Journal of Chromatography A, 1216 (2009) 4574–4581

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Hydrophilic interaction liquid chromatography – charged aerosol detection as a

straightforward solution for simultaneous analysis of ascorbic acid and
dehydroascorbic acid
Lucie Nováková a,∗ , Dagmar Solichová b , Petr Solich a
a
Department of Analytical Chemistry, Faculty of Pharmacy, Charles University, Heyrovského 1203, 500 05 Hradec Králové, Czech Republic
b
Department of Metabolic Care and Gerontology, Charles University, Faculty of Medicine and Teaching Hospital, Sokolská 581, 500 05 Hradec Králové, Czech Republic

a r t i c l e i n f o a b s t r a c t

Article history: Ascorbic acid (AA) and dehydroascorbic acid (DHA) are small polar molecules difficult to be retained in
Received 25 January 2009 conventional chromatographic RP systems. Hydrophilic interaction liquid chromatography (HILIC) using
Received in revised form 18 March 2009 Obelisk R (100 × 3.2 mm, 5 ␮m, Sielc) analytical column and isocratic elution by ammonium acetate buffer
Accepted 23 March 2009
pH 4.2 was found to be successful at this task, while other tested HILIC columns – Obelisk N (100 × 3.2 mm,
Available online 28 March 2009
5 ␮m, Sielc) and Luna HILIC (100 × 3.0 mm, 3 ␮m, Phenomenex) were unsuccessful for the purposes of
analysis. Charged aerosol detection (CAD) has recently become a new alternative universal detection sys-
Keywords:
tem in HPLC, and was extremely convenient for the simultaneous analysis of AA and DHA without the need
Ascorbic acid
Dehydroascorbic acid
of subtraction approach and oxidation/reduction step. CAD response was found linear in defined range
HILIC in spite of the fact that CAD is designated as non-linear detection method. A simple and fast HILIC-CAD
CAD method was applied for the analysis of pharmaceutical preparations containing AA. Method validation
Pharmaceutical analysis was performed including parameters of precision, accuracy, linearity, limit of detection and limit of quan-
titation (LOQ). The method was fast, accurate and precise for both detectors with LOQAA 5 ␮g/ml for UV
detection and 10 ␮g/ml for CAD, respectively. DHA was detected only by CAD within tested concentration
range with LOQDHA 1 ␮g/ml.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction reversed-phase chromatography, ion-exchange, ion-pair chro-

matography and ion-exclusion chromatography, as was recently
Simultaneous analysis of ascorbic acid (AA) and dehydroascorbic reviewed by our group [4]. The mobile phases are often very com-
acid (DHA) is a multidisciplinary task. The biologically active isomer plex, with more than two components containing various modifiers
of ascorbic acid (vitamin C) is l-ascorbic acid. However, there are or reagents. The approach of hydrophilic interaction liquid chro-
still some discussions about the activity of l-dehydroascorbic acid matography (HILIC) has become popular especially in recent period.
as well [1,2]. The role of AA in human metabolism is complex. Its While reversed-phase chromatography was not found very conve-
protecting action against the oxidizing effect of the free radicals is of nient for the purpose of AA and DHA analysis, ion-exclusion [5,6]
crucial importance. Its concentration, or more accurately, AA/DHA and HILIC approach seem to be fast and elegant solution for their
ratio can be an indicator of the redox state of an organism. That analysis.
is why sometimes AA, together with DHA, has been determined in HILIC is an alternative of conventional RP-HPLC or NP-HPLC
all kinds of biological materials using different analytical methods. (normal-phase HPLC) and it is very convenient for the analysis of
The content of vitamin C in fruits (sum of the content of AA plus small polar molecules being weakly retained in conventional RP-
DHA) is used as an index of the health-related quality of fruits, thus HPLC systems. NP-HPLC was often replaced by HILIC method. The
great interest has increased also in a field of food analysis [3]. main reasons for that are: bad reproducibility of NP analysis, low
AA and DHA belong to the group of very small polar molecules, solubility of polar compounds in NP mobile phases and great diffi-
which are difficult to retain in conventional reversed-phase culties of ionization in NP when connection with MS detection was
(RP) chromatographic systems (Fig. 1). There are four princi- required [7].
pal approaches for the determination of AA and DHA by HPLC: In HILIC the analyte retention is believed to be caused by parti-
tioning of the analyte between a water-enriched layer of stagnant
eluent on a hydrophilic stationary phase and a relatively hydropho-
∗ Corresponding author. Tel.: +420 495067345; fax: +420 495067164. bic bulk eluent, with the main components usually being 5–40%
E-mail address: [email protected] (L. Nováková). water in ACN (acetonitrile). Conceptually, it is the most ratio-

0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2009.03.060
 L. Nováková et al. / J. Chromatogr. A 1216 (2009) 4574–4581 4575

nal way to address very hydrophilic and uncharged compounds. analytical columns with differing HILIC stationary phases. Sub-
The mobile phase is composed of high percentage of an organic sequently, a CAD approach for simple, fast and one-step analysis
solvent (typically acetonitrile) and it is complemented by small without the need for an oxidation/reduction step and subtraction
percentage of water/volatile buffer part. The water-enriched liq- approach were needed to be developed.
uid layer is established within stationary phase, thus partitioning
of solutes from mobile phase into the hydrophilic layer occurs.
Primary mechanism of separation is the partitioning between the 2. Experimental
mobile phase eluent and a water-enriched layer in the hydrophilic
stationary phase. Secondary mechanism, which could influence 2.1. Chemicals and reagents
the selectivity, is electrostatic interaction with charged station-
ary phases. Elution is enabled by increasing of the polarity of Working standards of ascorbic acid and dehydroascorbic acid
mobile phase, thus the content of water component. Under the were used during this study. They were both obtained from
HILIC conditions, stationary phase is of polar character, contain- Sigma–Aldrich (Prague, Czech Republic).
ing usually hydroxyl-ethyl, diole or amino groups or it could be Acetic acid, ammonium, formic acid (all reagent grade)
a special kind of “zwitterionic” stationary phase and some others and acetonitrile (HPLC gradient grade) were purchased from
[7,8]. HILIC approach has been so far applied in two methods deter- Sigma–Aldrich. HPLC-grade water was prepared by Milli-Q reverse
mining AA without taking into account DHA unfortunately. AA was osmosis from Millipore (Bedford, MA, USA) and meets European
analyzed together with its more stable derivatives in food and bev- Pharmacopoeia requirements.
erages [9] and in pharmaceutical preparations containing vitamin The composition of tablets was as follows: ascorbic acid
C [10]. 100 mg, corn starch, disodium edetate dihydrate, talc, cel-
Simultaneous direct detection of AA and DHA is quite com- lulose microcrystalline, disodium stearyl fumarate, disodium
plicated analytical challenge. A simple analytical method for carboxymethylamylose and orange dye.
the determination of both acids without the need of oxida-
tion/reduction pretreatment step (to get only one of redox forms)
and subsequent two step analysis (double injection of sample) 2.2. Preparation of standard solutions and samples
needed for subtraction approach has not been developed so far.
There are many detection techniques available in HPLC. However AA The stock solutions of standards were prepared by dissolving
and DHA demonstrate very different properties in UV absorption, 1.0 mg of AA and DHA standard into 1.0 ml of acetonitrile/acetic
fluorescence and electrochemical detection. Thus, AA in contrast acid 0.1% (50:50), each one separately. Stock solutions were fur-
with DHA, has strong response in UV and also electrochemically. On ther diluted by acetonitrile/acetic acid 0.1% (50:50) to achieve a
the other hand, fluorescence detection is only possible, when DHA concentration 100 ␮g/ml for system suitability test (SST) measure-
form is subjected to the derivatization reaction. Generally, there is ments, and to get individual points of calibration curve in the range
a need to transform one form into the other one in order to be able 500–1.0 ␮g/ml, using nine calibration points (500, 250, 100, 50, 25,
to determine the amount of both compounds using one detection 10, 5, 2.5 and 1.0 ␮g/ml).
technique or it is necessary to connect two detection techniques in AA tablet samples (100 mg) were prepared by dissolution in
parallel. The former is called subtraction approach and it includes 0.1% acetic acid/acetonitrile (50:50) using an ultrasonic bath. Sub-
two step analysis. First the content of AA is determined in original sequent dilutions were by water/acetonitrile (50:50) in order to get
sample to get initial concentration of AA. Subsequently the reduc- concentration 40 ␮g/ml.
tion of DHA in the sample is performed, where any DHA should be
converted to AA. After the conversion the sample is analyzed for
total AA, thus the same sample is injected in duplicate. The content 2.3. Chromatography
of DHA is calculated by subtracting the initial AA content from the
total AA content after the conversion [4,3,11,6]. A Shimadzu Prominence LC 20 system (Shimadzu, Kyoto, Japan)
Charged aerosol detection (CAD) belongs among the universal was used to perform all of the analyses. The instrument was
detection methods and it operates independent of the physic- equipped with a column oven SIL-20 AC enabling temperature con-
ochemical and spectral properties of non-volatile analytes. The trol. The built-in auto-sampler CTO-20 AC enabled cooling, which
eluent of the chromatographic system is first nebulized using a flow was important from stability reasons. Chromatographic software
of nitrogen. Created droplets are dried in a drift tube to remove Lab Solution (Shimadzu) was used for data collection and process-
mobile phase, producing analyte particles. A secondary stream of ing. Detection of AA and DHA was accomplished using a Corona CAD
nitrogen becomes positively charged as it passes through a high- system (ESA, MA, USA). An SPD-M20A diode array detection (DAD)
voltage, platinum corona wire. The charged nitrogen is then mixed system was connected in series in order to compare sensitivity of
with the stream of analyte particles where the charge migrates both.
to analyte. Charged analyte particles proceed to a collector region An Obelisk R (100 × 3.2 mm, 5 ␮m) analytical column (Sielc, IL,
where a highly sensitive electrometer produces a signal that is pro- USA), was used for the HPLC separation of AA and DHA. The column
portional to the weight of sample present, independent of chemical oven temperature was kept at 23 ◦ C. The binary mobile phase, which
structure [12,13]. Pharmaceutical applications of CAD are very rare. composed of acetonitrile and 75 mM ammonium acetate pH 4.2
It appears only two papers have been published so far [14,15]. CAD (15:85), was pumped at flow-rate 1.0 ml/min. The injection volume
has been found to achieve a higher level of sensitivity than other was 10 ␮l and the auto-sampler was cooled to 4 ◦ C. Optimal CAD
universal detection methods, such as evaporative light scattering was performed using a nitrogen pressure of 35 psi and range of 20
detection (ELSD) or refractive index detection (RI). Practical appli- pA. DAD (268 nm for AA and 210 nm for DHA) was performed in
cations in the field of pharmacy, using CAD on real samples include order to compare sensitivity of both detectors in analysis of AA and
e.g. analysis of paclitaxel and its impurities [16] or the analysis of DHA.
the compounds without UV chromofore – etidronate and biphos- In this study a comparison of HILIC stationary phases was also
phonate [17]. performed, thus Obelisk N (100 × 3.2 mm, 5 ␮m, Sielc) and Luna
The aim of this work was to optimize HILIC approach for simul- HILIC (100 × 3.0 mm, 3 ␮m, Phenomenex, Prague, Czech Republic)
taneous determination of AA and DHA and to compare various were also included in the study.
4576 L. Nováková et al. / J. Chromatogr. A 1216 (2009) 4574–4581

2.4. System suitability test and validation

An important part of method validation is the SST, details of

which are usually given in Pharmacopoeias [18,19]. The SST was
performed under optimized chromatographic conditions. Theoret-
ical plates, peak asymmetry, resolution of individual compounds
and the repeatability of reference standard solution injections have
been established (retentions times and peak areas were checked).
Calibration curves of ascorbic acid in the concentration range
Fig. 1. Chemical structures of ascorbic acid and dehydroascorbic acid. 500–1.0 ␮g/ml were measured. The applicability of the method
was verified on real samples of pharmaceutical tablets containing
100 mg ascorbic acid per tablet. Method precision and accuracy was
established. For precision, six samples of tablets were tested at 100%

Fig. 2. Retention behaviour of AA and DHA on Luna HILIC. AmAc: ammonium acetate, AcAc: acetic acid.
 L. Nováková et al. / J. Chromatogr. A 1216 (2009) 4574–4581 4577

of content, which corresponds to ICH (International Conference on of diol stationary phase and Obelisk N and R were chosen as rep-
Harmonization) requirements [20]. Accuracy was determined as a resentatives of zwitterionic stationary phases inversely charged.
method recovery using six samples spiked by AA solutions to get Obelisk N carries positive charge on its surface and negative one
100% concentration after dilution as well. inside of stationary phase, while Obelisk R contrariwise. The influ-
ence of water content in mobile phase to retention behaviour of AA
3. Results and discussion and DHA was determined for each of additives using always ace-
tonitrile as organic part of mobile phase. Miscellaneous additives
3.1. Optimization of HILIC approach according to column working pH range were examined as follows:
water, acetic acid, ammonium acetate buffer (pH 3.8, 4.8, 5.8, 6.8,
There are many HILIC stationary phases available nowadays 7.8) at various concentrations.
including aminopropyl-, cyanopropyl-, diol or zwitterionic groups
(e.g. sulphobetain). It is obvious that they differ in retention prop- 3.1.1. Luna HILIC
erties and other chromatographic characteristics as for example Analytical column containing diol groups as HILIC chemistry
efficiency. It is the reason why in our work multiple HILIC stationary demonstrated the lowest retention capability of all tested columns
phases were compared. Luna HILIC was chosen as a representative – Fig. 2. For AA the retention time was typically less then 4.0 min,

Fig. 3. Retention behaviour of AA and DHA on Obelisk N. AmAc: ammonium acetate, AcAc: acetic acid, AmF: ammonium formate.
4578 L. Nováková et al. / J. Chromatogr. A 1216 (2009) 4574–4581

Fig. 4. Retention behaviour of AA and DHA on Obelisk R. AmAc: ammonium acetate, AcAc: acetic acid.

unless the concentration of ammonium acetate was very high (more 3.1.2. Obelisk N
than 100 mM). The lowest retention agent was pure water. Ammo- Analytical column carries a positive charge on the surface of
nium acetate of pH values higher than 4.8 had positive influence zwitterionic stationary phase. A negative one is embedded inside
on the increase of AA retention. A decrease of water part concen- of stationary phase. Operating range of this analytical column is
tration up to 15% almost did not influence retention. To increase 2.0–4.5, which was the reason why higher pH value buffers were
the retention of AA it was necessary to lower the water part con- excluded from the experiment. Contrariwise to Luna HILIC, the
centration up to 5 %. The retention properties of Luna HILIC seem strongest retention agent was water followed by low pH agents as
to be quite similar to ZIC HILIC analytical column (Sequant, sul- was 0.1% and 0.5% acetic acid – Fig. 3. The higher was pH, the lower
fobetain zwitterionic stationary phase) as was published by our was the retention of AA. Neither significant change was observed
group [10]. However, Luna HILIC shows much lower retention capa- when buffer concentration was increased nor when buffer type
bility, which was not influenced by the change of aqueous part was modified and ammonium acetate was replaced by ammonium
concentration as it was in case of ZIC HILIC. Retention of DHA formate. An increase of water part content had significant effect
remained almost unaffected using Luna HILIC under tested HILIC on retention of AA especially in case of acetic acid. Concerning
conditions. retention DHA, it was positively influenced by higher pH buffers
 L. Nováková et al. / J. Chromatogr. A 1216 (2009) 4574–4581 4579

Fig. 5. Typical chromatogram of the separation of standard mixture of AA and DHA-CAD analysed on Obelisk R using 75 mM AmAc buffer, pH 4.2/ACN (15:85).

contrariwise to AA, the best retention was observed at pH 4.8 or Ammonium acetate buffer at different pH values around 4
when ammonium acetate was replaced by ammonium formate. (which was pH, where AA was still well retained and DHA
Unfortunately, at these conditions the peak shape of DHA was unac- retention and peak shape was acceptable) and at different con-
ceptable. centrations were tested and the separation was finely tuned.
Finally, ammonium acetate 75 mM at pH 4.2 was able to retain
both compounds and separate DHA from the dead volume. The
3.1.3. Obelisk R only drawback was a peak shape of DHA, which was not opti-
Analytical column contrariwise to Obelisk N carries a negative mal. However, SST results of repeatability showed that it could be
charge on the surface of its zwitterionic modifier and a positive acceptable.
one is embedded inside. Operating range of this analytical column
is 2.0–6.5, which provides more possibilities for method optimiza- 3.2. Chromatographic conditions – UV detection and CAD
tion. The retention behaviour of AA was kind of similar to Obelisk N
– Fig. 4. However, the retention capability was strongly amplified. The individual parameters which could influence the response
The strongest retention agent was again water (data not shown, of CAD are as follows: the range on CAD, the ratio of water/organic
because retention times there were too long). Similarly, the lower part of mobile phase, the flow rate changes and various additives
was the pH, the higher was AA retention, thus after water acetic acid including formic acid, acetic acid and ammonium acetate buffers.
was the strongest retention agent followed by ammonium acetate The influence of pH is also significant, typically with higher pH
buffer pH 3.8. Using Obelisk R analytical column it was observed (about 7.0) the detector response decreases significantly. The range
that the increase of buffer concentration had negative influence set-up is a key parameter which influences the CAD sensitivity. The
to AA retention contrariwise to Luna HILIC or ZIC HILIC station- best responses (measured as peak area) were obtained within the 1
ary phase. Finally, it was possible to tune retention properties of and 2 pA range. However, these values were practically impossible
DHA too. Since the pH was decreased to about 4 and more, the to use due to the very high signal to noise ratio. Therefore the value
retention of DHA was increased, which unfortunately did not cor- of 50 pA was chosen for further experiments.
respond to the retention of AA. Negative influence to peak shape of Possible changes of mobile phase composition or in additives
DHA was observed with increasing pH as was already mentioned including formic acid, acetic acid and ammonium acetate buffer at
using Obelisk N. various concentrations were quite limited as was the possibility of
The compromise enabling retention of both AA and DHA, sep- pH changes due to separation reasons, see Section 3.1.
aration of DHA from the dead volume and separation one from
each other was needed with the regard to peak shape, especially 3.3. SST and validation
concerning DHA. As retention properties of AA and DHA differed
significantly, reasonable compromise was chosen with the regard The SST was performed by 10 subsequent injections of mixed
to DHA, which was difficult to separate from the dead volume. The solutions of AA and DHA which were analyzed under optimum con-
conditions for the Obelisk R were further finally tuned based on ditions. A typical chromatogram could be seen in Fig. 5. Parameters
the results obtained by experimental study shown in Fig. 4. Various such as number of theoretical plates, peak asymmetry, resolution
pH around the value 4.5, which appeared to be convenient were of individual compounds and the repeatability of reference stan-
experimentally tested together with the change of mobile phase dard solution injection were established (retentions times and peak
composition, especially around 15% of ACN, which gave good results areas were checked, the repeatability was expressed as RSD in %).
in terms of DHA and AA retention. SST results could be seen in Table 1.
4580 L. Nováková et al. / J. Chromatogr. A 1216 (2009) 4574–4581

Table 1
System suitability and method validation results – analysis of AA and DHA acid using CAD approach and DAD for the comparison of sensitivity.

Ascorbic acid Dehydroascorbic acid Limits

SST
Theoretical platesa 838 380 N > 200
HETPa 119 263 Not given
Asymmetrya 1.23 2.4 As < 2.0
Resolutiona 6.84 – Rij > 1.5
Repeatability-tr a (% RSD) 0.08 0.05 RSD < 1%
Repeatability-Aa (% RSD) 0.82 0.64 RSD < 1%

Validation

Accuracyc (%) CAD 100.72 98.50 Recovery = 100 ± 5%

Precisionc (% RSD) CAD 2.23 4.63 RSD < 5%
Accuracyc (%) UV 100.34 NA Recovery = 100 ± 5%
Precisionc (% RSD) UV 2.52 NA RSD < 5%
Linearity b (correlation coefficient) UV 0.9990 – R > 0.9990
Linearityb (equation) UV y = 51859x−579597 – –
Linearity b (correlation coefficient) CAD 0.9997 0.9999
–
Linearityb (equation) CAD y = 17779x−74286 y = 14272x + 12926
LOQ UV 5.0 ␮g/ml Not detected –
LOD UV 1.5 ␮g/ml Not detected –
LOQ CAD 10 ␮g/ml 1.0 ␮g/ml –
LOD CAD 3.0 ␮g/ml 0.33 ␮g/ml –

NA – not available, UV detector was not able to detect DHA in tested concentration range.
a
Made in 10 replicates.
b
Seven calibration levels were obtained for AA using UV detection, 5 using CAD, 7 calibration levels were obtained for DHA using CAD, using UV detection DHA did not
give any signal at tested concentration range (500–1.0 ␮g/ml).
c
Made in six replicates.

CAD measurements produced results which met the require- by RSD of peak area and thus the method was convenient for quan-
ments of the appropriate authorities (see the last column in Table 1) titative purposes.
concerning all SST parameters. Excellent repeatability of injection,
noted by the peak retention time and peak area, was observed for 3.3.1. Linearity – calibration range
both detectors (RSD < 1%, UV data not shown, as they were available Calibration curves of AA and DHA in the concentration range
for AA only). The only parameter, which did not meet requirements 500–1.0 ␮g/ml were measured. Results can be seen in Table 1. Five
was the parameter of peak asymmetry of DHA, which was above calibration points were obtained for AA (250–10 ␮g/ml) and seven
2.0, however, peak area integration was repeatable as was proven calibration points for DHA (100–1.0 ␮g/ml) using CAD. UV detec-

Fig. 6. Typical chromatogram of analysis of tablet sample containing AA and DHA as degradation product – CAD analysed on Obelisk R using 75 mM AmAc buffer, pH 4.2/ACN
(15:85).
 L. Nováková et al. / J. Chromatogr. A 1216 (2009) 4574–4581 4581

tion enabled detection of AA only (268 nm). Seven calibration points traction process and thus repeated analysis of the same sample,
were obtained in the range of 500–5.0 ␮g/ml, while DHA was not which was necessary using all UV, fluorescence and electrochemi-
detected within whole tested range (500–0.1 ␮g/ml) using 210 nm cal detection approaches. Nor reduction/oxidation step of AA/DHA
as detection wavelength. The calibration curves were linear in the prior to analysis is needed, as both compounds are determined
defined range, thus it can be concluded that the method was appro- directly within one analytical run. Various HILIC stationary phases
priate for quantitative purposes for both UV detection and CAD were tested including Obelisk N, Obelisk R (zwitterionic phases
approach in spite of statement, that CAD is not a universally linear with inverse polarity) and Luna HILIC (diol phase).
detection method. Obelisk R was found to be most convenient using 75 mM ammo-
nium acetate buffer pH 4.2/ACN (15:75) as a mobile phase. CAD
3.3.2. Accuracy and precision detection was performed using 50 pA range, UV detection of AA
Method accuracy and precision were tested using tablet sam- was performed at 268 nm, while DHA at 210 nm.
ples of AA containing 100 mg of active substance. A typical The SST and validation results of the method were in good agree-
chromatogram is shown in Fig. 6. Six samples of tablets were pre- ment with validation requirements. The method repeatability in the
pared for each experiment. Precision was expressed as RSD of six frame of SST showed a RSD lower than 1%. Both detectors, UV and
determinations. Accuracy was expressed as % of recovery of six CAD gave linear response in the tested range, CAD in spite of belong-
determinations – see Table 1. All the results were in correspondence ing to the non-linear detection methods. The sensitivity of CAD was
with the requirements for method validation in pharmaceutical two times lower than the UV detection when applied to AA. How-
application. ever it was much more sensitive in detection of DHA, which was
The amounts found in pharmaceutical preparation corre- critical task of the analysis.
sponded well to the amount declared – tablets should contain
100 mg of vitamin C. The assay was determined as 96.5% of declared Acknowledgement
amount with RSD 2.23 %. These findings fully correspond to the
requirements of USP, which requires tablet assay 100 ± 10%. DHA The authors gratefully acknowledge the financial support of
was found in drug formulation at LOQ values with the average value GAČR 203/07/P370, MZO 00179906 and MSM 0021620822.
2,44 ␮g/ml.
References
3.3.3. Limits of detection and quantitation
Limits of detection and quantitation were established for both [1] K.L. Retsky, M.W. Freeman, B. Frei, J. Biol. Chem. 268 (1993) 1304.
[2] J.C. Deutsch, J. Chromatogr. A 881 (2000) 299.
detection approaches. They were found to be two times lower for [3] I. Odriozola-Serrano, T. Hernandez-Jover, O. Martin-Belosso, Food. Chem. 105
AA using UV detection approach, see Table 1. However, CAD gave (2007) 1151.
much better sensitivity for DHA, which was not detected by UV any [4] L. Nováková, D. Solichová, P. Solich, Trends Anal. Chem. 27 (2008) 352–367.
[5] B. Ribeiro, J. Rangel, P. Valentao, P.B. Andrade, J.A. Pereira, H. Bolke, R.M. Seabra,
way at all in tested concentration range. This is the main advantage
Food. Chem. 100 (2007) 504.
of CAD approach, because lower LOQ are required for DHA, which [6] Y. Hernandez, M.G. Lobo, M. Gonzales, Food. Chem. 96 (2006) 654.
is typically present in lower concentration than AA. [7] P. Hemström, K. Irgum, J. Sep. Sci. 29 (2006) 1784.
[8] A Practical Guide to HILIC - A Tutorial and Application Book, Sequant, Sweden,
2005, pp. 3–12.
3.4. Method selectivity [9] A. Tai, E. Godha, J. Chromatogr. B 853 (2007) 214.
[10] L. Nováková, D. Solichová, S. Pavlovičová, P. Solich, J. Sep. Sci. 31 (2008) 1634.
Peak purity index calculated by software Lab Solution was used [11] A. Karlesen, R. Blomhoff, T.E. Gundersen, J. Chromatogr. B 824 (2005) 132.
[12] R.W. Dixon, D.S. Peterson, Anal. Chem. 74 (2002) 2930.
for the evaluation of method selectivity. For AA peak purity index [13] I. Sinclair, R. Gallagher, Chromatogr. Today 1 (2008) 5.
value was 1.000 in tablet sample, which was in accordance with [14] C. Brunelli, T. Gorecki, Y. Zao, P. Sandra, Anal. Chem. 79 (2007) 2472.
the measurement of standard solution of AA. Unfortunately, peak [15] N. Vervoort, D. Daemen, G. Torok, J. Chromatogr. A 1189 (2008) 92.
[16] P. Sun, X. Wang, L. Alquier, C. Maryanoff, J. Chromatogr. A 1177 (2008) 87.
purity index could not be evaluated for DHA, because UV detection [17] X.-K. Liu, J.B. Fang, N. Cauchon, Z. Pnegzu, J. Pharm. Biomed. Anal. 46 (2008)
did not show any response at tested concentrations of real samples. 639.
[18] European Pharmacopoeia (Ph. Eur. 5), Council of Europe, Strasbourg, 5th ed.,
2004.
4. Conclusions [19] United States Pharmacopoeia 30, United States Pharmacopoeial Convention,
Rockville, MD, 2007.
The new analytical method for simultaneous determination of [20] International Conference on Harmonization (ICH), Q2 (R1): Validation of Ana-
lytical Procedures: Text and Methodology, US Food and Drug Administeration
AA and DHA was developed. The method uses HILIC approach and (FDA), Fed. Reg. 60 (March 1995) 11260.
CAD which is highly advantageous. There is no more need of sub-