Anda di halaman 1dari 8

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/10772541

Comparison of three spectrophotometric


methods for the determination of ?-oryzanol in
rice bran oil
Article in Analytical and Bioanalytical Chemistry May 2003
DOI: 10.1007/s00216-002-1700-5 Source: PubMed

CITATIONS

READS

30

540

4 authors, including:
Federico Marini
Sapienza University of Rome
128 PUBLICATIONS 1,437 CITATIONS
SEE PROFILE

Available from: Federico Marini


Retrieved on: 08 October 2016

12

Springer

Dear Author:
Please find attached the final pdf file of your contribution, which can be viewed using the
Acrobat Reader, version 3.0 or higher. We would kindly like to draw your attention to the
fact that copyright law is also valid for electronic products. This means especially that:

You may not alter the pdf file, as changes to the published contribution are
prohibited by copyright law.

You may print the file and distribute it amongst your colleagues in the scientific
community for scientific and/or personal use.

You may make your article published by Springer-Verlag available on your personal
home page provided the source of the published article is cited and Springer-Verlag is
mentioned as copyright holder. You are requested to create a link to the published
article in Springer's internet service. The link must be accompanied by the following
text: The original publication is available at http://link.springer.de or at http://link.springerny.com . Please use the appropriate URL and/or DOI for the article. Articles
disseminated via SpringerLink are indexed, abstracted and referenced by many
abstracting and information services, bibliographic networks, subscription agencies,
library networks and consortia.

Without having asked Springer-Verlag for a separate permission your


institute/your company is not allowed to place this file on its homepage.

Please address any queries to the production editor of the journal in question, giving
your name, the journal title, volume and first page number.

Yours sincerely,

Springer-Verlag

Anal Bioanal Chem (2003) 375 : 12541259


DOI 10.1007/s00216-002-1700-5

S P E C I A L I S S U E PA P E R

Remo Bucci Andrea D. Magr Antonio L. Magr


Federico Marini

Comparison of three spectrophotometric methods


for the determination of -oryzanol in rice bran oil
Received: 29 July 2002 / Revised: 6 November 2002 / Accepted: 13 November 2002 / Published online: 29 January 2003
Springer-Verlag 2003

Abstract A comparison of the results obtained by applying three spectrophotometric methods (at fixed wavelength,
second-derivative and multicomponent analysis) to the
determination of -oryzanol in rice bran oil is reported.
At fixed wavelength the results are more accurate when
using isopropyl alcohol, rather than n-heptane, to dilute
the oil samples, because the absorption bands of -oryzanol
are red-shifted and the absorbance, measured at max=
327 nm, is less affected by the interference of the oil matrix (max=314 nm in n-heptane).
However, to obtain accurate results also in oils with a
low content of -oryzanol, it is necessary to perform the
analysis using second-derivative (2D330.365) or multicomponent (=310360 nm) methods. The first one fully
removes the interference of oil matrix whilst the second,
which needs a specific computational program to process
the spectrophotometric data, furnishes evidence the presence of some unexpected interference in the analysis
and/or standards which are not representative of the
analysed samples, from the square root of the sum of the
squared differences at each point between the linear combination of the standards and the unknown spectra (RMS
error).
Finally, some aspects of the chemical, spectroscopic
(UV, IR) and thermoanalytical (TG, DSC) behaviour of
-oryzanol and the values of the parameters which enable
computation of its UV spectra are reported.
Keywords -Oryzanol Spectrophotometry Rice bran
oil

Introduction
Owing to its anti-oxidant action in comparison with tocopherol, -oryzanol, a mixture of ferulic acid esters of

R. Bucci A. D. Magr A. L. Magr F. Marini ()


Dipartimento di Chimica, Universit di Roma La Sapienza,
P. le Aldo Moro 5, 00185 Roma, Italy
e-mail: fmmonet@hotmail.com

sterols (campesterol, stigmasterol, -sitosterol) and triterpene alcohols (cycloartanol, cycloartenol, 24-methylenecycloartanol, cyclobranol)

is exciting very much interest in scientific world for its


use as an additive in foods (it is listed as oxidation inhibitor in the Food Additive List), in cosmetics (application to creams and sunscreens) and especially in pharmaceuticals. In fact, recent studies have demonstrated that
-oryzanol has several important physiological effects
which include the ability to lower plasma cholesterol [1],
to reduce cholesterol absorption and decrease early atherosclerosis [2], to inhibit platelet aggregation [3] and to increase faecal bile acid excretion [4]. The reduction in total
serum cholesterol seems to be due to the decrease of lowdensity lipoproteins (LDL) and very-low-density lipoproteins (VLDL), while high density lipoprotein (HDL) content remains quite constant [5].
-Oryzanol is a natural component of rice bran oil, so its
functional and nutritional properties could increase the commercial importance of this foodstuff, whose total world production in 1997 was just 1% of the total annual production of
vegetable oils [6]. In fact, rice bran oil can be used in practically any application to replace other vegetable oils, but its
characteristics make it particularly well suited to be used
where both functionality and health benefits are required.
The most commonly used techniques for the determination of -oryzanol in rice bran oil are HPLC and UV
spectrophotometry. The former is generally applied to investigate the chemical composition of the mixture (e.g., in
1957 Shimizu et al. separated three major components of
-oryzanol (designated A, B, and C) [7]; subsequently
(19681969) Endo et al. were able to separate well seven
components [8, 9] and, more recently (1999), Xu and
Godber isolated and identified up to ten components
[10]). UV spectrophotometry [11, 12, 13, 14, 15], is preferred to determine its total content (the absorption spec-

1255

tra of the individual components of -oryzanol are almost


identical [10]) because it is more simple, rapid and inexpensive than HPLC. However these methods, measuring
the absorbance at a single wavelength where the contribution from the oil matrix could be significant, may produce
wrong results, particularly for samples with low concentrations of the analyte. Besides, they do not furnish evidence the presence of unexpected interferents, as the accuracy of the results cannot be verified.
Therefore, in this paper, we propose two new spectrophotometric methods for the determination of the total
-oryzanol content of rice bran oil. We intend showing
that the methods based on the second-derivative UV spectrum and/or on the processing of spectrophotometric data
in a selected range by a multi-wavelength computational
program are more accurate than that based on measurement of absorbance at a fixed wavelength.
In addition, we have examined closely some aspects of
the chemical, spectroscopic (UV, IR) and thermoanalytical (TG, DSC) behaviour of -oryzanol.

Experimental
Instruments
Zero-order and second-derivative UV spectra were acquired on a
PerkinElmer 320 UVvisible spectrophotometer equipped with
1 cm quartz cells and connected to a model 3700 Data Station. The
software IF-320 and CDS-13, respectively, was used for instrument control and for the mathematical treatment of the spectra;
multi-wavelength analysis was performed with the package
QUEST, which minimizes the square root of the sum of the
squared differences at each point between a linear combination of
the standards and the unknown spectra (RMS error), to find a
least-squares best fit. In the latter case the spectra of at least a sample of rice bran oil with a known content of -oryzanol for each
geographical origin (randomly selected among those analysed
by second derivative spectrophotometry) were chosen as standards
and the analysis were performed in the wavelength range 310
360 nm. The scan speed was adjusted at 30 nm min1, slit width=
1 nm, response time=0.5 s and =5 nm (in the case of the derivative spectra). The spectrophotometer was connected with a Colora ultra-thermostat.
Liquid chromatography was performed on a ThermoQuest
SpectraSeries HPLC (Thermo Quest, San Jos, CA, USA), equipped
with an Inertsil ODS-2 column (25 cm4.6 mm5 m; Varian Inc.,
Palo Alto, CA, USA). Individual ferulic acid esters were detected
with a SpectraSeries UV6000 photodiode array detector at =
327 nm, with isocratic elution (methanolacetonitrile 50:50) at a
flow rate of 0.8 mL min1. In addition, wavelength scans were performed in the range 230 to 350 nm. Retention times in the oil matrix were confirmed by chromatography and comparison of the
wavelength scans with the -oryzanol standard. The amount of total -oryzanol was quantitated by peak analysis using ChromQuest
(Thermo Quest) software.
IR spectra were recorded using a PerkinElmer 1600 FTIR
spectrometer.
The thermal measurements were carried out using a Perkin
Elmer TGS-2 Thermogravimetric Analyzer, connected to a model
3700 Data Station, and a PerkinElmer DSC-7 Differential Scanning Calorimeter (PerkinElmer, Shelton, CT, USA) equipped
with a multitasking software for instrument control and data analysis. Unless otherwise specified, thermogravimetry (TG) and differential scanning calorimetry (DSC) runs were made on sample of
about 12 mg (TG) and 0.30.5 mg (DSC) in a stream of N2 or O2
(flow rate, 50 mL min1), heating rate 10 C min1.

Materials
Pure -oryzanol was obtained from Pure -oryzanol Capsules
(Vitamin Power, USA) by solidliquid extraction in a Soxhlet apparatus, using diethyl ether as solvent, in which all the excipients
(calcium sulfate, magnesium stearate, talc and silica) were not soluble; the solvent was evaporated under reduced pressure and the
white powder obtained was checked for purity by means of HPLC
and TG analyses.
The rice bran oils [30 samples produced in three geographical
areas: Italy (labelled I01I10), Thailand (T01T10) and Switzerland
(S01S10)] were stored in dark bottles at 4 C until chemical analyses were performed. Each sample was appropriately diluted (n-heptane or i-propanol) for spectrophotometric analysis, so that the absorbance would lie in the range of minimum instrumental error.
All chemicals were analytical-reagent grade.
Standard solutions
About 100 mg L1 stock standard solutions were prepared by dissolving the required amount, accurately weighed, of pure and dry
-oryzanol in n-heptane (Sh1) and isopropyl alcohol (Si2) (calibrated flask). The concentrations required for the analysis were obtained by suitable dilution of these stock solutions.
The linearity of the response in spectrophotometric measurements was assessed using ten standard solutions in the analyte concentration range 4 to 50 mg L1.
Three of the oil samples, one for each geographical area, randomly chosen, (I05, T03, S08) were also analysed for total -oryzanol
content (sum of the concentration of the individual components,
identified by their UV spectra) by HPLC and the results were used
to check the accuracy of the spectrophotometric methods.

Results and discussion


Chemical, spectroscopic and thermoanalytical properties
-Oryzanol is a white or yellowish white odourless crystalline powder insoluble in water, slightly soluble in diethyl ether and n-heptane, more in isopropyl alcohol and
practically soluble in chloroform.
A solution of -oryzanol in n-heptane shows a complex
spectrum (Fig. 1a) with several partially overlapped ab-

Fig. 1 UV absorption spectra of -oryzanol (25 mg L1) in n-heptane (a) and in isopropyl alcohol [experimental (b) and computergenerated (c)]

1256
Table 1 Values of the parameters used to generate the UV spectrum of -oryzanol according to Eqs. (1) and (2) (=400230 nm;
C=25.0 mg L1 in isopropyl alcohol)
Gaussian
curve

amax,k C

max,k

1
2
3
4
5
6

0.154148
0.790140
0.386488
0.128048
0.0262819
0.485404

344.248
325.444
292.883
276.601
246.825
235.544

10.9962
16.3977
10.7451
7.57989
3.41123
16.2633

sorption bands but only two evident maxima at =290 nm


and =314 nm. In the concentration range from about 4 to
50 mg L1 the spectra obey Beers law and the calibration
curve at 314 is a straight line passing through the origin
(R2=0.99876) so the slope represents the specific extinction
coefficient [E(314 nm)=(30.880.28) g1 L cm1]. In isopropyl
alcohol the spectrum is quite simple (Fig. 1b) and the above
reported absorbance bands are red-shifted; the specific extinction coefficient is slightly increased (hyperchromic effect) [E(327 nm)=(33.820.05) g1 L cm1; R2=0.99995].
In the same Fig. 1 (curve c) the spectrum of -oryzanol
in isopropyl alcohol, calculated via computer, is also reported.
It is well known [16], in fact, that UVvisible spectra
may be considered as the sum of some absorption bands
each exactly described by a suitable analytical function
(in practice all the bands are contained between the contours of the Gaussian and Lorentzian curves):

empirical spectrum which satisfactorily fits an experimental one within the limits of the experimental error. In
our case, the values of the parameters (see Table 1) were
determined applying a computer-assisted deconvolution
process to the analysis of an experimental spectrum of
-oryzanol [obtained by the weighted mean (C=25 mg L1)
of six independent spectra of -oryzanol standard solutions, in the concentration range 4 to 40 mg L1], on the
assumption that each individual band fits a symmetrical
bell-shaped curve described by the Gaussian equation:
2
(i max,k )

2 2
k
ai,k = amax,k e
(2)

where Ai and ai,k are, respectively, the total absorbance


and that of the kth band, at the ith wavelength, and C is
the concentration (mg L1). The calculation of the true
spectrum may be a tremendous undertaking but, for most
of the analytical purposes, it is sufficient to determine the
minimum number (N) of absorption bands to generate an

where ai,k is the amplitude of kth Gaussian curve at the ith


wavelength, amax,k, max,k and k are, respectively, the
maximum amplitude, centre and width of the kth Gaussian curve [the Peak-Fit software applies to the above
equation rather than to the similar equation [16] which
uses the wavenumber ( ) and the band half-width ()].
The spectrum of -oryzanol, calculated from the values
reported in Table 1, satisfactorily fitted the experimental one
[the spectra are perfectly overlapped in the range 400
230 nm (R2=0.999979)] and it may be used as a -oryzanol
standard spectrum in the absence of a pure product.
The IR spectrum of -oryzanol (the most significant region is reported in Fig. 2) shows several bands, consistent
with the chemical structure of this mixture of ferulic esters. The ester linkage is witnessed by a smooth carbonyl
stretching peak in the range 1732 to 1654 cm1, redshifted with respect to an aliphatic ester C=O stretching
band because of extensive conjugation, and by the two
CO stretching absorptions at 1266 and 1155 cm1. The
two bands at 1626 cm1 (symmetric stretching, weak) and
1599 cm1 (antisymmetric stretching, strong) are typical
of the trans double bond, conjugated to an aromatic ring
(ferulic acid moiety). Finally, the broad band at about
3400 cm1 can be ascribed to the hydrogen-bonded phenolicOH stretching.
The thermal stability of -oryzanol was studied by TG
and DSC in oxygen and nitrogen atmospheres. TG curves

Fig. 2 IR spectrum of -oryzanol in the region from 1800 to


1000 cm1 (KBr disc)

Fig. 3 TG of -oryzanol in nitrogen (a) and oxygen (b)

Ai =

N


ai,k C

(1)

k=1

1257

(Fig. 3) show that in O2 the anhydrous compound is stable


up to about 265 C and then it decomposes, rapidly,
by several overlapped steps (tmax=335 C and 475 C); at
500 C no residue remains. In N2 the first decomposition
process is shifted to higher temperature (tmax=360 C) but,
also in this case (t=530 C) no residue remains.
According to DSC curves (Fig. 4) the decomposition of
-oryzanol occurs, in O2, with evident exothermic reactions and, on the contrary, in N2 without any oxidative
process. In this case, in fact, is evidenced only one endothermic process at 150 C (melting peak).
Spectrophotometric determination in rice bran oil

Fig. 4 DSC of -oryzanol in nitrogen (a) and oxygen (b)

Table 2 Validation of the accuracy of the spectrophotometric determination of -oryzanol in rice bran oil
Sample -Oryzanol concentration (mg g1)
Spectrophotometric methods

I05
T03
S08

HPLC

=314 nma

=327 nmb

2D

1.0220.023
1.0610.025
10.990.20

0.8220.018
0.5920.013
10.280.23

0.4120.010
0.2540.010
9.790.16

330365 b
0.4080.018
0.2540.012
9.800.21

an-heptane
bi-propanol;

the precision is reported as the standard deviation of


the mean (N=5; p=0.05)

Fig. 5 UV and UV-2D spectra (=


5 nm) of a rice bran oil sample (a 4%
solution of I05 in isopropyl alcohol;
-oryzanol concentration=1.64 mg L1):
as it is (a) and spiked with -oryzanol
[3.03 mg L1 (b), 6.06 mg L1 (c),
9.09 mg L1 (d)]

Usually [15], the total concentration of -oryzanol in rice


bran oil is determined measuring the absorbance of an oil
solution in n-heptane (max=314 nm). However, at this
wavelength, the absorbance of the oil matrix may be not
negligible and the results inaccurate. Taking into account
the data reported in the previous paragraph, using isopropyl
alcohol (max=327 nm) it was possible to improve the accuracy of this method, but in samples with low content of
-oryzanol, the matrix interference (absorbance of -oryzanol
free rice bran oil) is very critical (Table 2, columns 2, 3, 5).
As, in the measurement range, the rice oil matrix
spectrum may be considered a quadratic curve, we supposed the use of the second derivative spectra (2D) may
resolve this interference. In fact, comparing the UV spectra of a rice oil (as it is and spiked with -oryzanol) and
those of the respective 2D, it is evident that only the last
spectra correlates linearly with analyte concentration
(Fig. 5). In fact, in the first case, the curves Ai vs.
-oryzanol concentration are straight lines but with an intercept value depending on the selected i. On the contrary, as rice oil matrix 2D spectrum is null in the con-

1258
Table 3 Determination of -oryzanol in rice bran oil: a comparison of the accuracy of the proposed spectrophotometric methods
Sample

-Oryzanol concentration (mg g1)


Spectrophotometric methods (i-PrOH)

I01
I02
I03
I04
I05
I06
I07
I08
I09
I10
T01
T02
T03
T04
T05
T06
T07
T08
T09
T10
S01
S02
S03
S04
S05
S06
S07
S08
S09
S10

=327 nm

2D

0.806
0.790
0.786
0.790
0.823
0.786
0.802
0.791
0.802
0.788
0.607
0.612
0.592
0.606
0.595
0.790
0.471
0.609
0.626
0.577
10.62
13.02
10.40
10.30
10.78
10.39
10.40
10.28
9.57
10.38

0.406
0.403
0.396
0.392
0.410
0.394
0.401
0.403
0.381
0.403
0.264
0.268
0.254
0.266
0.264
0.317
0.209
0.276
0.269
0.241
10.23
11.27
10.06
9.87
10.18
10.01
9.94
9.79
8.98
9.94

330365

Multi- (310360)
0.407 [0.001]
0.399 [0.000]
0.389 [0.000]
0.392 [0.000]
*
0.398 [0.001]
*
0.393 [0.001]
0.383 [0.001]
0.396 [0.001]
0.264 [0.001]
0.265 [0.001]
*
0.256 [0.000]
0.261 [0.000]
0.334 [0.006]
0.211 [0.000]
0.263 [0.002]
0.262 [0.002]
0.253 [0.000]
10.47 [0.002]
11.62 [0.001]
9.97 [0.001]
9.89 [0.002]
10.22 [0.001]
10.11 [0.003]
10.08 [0.003]
*
9.02 [0.000]
10.09 [0.002]

The relative standard deviation (%) was always less than 5%;
(N3; p=0.05);
In the multi-wavelength method: * spectrum selected as standard;
[RMS]

sidered range of wavelengths, the calibration curve (for


C=4 to 50 mg L1, solvent=i-PrOH, =5 nm) is a straight
line passing through the origin [Y(2D330.365)=(5.27
0.01)C (g L1); R2=0.99995] (where Y is the 2D330.365
peak-to-peak amplitude).
The results obtained applying this new method to the
analysis of I05, T03 and S08 samples agreed with those
found by HPLC (Table 2, columns 4,5), proving 2D-spectrophotometry, in this case, fully removes the interference
of the oil matrix.
Using these two spectrophotometric methods, -oryzanol
was determined in all the 30 rice bran oil samples. The results (Table 3) prove that the first method (fixed wavelength) produces results always higher than 2D and that
the relative error is inversely proportional to the -oryzanol concentration.

In Table 3 (column 4) the -oryzanol content in each oil


sample, as determined by multicomponent analysis method,
is also reported. The excellent agreement of these results
with those obtained by 2D method and the low values of
RMS error, prove it may be an alternative method for a
rapid and accurate determination of -oryzanol in rice
bran oil. However, we could obtain this accuracy in all the
analysed samples only if there was, among the standards
selected randomly, at least a spectrum of an oil from each
region. Otherwise, probably because of (even slight) matrix differences, some results were inaccurate (up to 20%),
but always marked with high RMS values (up to 0.012).
The RMS error, in fact, is an internal index of accuracy so
a high value proves the presence of some unexpected interference in the analysis and/or standards not representative of the analysed samples.

Conclusions
The results of our study show that UV spectrophotometry,
in its various aspects, is still a valid technique for the determination of analytes in real matrices (even rather complex), as it combines the simpleness and inexpensiveness
of these methods with a great precision and accuracy of its
results.
In particular, it has been pointed out that -oryzanol
can be determined in rice bran oil, of which it is a valuable
natural component, achieving satisfactory results even when
the concentration of this analyte is very low (0.10.2 mg
g1). In fact, using second derivative spectrophotometry it
is possible to get rid of the interference of the oil matrix,
which modifies, often in a not negligible way, the absorbance at the wavelength selected for the measurement
(314 or 327 nm in n-heptane or i-propanol, respectively).
Moreover, the application of the multicomponent method
enables control of the accuracy of the result obtained by
means of the RMS error value, which increases considerably in the presence of unexpected interferences and/or of
standard not fully representative of the analysed samples.
Acknowledgement The authors acknowledge the Italian Research Council (CNR) for financial support.

References
1. Seetharamaiah GS, Krishnakantha TP, Chandrasekhara N (1990)
J Nutr Sci Vitaminol 36:291297
2. Seetharamaiah GS, Chandrasekhara N (1990) Indian J Med
Res 92:471475
3. Eitenmiller RR (1997) Food Technol 51:7881
4. Nesaretnam K, Stephen R, Dils R, Darbre P (1998) Lipids 33:
461469
5. Sayre B, Saunders R (1990) Lipid Technol 2:7276
6. McCaskill DR, Zhang F (1999) Food Technol 53:5053
7. Ohta G, Shimizu M (1957) Chem Pharm Bull 5:4044
8. Endo T, Ueno K, Ianba Y (1968) Yukagaku 17:344348
9. Endo T, Misu O, Inaba Y (1969) Yukagaku 18:6367
10. Xu Z, Godber JS (1999) J Agric Food Chem 47:27242728

1259
11. Tsuchiya T, Kaneko R (1954) Rep Govt Chem Ind Res Inst
Tokyo 49:141145
12. Tsuchiya T, Kaneko R, Okubo O (1957) Tokyo Kogyo Shikensho Hokoku 52:13
13. Qiu Y, Tang B (1998) Huaxi Yaoxue Zazhi 13:3638
14. Gong Y, Yao H (2001) Zhengzhou Gongcheng Xueyuan Xuebao 22:2831

15. De BK, Bhattacharyya DK (2000) J Oil Technol Assoc India


32:1011
16. Wieteska E (1986) In: Svehla G (ed) Wilson and Wilsons
comprehensive analytical chemistry, Vol XIX, Analytical visible and ultraviolet spectrometry. Elsevier, Amsterdam, pp 63
68