hantzsch reaction under supercritical
conditions, the selected parameters were CO pres-
sure (P), extraction temperature (T), static and
dynamic extraction time (s and d) and volume of
acetylacetone (c). The low and high values for each
parameter were selected according to the experimen-
tal limitations and coded to be 21 and 11 from the
center of the design (0 for each parameter).With this
transformation all parameters are independent of the
units. The evaluated response parameters were res-
olution (R ) and sensitivity (S). The former was
chosen to avoid the overlapping of the peak corre-
sponding to DADHL and the co-extracted substances
in the UV–Vis spectrum, as can be seen in Fig. 1.
Resolution was calculated as a ratio between several
terms. In the numerator: the absorbance at 410 nm,
2.4. UV–Vis spectrophotometry
Spectra from 190 to 1100 nm and the absorbances
at 410 nm of the final extracts with distilled water as
reference were considered. The absorbance is associ-
ated with the product of the condensation of form-
aldehyde with ammonia and acetylacetone to form
DADHL. UV–Vis detection was carried out with a
For the second part of this work, the standard
addition method was employed to determine the
content of formaldehyde in several finger-paints. The
SFDE conditions were those of the optimal point
found in the response surface defined by a central
2.5. Supercritical fluid extraction and subsequent
In order to compare the results with those obtained
by the use of supercritical fluid extraction and
subsequent derivatization (SFE1D), experiments
were run in a similar way. From the optimum results
of a previous factorial design, SFE was carried out at
13.8 MPa, 1208C, 15 min of static extraction time,
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–59
Fig. 1. UV–Vis spectra of maximum resolution (left) and maximum sensitivity (right) from the 2 factorial design.
the difference between the absorbance at 410 nm and
the absorbance in the valley between 320 and 410
nm and the ratio between minimum and maximum
absorbances around 410 nm. In the denominator: the
absorbance at 320 nm and the absorbance in the
valley between 320 and 410 nm. The second parame-
ter was selected because of its interest as an ana-
lytical variable to give the highest signal per mol of
formaldehyde present in the sample. Therefore,
sensitivity was defined as absorbance at 410 nm.
Table 1 lists the values of each factor and the results
for each response in the different experiments.
In order to stabilize the variance of results, an
appropriate potential transformation (Y5y ) of either
response (y5R or y5S) was carried out before the
analysis of results. A suitable transformation of
response is recommended when big differences
between the values of response are found . The
best transformation is achieved when the sum of
squares of residuals is the lowest as a function of the
List of experiments in the 2
factorial design with resolution V
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–5955
exponent l. The optimum transformation for res-
olution (Y5R ) was obtained for l between 20.85
and 1.65. The significant factors and interactions
were identified by using normal probabilistic plots,
where all negligible factors and interactions are
expected to be located along an straight line and used
to estimate the variance of the design. By contrast,
points that fall well off the line would suggest the
existence of a significant influence. These plots were
employed because the degrees of freedom of the
design are not enough to calculate the error. Fig. 2
shows the normal probabilistic plot for resolution,
where l51.6 was selected because of the good
correlation coefficient of the straight line. This figure
shows two main factors, pressure and dynamic
extraction time, which are significant with positive
effects. Moreover, two negative interactions are
acetylacetone and between static and dynamic ex-
traction times. The effects of P and d can be
explained attending to the solubility and sweeping of
DADHL from the supercritical fluid extractor. How-
ever, the explanation of the two negative interactions
is not immediate. The main goal is the reduction of
the interfering reactions and improvement of the
production of DADHL. According to this, T with c
as well as s with d have opposite effects. The T*c
interaction can be justified considering that T and c
increase the velocity of the Hantzsch reaction as well
as the possibilities of finding lateral reactions. There-
fore, from Fig. 3, increasing T and reducing c seems
to be better than to increase the response. On the
other hand, the s*d interaction can be explained
considering that the bigger improvement in R is
obtained when s is short and d is long. In this
manner, the development of interfering reactions is
reduced and the sweeping of the DADHL formed is
A similar analysis was carried out for sensitivity
resulting in a potential transformation Y5S
cording to the same considerations. From the normal
probabilistic plot, shown in Fig. 4, pressure, tempera-
ture and volume of acetylacetone as well as the
interaction between pressure and temperature appear
to be significant with a positive effect. These results
can be explained in terms of the assumption that the
Fig. 2. Normal probabilistic plot of cumulative probability density function vs. calculated effects for Y5R
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–59
Fig. 3. Significant interaction for resolution from the 2factorial design.
three main factors and the significant interaction can
be directly associated with an increase in the velocity
of the Hantzsch reaction. Moreover, P can contrib-
ute, as in R, in the polarity of supercritical CO and
so, in the solubility of the DADHL. The positive sign
of the P*T interaction could be related with a
synergetic effect produced by a simultaneous in-
crease of the probability of efficient collisions be-
tween the reagents and their thermal agitation.
Further optimization is not easy to plan in terms of
Fig. 4. Normal probabilistic plot of cumulative probability density function vs. calculated effects for Y5S
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–59 57
R because two second-order interactions are present
and only the significant main factor d is included in
one of them (s*d). By contrast, S presents only one
interaction (P*T) directly related with two signifi-
cant main factors. Moreover, absorbance at 410 nm
defines specifically S and is included in R. For this
reason, an enhancement in S results in an improve-
ment in R, making possible a better comprehension
of the supercritical phenomena.
As the increase of volume of acetylacetone could
lead to the loss of the supercritical condition of the
fluid into the extractor, this parameter as well as
those considered non-significant, were maintained in
their high levels to study the effect of pressure and
temperature in more detail. Therefore, an extra
central composite design was carried out for these
two factors. This design was constructed by addition
of a factorial design 2
experiments in the centre point and others located at
an adequate distance from the center point. This
distance was selected to be equal to 21.414 and
11.414 to assure the rotatability condition of the
central composite design. Rotatability makes the
uncertainty of the design only dependent on the
distance to the center of the working range .
Furthermore, to obtain a rotatable design with uni-
form precision, five center points were used .
Contraction on the initial levels of pressure and
temperature were done to avoid that the expansion in
the star of the central composite design falls out the
experimental limits of the extractor. In this manner,
estimation of the response surface can be done
without instrumental limitations. The required ex-
periments and results are presented in Table 2. The
obtained results are shown as a response surface in
Fig. 5 where a maximum is found at 45.4 MPa,
1058C, using 15 min static and 15 min dynamic
extraction times, 100 ml of acetylacetone and 100 ml
of the acetic acid–ammonium acetate buffer. There-
fore, it can be considered that these are the optimum
conditions for the supercritical extraction of form-
aldehyde in finger-paints. Fig. 6 shows that an
improvement in resolution and sensitivity was ob-
tained when compared to the best results for each
one in the first design.
These optimum experimental conditions were
applied to the determination of formaldehyde in real
finger-paints by the standard addition method. Table
3 presents the results for several commercial finger-
paints. Four aliquots of each sample were fortified
with different known quantities of formaldehyde
(between 7.5 and 600?10
suitable solutions) and measured in triplicate. Ab-
sorbances were then related to the quantities of
formaldehyde added. These results were compared
with those obtained by derivatization of formalde-
hyde after the SFE using the standard addition
method in the same way. Although no significant
differences were observed between the two pro-
cedures, the simultaneous supercritical derivatization
and extraction permits the reduction of handling and
time of analysis, saving of expensive reagents and
reduction of solvent waste.
and a star design, with
mol with 0.25 ml of the
List of experiments in the central composite design for pressure
(P) and temperature (T)
Determination of formaldehyde by standard addition method in
Formaldehyde in finger-paint (mg/
aSFE1D: Supercritical fluid extraction and subsequent de-
rivatization. SFDE: In situ supercritical fluid derivatization and
extraction. Confidence intervals are calculated for 95% probabili-
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–59
Fig. 5. Response surface for pressure (P) and temperature (T) from the central composite design.
SFDE is a simple and effective method to get a
reliable and fast analysis with a reduction in solvent
waste and time. The Hantzsch reaction is a selective
way of derivatization for formaldehyde from finger-
paints in supercritical conditions. Factorial designs
can be considered as an effective method to study the
influence of several parameters and permit the
acquisition of more robust results with a reduced
number of experiments when compared to the classi-
cal one-to-one parameter approach. The fluidity of
finger-paints permits the application of the standard
addition method including the variability along the
whole procedure for the determination of formalde-
hyde. Relatively reproducible and reliable results are
obtained with independence of the differences
between matrices and the possible shortcomings of
any step included in the determination.
Fig. 6. UV–Vis spectrum of optimum sensitivity from the central
F. Reche et al. / J. Chromatogr. A 896 (2000) 51–59 59
191 (1980) 239.
 T. Nash, Biochem. J. 55 (1953) 416.
 S.B. Hawthorne, Anal. Chem. 62 (1990) 633A.
 W. Fiddler, J.W. Pensabene, J. AOAC Int. 79 (1996) 895.
 J.J. Langenfeld, M.D. Buford, S.B. Hawthorne, D.J. Miller, J.
Chromatogr. 594 (1992) 297.
 J.J. Langenfeld, S.B. Hawthorne, D.J. Miller, J. Pawliszyn,
Anal. Chem. 66 (1994) 909.
 M.L. Marın, A. Jimenez, J. Lopez, J. Vilaplana, J. Chroma-
togr. A 750 (1996) 183.
 M.C. Garrigos, F. Reche, K. Pernıas, A. Sanchez, A.
Jimenez, J. Chromatogr. A 819 (1998) 259.
 J.A. Field, J. Chromatogr. A 785 (1997) 239.
 J.W. Hills, H.H. Hill, T. Maeda, Anal. Chem. 63 (1991)
 S.B. Hawthorne, D.J. Miller, D.E. Nivens, D.C. White, Anal.
Chem. 64 (1992) 405.
 S.B. Hawthorne, D.J. Miller, J.J. Langenfeld, in: F. Bright,
M.E. McNally (Eds.), Supercritical Fluid Technology, Amer-
ican Chemical Society, Washington, DC, 1992, p. 165,
 D.C. Montgomery, Design and Analysis of Experiments,
Wiley, New York, 1976.
 G.E.P. Box, N.R. Draper, Empirical Model-Building and
Response Surfaces, Wiley, New York, 1987.
 J.S. Ho, P.H. Tang, J. Chromatogr. Sci. 30 (1992) 344.
 K. Li, C.P. Ong, S.F.Y. Li, J. Chromatogr. Sci. 32 (1994) 53.
 M.P. Llompart, R.A. Lorenzo, R. Cela, J. Chromatogr. A 723
 R. Cela-Torrijos, M. Miguens-Rodrıguez, A.M. Carro-Dıaz,
R.A. Lorenzo-Ferreira, J. Chromatogr. A 750 (1996) 199.
 M.P. Llompart, R.A. Lorenzo, R. Cela, J. Chromatogr. Sci.
34 (1996) 43.
 M. Kane, J.R. Dean, S.M. Hitchen, C.J. Dowle, R.L. Tranter,
Anal. Chim. Acta 271 (1993) 83.
 E. Janos, J. Balla, E. Tyihak, R. Gaborjanyi, J. Chromatogr.
The authors wish to express their appreciation to
the CICYT (Spain) (Project 1FD97-1080-C02) for
their financial support.
 R.E. Kirk, D.F. Othmer, Encyclopedia of Chemical Tech-
nology, 3rd ed., Wiley, New York, 1978.
 F. Van Hoof, A. Wittocx, E. Van Buggenhout, J. Janssens,
Anal. Chim. Acta 169 (1985) 419.
 K. Takami, K. Kuwata, A. Sugimae, M. Nakamoto, Anal.
Chem. 57 (1985) 243.
 R.J. Kieber, K. Mopper, Environ. Sci. Technol. 24 (1990)
 E. Cotsaris, B.C. Nicholson, Analyst 118 (1993) 265.
 S.Velikonja, I. Jarc, L. Zupanc ¸ic ¸-Kralj, J. Marsel, J. Chroma-
togr. A 704 (1995) 449.
 R.R. Miksch, D. WAnthon, L.Z. Fanning, C.D. Holowell, K.
Revzan, J. Glanville, Anal. Chem. 53 (1981) 2118.
 M. Pedrero-Munoz, F.J. De Villena-Rueda, L.M. Polo-Dıez,
Analyst 114 (1989) 1469.
 M. Possanzini, V. Di Palo, Chromatographia 46 (1997) 235.
 W. Tashkov, Chromatographia 43 (1996) 625.
 M. McGuire, S.H. Nahm, J. High Resolut. Chromatogr. 14
 H. Engelhardt, R. Klinkner, Chromatographia 20 (1985) 559.
 S.C. Rastogi, Contact Dermatitis 27 (1992) 235.
 CEN/Technical Committee 52/Working Group 5, Revised
Version of European Standard Draft prEN 71-7, Safety of
Toys, Part 7: Finger Paints – Requirements and Test
Methods, November 1999.