Anal. Chem. 1999, 71, 86-91
Quantitative RP-HPLC Determination of Some Aldehydes and Hydroxyaldehydes as Their 2,4-Dinitrophenylhydrazone Derivatives Eija Koivusalmi,* Elisa Haatainen, and Andrew Root
Analytical Research PB 310, Corporate Technology, NESTE OY, FIN-06101 Porvoo, Finland
A high-performance liquid chromatographic (HPLC) method is described for the quantitative determination of some aliphatic aldehydes and β-hydroxyaldehydes as their 2,4-dinitrophenylhydrazone derivatives. A method is described for the preparation of derivatives for those β-hydroxyaldehydes where no reference compounds of known purity are available. The detection limit of the method was 4.3-21.0 µg/L, depending on the aldehyde. Aldehydes such as formaldehyde (FA), propionaldehyde (PAL), n-butyraldehyde (nBAL), isobutyraldehyde (iBAL), and 2-ethylhexanal (EHAL) are widely used in different processes in the chemical industry. Identification of different compounds and their quantitative analysis are essential for the optimization of synthesis experiments. Several gas chromatographic (GC) methods have been described for the determination of aldehydes in water or air. Classically 2,4-dinitrophenylhydrazine (DNPH) has been used for derivatization despite the low volatility of these derivatives.1,2 Another GC method involves oxime derivatization and detection of the highly volatile derivatives by a nitrogen-selective detector.3-6 In one paper, the different forms of 2,2-dimethyl-3-hydroxypropionaldehyde (DHPAL) were determined by GC as oxime derivatives. After trimethylsilylation of the oxime, the total amount of DHPAL was determined.7 The most common method in recent years has been reversedphase HPLC where aldehydes, ketones, and hydroxycarbonyl compounds in exhaust gases and environmental samples are determined as their 2,4-dinitrophenylhydrazone derivatives by UV detection.8-14 Hydrazones are separated by HPLC and monitored (1) Papa, L. J.; Turner, L. P. J. Chromatogr. 1972, 10, 747-750. (2) Darlene, M.; Persson, P.; Skarping, G. J. Chromatogr. 1992, 626, 284288. (3) Magin, D. F. J. Chromatogr. 1979, 178, 219-227. (4) Magin, D. F. J. Chromatogr. 1980, 202, 255-261. (5) Levine, S. P.; Harvey, T. M.; Waeghe T. J.; Shapiro, R. H. Anal. Chem. 1981, 53, 805-809. (6) Nambara, T.; Kigasawa, K.; Iwata T.; Ibuki, M. J. Chromatogr. 1975, 114, 81. (7) Oesterhelt, G.; Pozeg, M.; Bubendorf, A.; Bartoldus, D. Fresenius Z. Anal. Chem. 1985, 321, 553-555. (8) Selim, S. J. Chromatogr. 1977, 136, 271. (9) Whittle, P. J.; Rennie, P. J. Analyst 1988, 133, 665-666. (10) Van Hoof, F.; Wittcox, A.; Van Buggenhout, E.; Janssens, J. Anal. Chim. Acta 1985, 169, 419. (11) Chiavari, G.; Bergamini, C. J. Chromatogr. 1985, 318, 427-432. (12) Barnes, A. R. Pharm. Acta Helv. 1993, 68, 113. (13) Edelkraut, F.; Brockmann, U. Chromatographia 1990, 30, 432-435.
86 Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
using a UV/visible-detector at a wavelength between 340 and 380 nm, depending on the absorption maximums of the relevant hydrazone.15 The qualitative analysis without derivatization has been described for formaldehyde, propionaldehyde, and 2,2bis(hydroxymethyl)butyraldehyde (BHBAL) among other compounds in synthesis mixtures by reversed-phase HPLC using a refractive index detector.16 This paper deals with the HPLC analysis of the aldehydes and β-hydroxyaldehydes as their 2,4-DNPH derivatives using an ultraviolet detector. The samples were obtained from reaction mixtures of FA with different aliphatic aldehydes that produced β-hydroxyaldehydes. The DNPH derivatization of DHPAL has been used earlier for elemental analysis for compound identification17 and for polarographic analysis,18 but no applications for HPLC are known to us. No references were found in the literature either concerning derivatization or HPLC analysis of β-hydroxyaldehydes BHBAL and 2,2-bis(hydroxymethyl)propionaldehyde (BHPAL) as their DNPH derivatives. The determination of hydroxyaldehydes as their DNPH derivatives is not as straightforward as for other aldehydes, because of the numerous side reactions in derivatization. The acid- or basecatalyzed dehydration reactions19 do not interfere with the determination of the β-hydroxyaldehydes used in this study, due to the absence of R-hydrogen. The formation of cyclic or oligomeric acetals typical for β-hydroxyaldehydes20,21 is minimized in the acidic derivatization conditions. The acid- or base-catalyzed dehydration reactions, as well as the reported rearrangement reactions of the derivatives of R-hydroxyketones or unsaturated aldehydes,14,22 may interfere with the determination of the side reaction products of BHPAL and BHBAL: 2-(hydroxymethyl)propionaldehyde (HPAL), 2-(hydroxymethyl)butyraldehyde (HBAL), 2-methyl-2-propenal (R-methylacrolein, AMA), and 2-ethyl-2-propenal (R-ethylacrolein, AEA). (14) Ferioli, F.; Vezzalini, F.; Rustichelli, C.; Gamberini, G. Chromatographia 1995, 41, 61-65. (15) Rappoport, Z.; Sheradsky, T. J. Chem. Soc. B 1968, 277-291. (16) Cairati, L.; Cannizzaro, M.; Comini, A.; Gatti, G.; Vita-Finzi, P. J. Chromatogr. 1981, 216, 402-405. (17) Spa¨th, E.; Szilagyi, J. Ber. Dtsch. Chem. Ges. 1943, 76, 949-956. (18) Tikhomirova, G. P.; Belous, G. C. Zh. Anal. Khim. 1972, 27, 173-177. (19) Shine, H. J. J. Org. Chem. 1959, 24, 1790-1791. (20) Santoro, E.; Chiavarini, M. J. Chem. Soc., Perkin Trans. 1978, 2, 189-192. (21) Cairati, L.; Cannizzaro, M.; Comini, A.; Gatti, G.; Pasquon, I.; Vita-Finzi, P. Chim. Ind. (Milan) 1981, 63, 723-725. (22) Braddock, L. I.; Garlow, K. Y.; Grim, A. F.; Kirkpatrick, A. F.; Pease, S. W.; Pollard, A. J.; Price, E. F.; Reissmann, T. L.; Rose, H. A.; Willard, M. L. Anal. Chem. 1953, 25, 301-306. 10.1021/ac980699f CCC: $18.00
© 1998 American Chemical Society Published on Web 12/01/1998
In the method developed here, the chromatographic conditions are selected so that the cis and trans isomers of DNPH derivatives23,24 are not separated. This is more sensitive to small aldehyde concentrations than the direct HPLC determination with RI detector described by Cairati et al.16 for similar sample matrixes. The main advantage of the DNPH method is that the derivatization minimizes the volatility of the low-molar-mass compounds. Also different aldehydes and ketones can be analyzed simultaneously in complex mixtures since the derivatization reaction enhances the detector response and the detection limit of the chromatographic analysis for aldehydes, but the other, otherwise interfering compounds, cannot be detected. EXPERIMENTAL SECTION Apparatus. In the study, a Hewlett-Packard model 1090 HPLC with a diode array detector (6-mm flow cell, 4-µm slit) was used. The column was Waters Nova-Pak C18, 150 × 4 mm, 60-Å pore size, 4-µm particle size at 40 °C. The detection wavelength was 360 nm and reference wavelength 550 nm. The injection volume was 5 µL. The eluents were (A) ultrapure water and (B) acetonitrile (ACN). The flow rate of the eluent was 1 mL/min and the gradient program as follows:
time/min %B
0 40
2 40
10 98
15 98
16 40
The melting points of the solid DNPH standards were determined by a Mettler DSC 30 differential scanning calorimetry under nitrogen atmosphere. The sample amount was 5 mg and the heating rate 10 °C/min. The structures of the BHBAL-DNPH derivatives were also characterized with solid-state NMR using a Chemagnetics CMX Infinity NMR spectrometer operating at 270 MHz. The samples were placed in 6-mm zirconia rotors and spun at 5 kHz. The crosspolarization (CP) experiment was carried out using 65-kHz rf fields on both channels, and a contact time of 2 ms was used. The recycle delay was 5 s and around 1000-2000 transients were acquired. The dipolar dephasing experiments were carried out with a dipolar dephasing delay of 50 µs. This experiment was the same as the CPMAS experiment except that acquisition of the FID was carried out after a 50-µs window after the contact pulse during which the decoupler was turned off.25 The acquired FIDs were Fourier transformed without any weighting function. Reagents. The water used in HPLC and sample preparation was deionized and further purified via a Milli-Q Water System (Millipore). ACN was HPLC grade from Rathburn (Walkerburn, Scotland), 2,4-DNPH (H2O = 50%) was p.a. grade from Fluka (Buchs, Switzerland), H2SO4 was 95-97% p.a. from Merck (Darmstadt, Germany), H2SO4 solution was 1.0 M from FF Chemicals (Yli-Ii, Finland), H3PO4 was 85% p.a. from J. T. Baker (Deventer, Holland), ethanol was 95% from Primalco (Rajama¨ki, Finland), formaldehyde was 37-38% solution in water from Merck, propionaldehyde 97%, butyraldehyde 99%, isobutyraldehyde 98%, and 2-ethylhexanal 98% were from Aldrich (Steinheim, Germany),
2-methyl-2-propenal was 95% practical grade from Fluka, and 2-ethyl-2-propenal was 85% technical grade from Aldrich. Hydroxyaldehydes were in-house synthesis products for analytical purposes. Preparation of 2,4-DNPH Solution and Carbonyl-DNPH Derivatives for Calibration Standards. The derivatizing agent for samples was prepared by dissolving 7.6 g of DNPH to 500 mL of ACN in a 500-mL volumetric flask. The solution was kept in an ultrasonic bath for 20 min to obtain a saturated solution. Hydrazone standards of all the other aldehydes but hydroxyaldehydes were prepared by the method described by Shriner et al.26 In this method, 4 g of DNPH was dissolved in 20 mL of concentrated sulfuric acid and then added to a solution of 28 mL of water in 100 mL of absolute ethanol. About 1.5 g of each aldehyde was dissolved in 60 mL of 95% ethanol and then 50 mL of the above prepared acidic DNPH solution was added. In the case of DHPAL, ∼1.0 g was dissolved in 60 mL of 95% ethanol and 2 mL of concentrated sulfuric acid and then 50 mL of the above prepared DNPH solution in sulfuric acid was added. Less aldehyde and an addition of sulfuric acid in ethanol was used for DHPAL in order to ensure that the molecule was mainly in the monomeric form before the addition of the DNPH reagent. The prepicitated DNPH derivatives were filtered off by suction on black ribbon filter paper and aldehydes were recrystallized from ethanol. The standard method described above was not suitable for the preparation of derivatives of BHPAL and BHBAL even though it worked for DHPAL, since in addition to the hydroxyaldehydeDNPH derivative, an insoluble precipitate was also formed. Thus, a new derivatization procedure was developed. The synthesis product of BHPAL or BHBAL was concentrated in a rotary evaporator and FA, IBAL, or PAL were steam distilled. The hydroxyaldehyde concentration was then evaluated to be ∼10%. About 12 g of steam-distilled hydroxyaldehyde solution, 2 mL of concentrated phosphoric acid, and 200 mL of 99.9% EtOH were mixed and heated to 55 °C. Approximately 3 g of DNPH was added slowly over 2 h to the hot solution with magnetic stirring. The obtained solution was put under nitrogen purge and evaporated to 100 mL. About 50 mL of water was added, and a bright yellow precipitate was allowed to settle in the refrigerator over the weekend. The precipitated DNPH derivative was filtered off by suction on black ribbon filter paper and the aldehyde derivative was recrystallized from a 2/1 water/ethanol mixture. The derivatives were dried in a vacuum desiccator. The standard solutions for HPLC analysis of the solid DNPH derivatives were prepared by weighing 25 mg of each standard into a 100-mL volumetric flask and filled to the mark with ACN. The stock solutions were diluted further to four different concentrations with a 50/50 water/ACN mixture just prior to use. The aldehyde concentration of the standard solutions were calculated using eq 1 and molar masses from Table 1.
aldehyde (mg/L) ) purity (%) ×
Mw(aldehyde) Mw(aldehyde-DNPH)
×
aldehyde-DNPH (mg/L) (1) (23) Kiebler, D. J.; Mopper, K. Mar. Chem. 1987, 21, 135-149. (24) Terada, H.; Hayashi, T.; Kawai, S.; Ohno, T. J. Chromatogr. 1977, 130, 281286. (25) Opella, S. J.; Frey, M. H. J. Am. Chem. Soc. 1979, 101, 5856.
(26) Shriner, R. L.; Fuson, R. C.; Curtin, D. Y.; Morrill, T. C. The Systematic Identification of Organic Compounds, 6th ed, Wiley: New York, 1980.
Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
87
Table 1. Melting Points of the Recrystallized DNPH Derivatives, Molar Masses, and Limits of Detection of the Compounds Tm (°C) standard
detnd
formaldehyde propionaldehyde n-butyraldehyde isobutyraldehyde R-methylacrolein R-ethylacrolein 2-ethylhexanal DHPAL BHPAL BHBAL DNPH
165.8 155.0 120.6 185.3
lit.32 166 155 123 187 237-238 237-238
189.5 191-19217
limit of detection aldehyde ald-DNPH (µg/L) Mw (g/mol)
30.026 58.080 72.107 72.091 70.091 84.118 128.214 102.133 118.133 132.159 (198.14)
210.166 238.220 252.247 252.247 250.231 264.258 308.354 282.273 298.273 312.299
4.3 6.6 8.4 7.7 6.6 10.0 14.0 11.1 13.6 21.0
Figure 2. Structures of BHBAL, its dimer, and oligomer.
Figure 1. Gradient run of a standard mixture with 60/40 water/ACN to 2/98 water ACN. Concentrations in mg/L: BHPAL 7.21, BHBAL 8.13, FA 3.37, DHPAL 7.99, PAL 4.27, AMA 4.62, BAL 7.38, AEA 5.44, and EHAL 9.06.
Preparation of Samples. Samples were prepared by weighing 50-100 mg of reaction mixture into weighed solution of 10 mL of 0.01 M H2SO4 and 2 mL of MeOH. Out of this diluted sample solution, a 100-mg portion was weighed into a 100-mL volumetric flask with 20-30 mL of DNPH solution and 2 mL of concentrated phosphoric acid. After a derivatization reaction of 1 h, the flask was filled to the mark with 50/50 water/ACN mixture. A 2-mL aliquot of the sample solution was diluted further to 100 mL with 50/50 water/ACN mixture and filtered through a 0.45-µm Millex HV filter for analysis. RESULTS AND DISCUSSION Optimization of the HPLC Method. The retention times of different aldehydes were determined at first with isocratic conditions using 35/65 water/ACN as eluent. Under these conditions, the polar aldehyde and ketone hydrazones, which contain additional hydroxy groups, eluted with even lower retention times than the formaldehyde derivative as described in the literature.11 Actually, the most polar compound, the 2,4-DNPH derivative of BHPAL, eluted even before DNPH itself. No separation was obtained for DHPAL and formaldehyde. Using 60/40 water/ACN a better resolution was obtained, but the retention times of the other aldehydes were quite long and the peaks broad. The gradient method using 60/40 water/ACN to 2/98 water/ACN as 88 Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
Figure 3. 13C CPMAS and dipolar dephased spectra of insoluble component in BHBAL-DNPH derivatization with sulfuric acid.
eluent gave the best resolution for early-eluting hydroxyaldehydes and for late-eluting aliphatic aldehydes as shown in Figure 1. The method was not optimized to separate the cis and trans isomers from each other, but they are determined as one peak. The resolution for isobutyraldehyde and butyraldehyde was not optimized, since they do not exist in the same synthesis sample. Standard Preparation and Identification. The precipitate obtained with a Shriner’s26 method for BHBAL contained an insoluble part which was separated from ACN solution and dried. It was postulated that the structure was a DNPH derivative of the BHBAL-dimer shown in Figure 2. The precipitate was analyzed by solid-state NMR and the CPMAS and dipolar dephased NMR spectra are shown in Figure 3. The dipolar dephased spectrum reveals only those carbons that do not have Hs directly bonded to them, or methyl groups. Details of this experiment can be found elsewhere.25 It is clear from the CPMAS NMR spectrum that there are very few aromatic carbons present in this sample so a dimer DNPH structure is discounted. From the dipolar dephased spectrum, it can be deduced that the peak at 40 ppm is probably due to a quaternary carbon while that at 10 ppm is due to a methyl
Figure 5. Derivatization reaction of an aldehyde and DNPH.
Figure 4. 13C CPMAS and dipolar dephased spectra of soluble product in BHBAL-DNPH derivatization with phosphoric acid.
group. The peak at 103 ppm is characteristic for OCH2O type groups. Therefore, it seems likely that the main component in the sample is like BHBAL, and since it is insoluble it is tentatively assigned to the long-chain polymer as shown in Figure 2. Peak assignments for this structure are shown in Figure 3. In the developed method for preparation of BHPAL and BHBAL hydrazone derivatives, it was expected that an equilibrium reaction between monomer and dimers obeys the same rules as in the case of DHPAL.20 Thus, at elevated temperature and in dilute acidic conditions, the amount of monomer is high compared to the amount of acetals. Hydroxyaldehydes can be derivatized quantitatively because the reaction equilibrium lies on the side of the monomer, which is continuously consumed by the added DNPH. The CPMAS and dipolar dephased spectra of the BHBAL derivative obtained with mild phosphoric acid conditions are shown in Figure 4. Since there are basically five peaks in the aliphatic area and seven peaks in the aromatic area, the structure of BHBAL-DNPH is assumed. The peak assignments are as shown in Figure 4. Characteristic is the peak at 160 ppm due to the CHdN group. The absence of a peak at 90-100 ppm shows that there are no O-CH2-O structures in the molecule. Therefore the dimer-DNPH structure is discounted. Some of the peaks in the aromatic area are split. This can be due to inequivalent atoms, due to crystallographic inequivalences in the solid state or, for the case of atom 4, due to the dipolar coupling with a nitrogen nucleus close by. This is because the 13C-14N dipolar interaction
Figure 6. DSC melting curves of DHPAL-DNPH derivatives prepared by the method of (A) Shriner et al. (B) by the developed method.
cannot be completely averaged out by MAS since the quantization axis of the 14N nucleus is not aligned along the static magnetic field.27 Sample Preparation. The samples from synthesis were collected into an acidic solution. Methanol was added to keep the sample soluble during the time of storage before analysis. A total of 50% water was used in dilutions of standards and samples to enhance the peak resolution and peak height in a chromatogram. The water content of 60% that was used in the eluent was not used in sample and standard preparation, since some DNPH derivatives tend to precipitate. The derivatization reaction of an aldehyde and DNPH is an addition reaction followed by dehydration and needs an acid catalyst, as shown in Figure 5. The reaction rate increases with decreasing pH, and a large excess of the DNPH reagent is needed to shift the equilibrium of the reaction to the side of the derivative.28 Thus, a high concentration of (27) Groombridge, C. L.; Harris, R. K.; Packer, K. J.; Say, B. J.; Tanner, S. F. J. Chem. Soc., Chem. Commun. 1980, 4, 174.
Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
89
Table 2. Precision of the Method for Samples with a Low and a High Concentration of Hydroxyaldehyde no. 1 2 3 4 5 SD av RSD (%)
no.
phosphoric acid was used in the sample preparation and the presence of free DNPH was required in the chromatograms to ensure these requirements. The time needed for the derivatization reaction varies from 30 min to 1 h, depending on the pH and the aldehyde.28,29 In this study, a reaction time of 60 min was used, since the presence of a hydroxy group may sterically hinder the reaction of 2,4-DNPH with the aldehyde carbonyl. The stability of DNPH derivatives is good when prepared in ACN solution and stored in the refrigerator. Addition of water either in dilution of the stock standard solution or in the sample preparation causes the derivative to decompose relatively quickly, because the higher water content shifts the equilibrium of the derivatization reaction to the side of the reactants. For example, the peak height of the FA standard after 12 h is 5-7% smaller than that of the freshly diluted and analyzed FA standard. Purity Determination. The purity of the dried, solid derivatives was tested by determining the melting points with DSC. The (28) Lowe, D. C.; Schmidt, U.; Ehhalt, D. H.; Frischkorn, C. G. B.; Nurnberg, H. W. Environ. Sci. Technol. 1981, 15, 819-823. (29) Cotsaris, E.; Nicholson, B. C. Analyst 1993, 118, 265-268.
90 Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
13.17 12.52 13.17 13.02 12.72 0.2894 12.92 2.239
20.14 20.04 20.77 20.71 21.25 0.4968 20.58 2.413
PAL 0.03 0.03 0.03 0.03 0.03 0.0000 0.03 0.0000
Sample from BHBAL Synthesis, Distillated (wt %) BHBAL FA BAL AEA
1 2 3 4 5 SD av RSD (%)
Figure 7. (A) Typical UV spectra of FA (- - -), hydroxyaldehyde (- - -), saturated aldehyde (- -), and unsaturated aldehyde (s). (B) Chromatogram of a BHPAL synthesis sample. (C) Chromatogram of 2-methyl-2-pentenal.
Sample from BHPAL Synthesis (wt %) BHPAL FA
63.72 61.50 61.54 62.52 61.00 1.0811 62.06 1.742
19.84 18.53 18.93 18.70 18.36 0.5807 18.87 3.077
0.92 0.87 0.85 0.89 0.87 0.0265 0.88 3.006
0.24 0.28 0.35 0.26 0.24 0.0456 0.27 16.64
impurities present in a sample cause melting point depression. The results are collected in Table 1, except for unsaturated aldehydes which after one recrystallization degrade or undergo several rearrangement reactions before melting.22,30 The results are in good agreement with the values obtained from the literature. The melting behavior of DHPAL-DNPH derivatives prepared by Shriner’s26 method and the method with preadded sulfuric acid are compared in Figure 6. No sharp melting point can be detected in the derivative prepared by the conventional method (Figure 6a). Instead, the derivative prepared by the developed method melts sharply after first recrystallization at 189.5 °C (Figure 6b). Both the DHPAL derivatives showed similar behavior in HPLC runs. The purity of all prepared hydrazone derivatives was checked by HPLC. Water/ACN solutions of each derivative as well as reagent blanks were analyzed. The purities were determined by subtracting the peak area percentages of impurities from 100% since the UV responses of the aldehyde-DNPH derivatives are quite equal. Compound Identification in HPLC Run. The UV spectra of the standard derivatives were collected and stored in the library of the HPLC instrument. By comparing the UV/visible spectra, one can separate the peaks eluted in the samples into five groups: formaldehyde, aliphatic aldehyde, aromatic aldehyde, aldehyde with conjugated double bonds, and aldehyde with conjugated triple bonds.31 The aliphatic and hydroxyaldehydes gave identical spectra giving absorption maximums at 275 and 365 nm and an absorption minimum at 295 nm as shown in Figure 7a. Formaldehyde-DNPH gave absorption maximums at 265 and 355 nm and an absorption minimum at 295 nm. The unsaturated aldehydes gave absorption maximums at 255 and 373 nm, an absorption minimum at 310 nm, and an extra shoulder at 285 nm. The retention times increased with increasing chain length. The unsaturated aldehyde eluted slightly earlier than a saturated (30) Behforouz, M.; Bolan, L.; Flynt, M. S. J. Org. Chem. 1985, 50, 1186-1189. (31) Po ¨tter, W.; Karst, U. Anal. Chem. 1996, 68, 3354-3358.
aldehyde with the same carbon number. The highly polar hydroxyaldehydes eluted very fast as shown in Figure 1. So, by comparing the retention time of an unknown compound in a chromatogram to those of known aldehydes, and by comparing a spectrum to those in the library, one can estimate the structure of an unknown compound. For example, the peak in a BHPAL sample chromatogram in Figure 7b at retention time 8.3 min elutes in the time range of C6 aldehydes and its spectrum is similar to unsaturated aldehydes. Thus, the structure of a compound from an aldol reaction between two propionaldehydes is supposed, which was confirmed by a synthesized reference compound 2-methyl-2-pentenal in Figure 7c. A peak at sample chromatogram at 3.7 min, which elutes very fast before FA and gives a spectrum similar to a hydroxyaldehyde, is quite obviously HPAL. In the case of BHBAL, the retention time of HBAL is 4.8 min and the retention time of 2-ethyl-2-hexenal the same as for 2-ethyl-hexanal, 9.5 min. The last two mentioned compounds can be identified from their UV spectra despite the same retention time, since they do not exist in the same sample. Detection Limit, Linearity, and Precision. The detection limit of the HPLC method was calculated for the aldehyde (32) CRC Handbook of Chemistry and Physics, 57th ed.; Weast, R. C., Ed.; CRC Press: Boca Raton, FL, 1976-1977. (33) King, E. P. J. Am. Stat. Assoc. 1958, 48, 531.
hydrazone derivatives on the basis of the smallest calibration standards and a system noise of 0.06 mAU multiplied by 2. The limits of detection are collected in Table 1. The linearity of the DNPH calibration lines of aliphatic aldehydes is excellent even for very intensive peaks near 1000 mAU, if diluted with a water/ ACN 50/50 mixture. The calibration lines of hydroxyaldehydes tend to be curved at higher concentrations. The precision of the method was tested by analyzing a BHPAL synthesis sample and a concentrated BHBAL synthesis sample five times. The original values and the results are collected in Table 2. The method is quite precise for the main components and for the other side compounds except AEA. The spread of the AEA results is very large, indicating poor precision. The obviously different value of AEA, 0.35 wt %, was retained on the basis of an outlier test, since the calculated Dixons’s Q ) 0.636 does not exeed the critical value Q ) 0.717 for sample size 5 (P ) 0.05).33 ACKNOWLEDGMENT The authors thank Liisa Seitvuo and Jaana Niemi for analyzing melting points by DSC. Received for review June 30, 1998. Accepted October 15, 1998. AC980699F
Analytical Chemistry, Vol. 71, No. 1, January 1, 1999
91
