evier.com/locate/jbbm

J. Biochem. Biophys. Methods 69 (2006) 15–24

www.els

Screening and optimization of derivatization inheating block for the determination of aliphatic

aldehydes by HPLC

A. Stafiej a, K. Pyrzynska a,⁎, A. Ranz b, E. Lankmayr b

a Department of Chemistry, Warsaw University, Pasteura 1, 02-093 Warsaw, Polandb Graz University of Technology, Institute of Analytical Chemistry and Radiochemistry, Technikerstasse 4,

8010 Graz, Austria

Received 30 September 2005; received in revised form 16 November 2005; accepted 21 February 2006

Abstract

For the study of the derivatization behavior of aliphatic C1–C10 aldehydes with 2,4-dinitrophenylhy-drazine (DNPH) in a heating block a gradient elution HPLC method for separation and determination byUV detection at 360nm was applied. The influence of time, temperature, excess of reagent and stirring ontothe reaction yields were examined for conducting the reaction in a thermostated heating block. First, thederivatization procedure had been screened by experiments according to a complete factorial design inorder to evaluate the statistically significant variables. Those parameters were used to establish the optimumconditions for the reaction by means of a Box-Behnken design.© 2006 Elsevier B.V. All rights reserved.

Keywords: HPLC; Aldehydes; 2,4-Dinitrophenylhydrazine; Optimization; Hexanal

1. Introduction

The analysis of aldehydes is an important task especially in the fields of atmospheric chemistryand occupational medicine. Several of these compounds, such as formaldehyde, acetaldehyde,acrolein and benzaldehyde, are reported to be carcinogens [1]. Acetaldehyde in the presence ofalcohols reacts with the amino groups in nucleosides to yield mixed acetals which are claimed toincrease the risk of breast cancer in women [2]. Formaldehyde, probably the most extensively

Abbreviations: DNPH, dinitrophenylhydrazine; HPLC, high performance liquid chromatography.⁎ Corresponding author. Fax: +48 22 8223532.E-mail address: kryspyrz@chem.uw.edu.pl (K. Pyrzynska).

0165-022X/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.jbbm.2006.02.009

16 A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

studied aldehyde, is a common environmental pollutant from industrial activities, incinerators andcar exhaust emission [3]. Hexanal and propanal were the dominating toxic compounds emittedfrom linseed oil paint [4]. Temperature and humidity speed up the emission process, giving ahigher total amount of aldehydes. Estimation of aldehydes is also the most common approach forthe study of lipid peroxidation [5–7]. Malonaldehyde can be very useful as a presumptive markerfor the development of oxidative stress in tissues and plasmas [5]. Determination of hexanal as anindicator of the lipidic oxidation state in some food samples has been presented by Sanches-Silvaet al. [6].

Due to excellent separation efficiency and sensitivity high-performance liquid chromato-graphic (HPLC) analysis [6,10,11] is more frequently selected for a determination of aldehydesthan gas chromatography [8,9] . The most common method for the determination of aldehydesinvolves derivatization with an acidic solution of 2,4-dinitrophenylhydrazine (DNPH) to formcorresponding hydrazones followed by HPLC separation and UV–VIS detection [2,5,10,12–15]or mass spectrometry [13,16,17]. However, different derivatization conditions have been applied.A majority of the procedures reported for this purpose applies mixing of the analytes with theDNPH at room temperature [4–6,10–12,14,16] but at completely different interval times: from30min [12] to 24h [16]. DNPH-modified silica cartridges were also found to be useful, it helps todecrease analysis time [6,13,18]. Comparison of three derivatization procedures, e.g. with DNPHcartridges, ultrasonication and magnetic stirring in solution, showed that the last procedurerevealed the most effective results [6].

In the present work, the DNPH derivatization of aliphatic C1–C10 aldehydes in sealed injectionvials which were contained in a thermally controlled metal heating block providing also magneticstirring possibility has been studied. The significance of the influence of the experimentalparameters such as reaction temperature, reaction time, excess of reagent and stirring has beenestablished by statistical screening with a factorial approach. A Box-Behnken design was appliedthen in order to localize the optimum conditions in the space of variables [19]. The applicability ofthis derivatization has been demonstrated by the determination of hexanal in potato crisps.

2. Materials and methods

2.1. Reagents

Formaldehyde in solution min 35% extra pure was from Merck (Darmstadt, Germany).Acetaldehyde (99.5% pure), hexanal (98% pure), heptanal (>97% pure), octanal (98% pure) anddecanal (98% pure) were supplied from Fluka (Switzerland). Propanal, butanal, pentanal andnonanal for synthesis were from Merck (Darmstadt, Germany). Ethanol absolute (Chromasol forHPLC) was purchased from Sigma Aldrich and methanol (Rostol HPLC Gradient Grade) fromCarl Roth (Karlsruhe, Germany). Deionized water was obtained using a Milli-Q system(Millipore, USA). 2,4-Dinitrophenylhydrazine (DNPH) and hydrochloric acid fuming 37% weresupplied from Merck (Darmstadt, Germany).

Stock solutions of aldehydes were prepared in methanol and stored in a refrigerator.2,4-Dinitrohydrazine (0.20mmol) was dissolved in 10mL 2mol L−1 HCl. Samples for

analysis were prepared every day in 1mL of methanol in autosampler vials using BRANDcapillary micropipettes (Wertheim, Germany). The vials were hermetically sealed with screw capscontaining PTFE-faced silicone septa.

Potato crisps were purchased from a local supermarket and treated as described in theexperimental section.

Table 1Linear gradient profile used for the separation of the DNPH derivatives

Time (min) A (%) B (%)

0 100 04 55 4540 0 10045 0 10048 100 0

A=water–methanol 50%/50%, B=methanol 100%.

17A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

2.2. HPLC analysis

The analyses were carried with an HPLC Agilent 1100 series, equipped with a G 1313Aautosampler, G 1322 degasser and G 1311 quaternary pump (Agilent, Waldbronn, Germany).Five microlitres of standard and sample solution were injected each. Separation was performed ona C18 column, Nucleosil, 150×4.6mm by gradient elution with methanol and water at a flow rateof 1mL min−1 using a programme as listed in Table 1. For isocratic elution a mixture of methanoland water (75:25, v/v) was applied also with a flow rate of 1mL min−1. Detection wasaccomplished by recording the UV-absorption signal at 360nm.

2.3. Derivatization in heating block

The derivatization reaction with DNPH was carried in 1.5mL autosampler vials inserted inappropriate drill holes of a metal heating block which was placed on a heating plate andcontrolled by a fuzzy ETS-D4 thermal sensor from IKA Labortechnik (Stauffen, Germany).The vials containing the mixtures of analytes and reagent were heated according to the settingsof the experimental design. 5μL each of the derivatized solutions were injected into the HPLCsystem.

Hexanal was extracted from potato crisps by addition of 5mL of ethanol to 2g of asample. After 40 min extraction in an ultrasonic bath, 1.5mL of supernatant was removedand mixed with 250μL of reagent. The mixture was incubated for 35min in the heatingblock. All samples were filtered through a 0.45-μm filters before injection onto the HPLCcolumn.

2.4. Experimental design

A factorial design with a four-fold repetition of the center resulting in 24 experiments waschosen for the screening of the most influential parameters of the derivatization reaction. Theminimum and maximum settings of the four factors are given in Table 2. Maximum

Table 2Factors and levels used in the factorial design for derivatization of aldehydes

Factor Minimum Maximum

Reaction time (min) 10 90Temperature (°C) 25 60Excess of DNPH (mol) 10 360Stirring Off On

Table 3Box-Behnken design matrix with repetition of the center for derivatization in the heating block

Time (min) Temperature(°C)

Excess of DNPH(mol)

35 30 1.560 110 18.2535 70 18.2510 70 1,560 70 3510 110 18.2560 30 18.2560 70 1.510 70 3535 70 18.2535 110 3535 110 1.535 30 3535 70 18.2510 30 18.25

18 A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

chromatographic peak area of the derivatives was selected as a target function for optimization.Linear regression was computed from the response as function of the investigated factors. Themagnitude of the influence of the factors was established by analysis of variance (ANOVA) at the95% significance level.

The effect of the statistically most important factors was considered to obtain the finaloptimum conditions. The procedure was based on the calculation of a response hypersurfacewhich was derived from the experiments according to a Box-Behnken design matrix as shown inTable 3. For statistical operations, creation of the experimental design and interpretation of theexperimental results, the software package Startgraphics Plus Version 3 for Windows(Manugistics, Rockville, USA) was used.

3. Results

3.1. Separation and quantification of aldehydes

For separation of carboxylic acids, aldehydes and ketones by RP-HPLC either gradient elutionusing acetonitrile–water mobile phases [5,12,14,17–19] or isocratic elution with methanol–water[4,7] and acetonitrile–water [13] mixtures have usually been applied. In the present study amixture of ten C1–C10 aliphatic aldehydes was derivatized prior to chromatography with DNPHand first separated under isocratic conditions using a mobile phase consisting of MeOH–water(75:25, v/v/) at a flow rate of 1mL min−1. For detection the UV-absorption signal at 360nm wasrecorded (Fig. 1A). With this experimental HPLC setup, retention times for aldehydes with longchain became unduly long, retention time for decanal was above 50min. According to bandbroadening under isocratic conditions, the late eluting peaks were very wide diminishing alsodetectability.

Thus, several gradient programs were examined starting with methanol/water mixtures andending with pure methanol. The most efficient separation of DNPH derivatives of allinvestigated aldehydes was achieved following the gradient profile as presented in Table 1.Compared to isocratic elution the analytes were separated in much shorter time (Fig. 1B). The

Fig. 1. HPLC separation of the DNPHs of standard aliphatic aldehydes. (A) Isocratic elution using methanol–water (75:25,v/v). (B) Gradient elution according to programme given in Table 1. Flow rate at 1mL min−1. UV detection at 360nm.

19A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

calibration curve obtained by triplicate injection of standard solutions showed linearity in therange of 0.3–10mg L−1 with regression coefficient of 0.9987. The reproducibility, expressedas the relative standard deviation, for all analytes at concentration of 0.6mg L−1 was 1–4%(n=4).

3.2. Screening design

Reaction kinetics as well as reaction yields depend on a number of experimental variables.A careful investigation of influential parameters followed by an optimization of the entireprocedure is mandatory. Therefore, in a first approximation a screening procedure wasperformed. A two-level full factorial design consisting of 24 runs was carried out in order todetermine the main factors influencing the derivatization reaction of aliphatic aldehydes with2,4-dinitrophenylhydrazine. Extraction time, temperature, excess of reagent and stirring werechosen as experimental parameters for the screening procedure. In Table 2 the maximum andminimum values attributed to each factor are listed and the actual settings of the parametermatrix for the statistical evaluation of the best conditions is shown in Table 3. The significanceof the effects was checked by analysis of the variance (ANOVA) at the specified significancelevel.

Based on the ANOVA results for the derivatization of aldehydes a Pareto chart ofmain effects was produced and one example for acetaldehyde is given in Fig. 2. Barlengths are proportional to the absolute value of the estimated effects, which helps incomparing the relative importance of effects. Factor bars, which exceed graphically the95% significance line (at p=0.05), exert a statistically significant influence on the results.The Pareto chart demonstrates that excess of DNPH provides the most significant effectfor derivatization of acetaldehyde. The second part of Fig. 2 shows the main effects plotfor this aldehyde. The slope of the straight line is proportional to the magnitude of the

Fig. 2. Pareto chart and main effects of the screening design for the derivatization of acetaldehyde with DNPH.

20 A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

particular effect and its direction specifies a positive or a negative influence of the effect.As can be seen, an increase in the excess of reagent leads to higher derivatizationefficiency. A positive effect can also be deduced from an increase of reaction temperaturebut it was not so substantial. Stirring has no influence on the derivatization process sofar.

The screening experiments were performed with all aliphatic aldehydes from C1 to C10. In allcases the most significant effect was obtained for the excess of DNPH. An increase of the molarexcess of derivatization reagent from 1.5- to 18.5-fold resulted in a 20-fold increase of detectionsensitivity measured as peak areas in the chromatograms of the derivatives of the aliphaticaldehydes.

3.3. Final optimization of reaction conditions

The data from the factorial design experiments revealed that the variables at the levelsstudied need final optimization. The factor stirring was skipped as it has an insignificant effectfor any of the studied aldehydes. Beside the excess of reagent concentration and reactiontemperature also the time of the derivatization process was taken into consideration since it isan important analytical performance criteria. Extraction time and temperature varied from 10 to60min and from 30 to 110°C, respectively. The molar excess of DNPH was in the range of1.5- to 35-fold. Fifteen experiments were chosen for optimization of derivatization reaction for

Fig. 3. Pareto charts and main effects plots of surface response design for the derivatization of C2 and C6 aliphaticaldehydes.

21A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

all aldehydes with three-fold repetition of the center. The Pareto charts and the main effectsplots of the surface response design for the derivatization reaction of acetaldehyde and hexanalare shown in Fig. 3.

Fig. 4. HPLC-UV chromatogram of potato crisps after ethanol extraction and derivatization process for 35min at 75°C.Hexanal addition of 2mg L−1.

10 20 30 40 50 600,3

0,4

0,5

0,6

0,7

0,8

0,9

Hex

anal

con

tent

, mg/

g

Derivatization time, min)

in the day of purchaseafter 2 daysafter 3 days

Fig. 5. The effect of derivatization time (at 75°C) on the content of hexanal in potato crisps during storage.

22 A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

3.4. Determination of hexanal in potato crisps

The above described HPLC method using gradient elution and UV detection was applied fordetermination of hexanal in potato crisps. After 40min of ultrasonic extraction, derivatization ofthe analyte under optimum conditions (32-fold excess of DNPH, temperature 75°C, time 35min)was carried out. Fig. 4 shows the chromatogram of the derivatized extract. Identification ofhexanal was accomplished by standard addition proving both, retention time as well as UVspectrum. Recoveries from spiked samples ranged from 95.8% to 100.4%, indicating reliableresults from the analytical procedure.

The proposed method was also applied to evaluate the hexanal content during storage of potatocrisps. After purchasing the potato crisps were kept in a transparent bag exposed to light at roomtemperature. Results presented in Fig. 5 showed that the hexanal content was significantlyincreased from 0.34mg g−1 to 0.88mg g−1 after 3 days. Similar hexanal generation processeshave been reported for the storage of peanuts [21,22].

4. Discussion

HPLC and UV detection based on derivatization with DNPH seems to be a simple and usefulmethod for determination of aliphatic aldehydes C1–C10. Gradient elution according toprogramme presented in Table 1 allows to obtain the separation of all analytes in much shortertime. However, the derivatization process itself requires a careful consideration of thoseexperimental parameters which have an impact on the analytical performance.

Application of a factorial design for screening experiments allows to check the influence ofreaction time and temperature, excess of reagent as well as stirring as the main factors influencingthe derivatization reaction of aliphatic aldehydes with 2,4-dinitrophenylhydrazine.

The final optimum conditions were established by computation of a response surface from thedata obtained from experiments according to a Box-Behnken design matrix. Similar but notidentical plots were obtained for the target compounds process. As can be seen from the maineffects plots presented in Fig. 3, characteristic maxima could be obtained for the factors within theinvestigated ranges. The excess of reagent was identified as statistically most significant factor forthe derivatization With increasing chain length of the aliphatic aldehydes also the temperature ofthe derivatization process becomes rather influential exhibiting the significance line at a 95%

23A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

significance level. The factor stirring has no effect for the studied aldehydes. Considering theresults obtained from the systematic optimization experiments, the best conditions forderivatization with DNPH in a heating block was found for acetaldehyde with a 35-fold molarexcess of reagent and 65°C reaction temperature, while for hexanal these parameters were 32 and75°C, respectively. For both aldehydes optimum derivatization time of 35min was obtained. Anoverall optimum for the analysis of aliphatic aldehydes from C1 to C10 can be defined to be a 35-fold molar excess of DNPH and heating at 75°C for 35min.

The suitability of the proposed method, optimized formation of derivatives as well asseparation by gradient chromatography, could be proved with the analysis of hexanal in potatocrisps. This aldehyde is known as an indicator of the lipid peroxidation process [20,21]. It is amajor breakdown product of linoleic acid oxidation has an odor described as “greasy” whichcontributes to off-flavors. The temperature usually proposed in the literature for derivatization ofaldehydes with DNPH is 20–45°C and a variety of applications has been reported [2,6,14,15,18].Experiments performed at 45°C for 60min resulted however in much lower sensitivity for thedetermination of hexanal in potato crisps by HPLC compared with those obtained byderivatization at the optimized conditions of 75°C for 35min which increased sensitivity by afactor of 3.5.

References

[1] Kerns WD, Parkov KL, Donofrio DJ, Gralla EJ, Swenberg JA. Carcinogenicity of formaldehyde in rats and miceafter long-term inhalation exposure. Cancer Res 1983;43:4382–92.

[2] Nascimento RF, Marques JC, Lima Neto BS, De Keukeleire D, Franco DW. Qualitative and quantitative high-performance liquid chromatographic analysis of aldehydes in Brazilian sugar cane sprits and other distilled alcoholicbeverages. J Chromatogr, A 1997;782:13–23.

[3] Burini G, Coli R. Determination of formaldehyde in spirits by high-performance liquid chromatography with diode-array detection after derivatization. Anal Chim Acta 2004;511:155–8.

[4] Fjällström PB, Andersson B, Nilsson C, Andersson K. Drying of linseed oil paints: a laboratory study of aldehydesemission. Ind Crops Prod 2002;16:173–84.

[5] Cordis GA, Dias DK, Riedel W. High-performance liquid chromatographic peak identification of 2,4-dinitrophenylhydrazine derivatives of lipid preoxidation aldehydes by photodiode array detection. J Chromatogr,A 1998;798:117–23.

[6] Sanches-Silva A, Rodríguez-Bernaldo de Quirós A, Lopez-Hernández J, Paseiro-Losada P. Determination of hexanalas indicator of the lipidic oxidation state in potato crisps using gas chromatography and high-performance liquidchromatography. J Chromatogr, A 2004;1046:75–81.

[7] Wheatley RA. Some recent trends in the analytical chemistry of lipid preoxidation. Trends Anal Chem2000;19:617–28.

[8] Menet MC, Gueylard D, Fievet MH, Thuillier AA. Fast specific separation and sensitive quantification ofbactericidal and sporicidal aldehydes by high-performance liquid chromatography: example of glutaraldehydedetermination. J Chromatogr, B 1996;692:79–86.

[9] Dong JZ, Moldoveanu SC. Gas-chromatography-mass spectrometry of carbonyl compounds in cigarette mainstreamsmoke after derivatization with 2,4-dintrophenylhydrazine. J Chromatogr, A 2004;1027:25–35.

[10] Wu R, White LB. Automated procedure for determination of trace amounts of aldehydes in drinking water. JChromatogr, A 1995;692:1–9.

[11] Peters R, Hellenbrand J, Mengerink Y, Van der Wal SJ. On-line determination of carboxylic acids, aldehydes andketones by high-performance liquid chromatography-9 diode array detection-atmospheric pressure chemicalionisation mass spectrometry after derivatization with 2-nitrophenylhydrazine. J Chromatogr, A 2004;1031:35–50.

[12] Oliva-Teles MT, Paiga PP, Delevue-Matos CM, Alvim-Ferraz MCM. Determination of free formaldehyde in foundryresins as its 2,4-dinitrophenylhydrazone by liquid chromatography. Anal Chim Acta 2002;467:97–103.

[13] Uchiyama SE, Matsushima E, Aoyagi S, Ando M. Simultaneous determination of C1–C4 carboxylic acids andaldehydes using 2w,4-dinitrophenylhydrazine-impregnated silica gel and high-performance liquid chromatography.Anal Chem 2004;76:5849–54.

24 A. Stafiej et al. / J. Biochem. Biophys. Methods 69 (2006) 15–24

[14] Schulte. Determination of higher carbonyl compounds in used frying fats by HPLC of DNPH derivatives. AnalBioanal Chem 2002;372:644–8.

[15] Casset A, Marchand C, Le Calve S, Mirabel P, de Blay F. Human exposure chamber for known formaldehyde levels:Generation and validation. Indoor Built Environ 2005;14:173–82.

[16] Zwiener C, Gauner T, Frimmel FH. Method optimization for the determination of carbonyl compounds in disinfectedwater by DNPH derivatization and LC-ESI-MS-MS. Anal Bioanal Chem 2002;372:615–21.

[17] van Leeuwen SM, Hendriksen L, Karst U. Determination of aldehydes and ketones using derivatization with 2,4-dinitrophenylhydrazine and liquid chromatography-atmospheric pressure photoionization-mass spectrometry. JChromatogr, A 2004;1058:107–12.

[18] Uchiyama S, Aoyagi S, Ando M. Evaluation of a diffusive sampler for measurement of carbonyl compounds in air.Atmos Environ 2004;38:6319–26.

[19] Gfrerer M, Lankmayr E. Screening, optimization and validation of microwave-assisted extraction for thedetermination of persistent organochlorine pesticides. Anal Chim Acta 2005;533:203–11.

[20] Brunton NP, Cronin DA, Frank J, Monahan R. A comparison of solid-phase microextraction (SPME) fibers formeasurement of hexanal and pentanal in cooked turkey. Food Chem 2000;68:339–45.

[21] Grosso NR, Resurreccion AVA. Predicting consumer acceptance ratings of cracker-coated and roasted peanuts fromdescriptive analysis and hexanal measurements. J Food Sci 2002;67:1530–7.

[22] Lee SY, Krochts JM. Accelerated shelf life testing of whey-protein-coated peanuts analyzed by static headspace gaschromatography. J Agric Food Chem 2002;50:2022–8.