Research article

Received: 17 August 2014

Revised: 19 September 2014

Accepted: 21 October 2014

Published online in Wiley Online Library: 2 December 2014

(wileyonlinelibrary.com) DOI 10.1002/xrs.2572

Graphene oxide-based solid phase extraction of

vitamin B12 from pharmaceutical formulations
and its determination by X-ray fluorescence
Morteza Moradi,* Soraya Zenouzi, Kamran Ahmadi and Ali Aghakhani
In the present study, a novel quantitative method, namely solid phase extraction, was applied to extract vitamin B12 from pharmaceutical formulations. The technique involves the use of graphene oxide (GO) as an efficient adsorbent for solid-phase extraction of vitamin B12. Collection of GO from aqueous solution was simply achieved by applying filtration assembly. The extracted
analyte was directly analyzed by using X-ray fluorescence (XRF) spectroscopy. Factors affecting the extraction efficiency were investigated and optimized. Under the optimum conditions, enhancement factor of 46, linear dynamic range of 251000 g l
1 with
correlation of determination (R2 = 0.998) and limit of detection of 20 g l
1 were obtained for vitamin B12. The percent relative
standard deviation based on three-replicate determination was less than 8.1%. The method was successfully applied for extraction and determination of vitamin B12 in different types of pharmaceutical samples such as multivitamin tablet, effervescent tablet
and injection sample. The results showed that the proposed method based on GO was a simple, accurate, and highly efficient approach for analysis of vitamin B12. Copyright 2014 John Wiley & Sons, Ltd.

Introduction

16

Vitamin B12 (cobalamin) is an organic complex that contains a cobalt ion in its structure.[1] This vitamin is an important coenzyme
for cell development and growth, and its deficiency leads to weakness, fatigue, nausea, constipation, weight loss, pernicious anemia
and nerve degeneration.[2] Peculiarly, vitamin B12 is essential for
the human diet, but cannot be synthesized by the body. This vitamin has to be obtained from natural sources such as fish, dairy
products, egg, meat and poultry. However, compared to other vitamins, the daily requirement of vitamin B12 is relatively low.[3] The
determination of vitamin B12 in human serum at picogram level becomes necessary for the detection of its deficiency.[4] Hence, sensitive analytical detection devices are required. To counter the
deficiency of vitamin B12, many supplements are available in the form
of tablets, capsules and injections. Methods for determination of vitamin B12 remain limited due to their low sensitivity and poor selectivity.
Several methods have been so far applied for determination of
vitamin B12 including microbiological assay that uses Lactobacillus
leichmannii (ATCC 7830) as the test organism,[5] radioisotopic dilution assay,[6,7] high-performance liquid chromatography (HPLC)
with different detection methods including UV detection,[8,9] electrochemical detection,[8] flame atomic absorption spectrometry
(FAAS),[10] fluorometry,[11,12] inductively coupled plasma mass spectrometry (ICP-MS)[1315] and biosensor assay based on surface plasmon resonance (SPR).[16] These conventional methods have their
own advantages, but at the same time, they have certain drawbacks such as being laborious, time consuming, tedious, less safe,
too expensive and showing low sensitivity and selectivity. Other
disadvantages of these methods have been mentioned in the
literature.[17] But there is no report about determination of vitamin
B12 by using X-ray fluorescence (XRF).
XRF analysis is a powerful analytical tool for the spectrochemical
determination of almost all the elements present in a sample. XRF

X-Ray Spectrom. 2015, 44, 1623

radiation is induced when photons of sufficiently high energy, emitted from an X-ray source, impinge on a material. These primary
X-rays undergo interaction processes with the analyte atoms. This
allows the identification of the elements present in the analyte
and the determination of their mass or concentration. Measurement of the spectrum of the emitted characteristic fluorescence radiation is performed using wavelength dispersive (WD) and energy
dispersive (ED) spectrometers. In principle, XRF analysis is a multielement analytical technique and in particular, the simultaneous determination of all the detectable elements present in the sample is
inherently possible with EDXRF. In WDXRF both the sequential and
the simultaneous detection modes are possible.[18]
New idea for analysis of organic compounds by XRF methods extends the applicability of this methodology in further studies. So far,
only a few paper presented an indirect method for the determination XRF determination of organic compound. The method is based
on the precipitation of an ion-associate complex formed between
terazosin (as target analyte) and [Zn(SCN)4]2
and the formation
of a thin film on a membrane filter.[19] To the best of our knowledge,
this is the first report about the application of the XRF for the determination of vitamin B12 which have one cobalt atom in its structure.
XRF possesses a number of limitations when dealing with aqueous samples, such as short linear range, the requirement of closely
matching standards to overcome matrix effects, bubbles released
from solutions due to inadequate sample holder filling and heating
of the solution.[20] The extraction of analytes from a liquid onto a
solid (SPE) is an important field in sample handling of elements

* Correspondence to: Morteza Moradi, Department of Semiconductors, Materials

and Energy Research Center, Karaj, Iran. E-mail: 2m.alborzi@gmail.com
Department of Semiconductors, Materials and Energy Research Center, Karaj, Iran

Copyright 2014 John Wiley & Sons, Ltd.

Sample preparation
for their detection by XRF. The basic principle of such procedure is
based on transfer of analytes from the aqueous phase to the active
sites of the adjacent solid phase. One of the clear trends in science
and technology of the future is nanotechnology. Several sorbents
have been reported in the literature to preconcentrate metals prior
to their determination by XRF techniques, such as ion-exchange
materials,[21] polyurethane foam (PUF),[22] filter discs,[23] activated
carbon[24] and membrane filters.[25] Yamini et al.[26] reported determining the nanogram per milliliter level of V, Cr, Mn, Fe, Co, Ni, Cu
and Zn in water by XRF after using a silica gel powder. In another
work, the preconcentration of Co, Ni, Cu, Zn and Pb using graphene
oxide (GO) as solid sorbent without using chelating agent is proposed by Zawisza et al..[27] The proposed procedure is based on dispersive solid-phase microextraction (DSPME). Some new efficient
variations of these methods and new techniques extending the
possibilities of XRF for liquid solutions analysis have been proposed
in recent years. In 2010, Margu et al. reviewed practical aspects related to the use of extractive and non-extractive sample preparation
of elements before their detection by XRF.[28] However, the interest in
preconcentration methods for XRF has increasingly turned towards
the refinement of its modes for use in practical sample preparations.
In this research, GO are dispersed in aqueous samples, which promotes the immediate interaction between the vitamin B12 and GO
and shortens the time of sample preparation in comparison with a
classical SPE. After the adsorption process, the GO with adsorbed vitamin were collected onto a filter and measured directly using XRF
spectrometry without the necessity of analyte elution.

Germany). The morphology of the GO was characterized by

transmission electron microscopy (TEM, Philips EM 208, The
Netherlands).
The pH values were measured using a WTW Inolab 720 pH meter
(Weilheim, Germany). During the extraction, the sample containing
GO was stirred with a stirring speed of 1000 rpm by a heatermagnetic stirrer model 301 from Heidolph (Kelheim, Germany)
using a 10.0 mm 5.0 mm magnetic bar.
Synthesis of graphene oxide
Graphene oxide (GO) synthesis was carried out via the modified
Hummers method.[29] Graphite (5.0 g) and NaNO3 (2.5 g) were
mixed with 120 ml of H2SO4 (95%) in a 500-ml flask. The mixture
was stirred for 30 min within an ice bath. While maintaining vigorous stirring, KMnO4 (15 g) was added to the suspension. The rate
of addition was carefully controlled to keep the reaction temperature lower than 20 C. The ice bath was then removed, and the mixture was stirred at room temperature overnight. As the reaction
progresses, the mixture gradually became pasty, and the color
turned into light brownish. At the end, 150 ml of H2O was slowly
added to the pasty with vigorous agitation. The reaction temperature was rapidly increased to 98 C with effervescence, and the
color changed to yellow. The diluted suspension was stirred at
98 C for one day. Then, 50 ml of 30% H2O2 was added to the mixture. For purification, the mixture was washed by rinsing and centrifugation with 5% HCl and then deionized water for several times.
After filtration and dried under vacuum, the GO was obtained as a
gray powder.

Experimental
Sample preparation

Chemicals and reagents

All reagents used were of analytical grade. Vitamin B12 and methanol were obtained from Sigma-Aldrich (St. Louis, MO, USA). Sulfuric
acid (99.8%), hydrogen peroxide (H2O2, 30%), hydrochloric acid
(HCl, 37%), acetonitrile (ACN), sodium chloride (NaCl), sodium hydroxide and potassium permanganate (KMnO4) with the purity
higher than 99.9% were obtained from Merck Chemicals
(Darmstadt, Germany). (Graphite powder was purchased from
Samchun Pure Chemical Co Ltd (Pyeongtaek, Korea). Ultra-pure water was prepared by use of an Aqua Max-Ultra Youngling ultra-pure
water-purification system (Dongan-gu, South Korea). Stock solutions (1000 mg l
1) of the analyte were prepared by dissolving appropriate amounts of the compounds in deionized water. They
were stored under refrigeration at 4 C and were stable for at least
3 months. Working standard solutions were prepared daily by diluting the stock standard solutions to the required concentrations
with ultra-pure water. The samples of multivitamin tablet, effervescent tablet and injection sample were purchased from local markets.
Apparatus

X-Ray Spectrom. 2015, 44, 1623

Extraction procedure
2.3 ml of 1.0 mg ml-1 suspension of GO was added to 77.7 ml of
aqueous sample. The pH of the solution was adjusted to 2. Next,
the mixture was stirred by a magnetic stirrer for 10.5 min. After that,
the sample was collected onto the Millipore filter (cellulose esters,
0.2-mm pore size) using a vacuum filtration assembly. Loaded filters
were placed between two 6.0-mm-thick Mylar X-ray foils (supplied
by Chemplex Industries, Inc., New Cork, USA) mounted in special
liquid sample holders which incorporate a snap-on ring at the
end of the cell for attachment of thin-film supports. Afterwards,
samples were sealed in the sample holder of the equipment for
WDXRF analysis. The blank sample was prepared using the described
procedure and high purity water instead of the analyzed solution.
Optimization strategy
There are several factors like pH, sample volume, sorbent amount
and extraction time that affect the extraction process. In order to
obtain the optimum conditions for extraction of vitamin B12, effects

Copyright 2014 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/xrs

17

A wavelength-dispersive sequential X-ray spectrometer (8410, ARL,

Australia) equipped by super-Q soft-ware for quantitative analysis
was used for measuring the K line of all elements. The measurement parameter was 2 = 52.80 for Co. Rh tube, 60 kV 50 mA; LiF
(200) crystal; 150-m collimator; 25%75% window width; counting
time f or peak and background, a duplex (flow + Xe-sealed) detector
for the determination of Co. Prepared samples were characterized
using X-ray powder diffraction (XRD, Philips Xpert diffractometer
with Cu K radiation ( = 0.154 nm) generated at 40 kV and 30 mA
with a step size of 0.04 s
1) and Raman spectroscopy (Bruker,

For extraction and determination of vitamin B12 from multivitamin

tablet sample, one tablet was finely powdered and dissolved in
50 ml of deionized water. After filtering the solution, the supernatant was transferred into a 100-ml volumetric flask and diluted to
the mark with deionized water. Finally, the extraction procedure
was followed under the optimum conditions for 80 ml of this solution. For extraction of vitamin B12 from effervescent tablet sample,
each tablet was dissolved in 100 ml of deionized water and after filtering and pH adjustment, extraction was accomplished under the
optimum conditions.

M. Moradi et al.
of sample pH on the extraction of analyte by the proposed method
were optimized using one-variable-at-a-time (OVAT) process. In
OVAT strategy we can only control one variable at a time and other
factors are fixed. Optimization of the three other parameters on the
extraction efficiency was performed using a multivariate optimization. The software package Design-Expert 8.0.6 trial version
(Stat-Ease Inc., MN, USA) was used for experimental design, data
analysis and response surfaces.
The extraction recovery for solid phase extraction was expressed
as the percentage of total analyte amount, ns, initial (number of
moles of the analyte amoun toriginally present in the sample) transferred into the extraction phase at the end of the extraction, na, final
(number of moles finally collected in the extraction phase) was calculated by the following equation:
R% 100 na;final =na;initial

(1)

Results and discussion

Characterization of GO
Figure 1 shows the XRD patterns of natural graphite, and GO.
Graphite exhibits a strong and sharp peak at 26.40 in Fig. 1(a), indicating a higher ordered structure, that corresponds to a basal spacing d (002) = 3.3 . The pattern of GO, on the other hand, exhibits a
001 reflection at 9.07. The interlayer spacing of GO was calculated
to be 9.6 . This value is higher than interlayer spacing of graphite
(d-spacing = 3.3 , 2Theta = 26.4), due to the presence of oxygenated functional groups and intercalated water molecules.
Raman spectroscopy, as a non-destructive technique, plays a critical role in characterization of graphene materials[30] and has been
employed here to gain information about structure and defects of
the GO sheets in the composite and the precursor phase. The Raman spectrum for the graphite and GO is presented in Fig. 1B.
Two intense bands are clearly seen in GO sample at 1355 (D-band)
and 1581 cm
1 (G-band), respectively. The former peak is called D
band, which is activated once the defects are introduced into sp2
hybridized carbon networks, and the intensity of D band is proportional to the defects concentration. Thus the strong D band
intensity indicates the existence of large amount of defects in
sp2-carbon-based materials. The D-band is associated with inplane bond stretching motion of the pairs of C sp2 atoms (the
E2g phonons) and shows defects of the sp2 domain, while the
G-band is related to breathing modes of rings or -point phonons of A1g symmetry.[31]

Figure 1C displays the typical TEM image of the synthesized GO

that represent a layered and sheet-like structure with the large surface and wrinkled edge. These results indicate a good exfoliation of
graphite during the oxidation process.
Optimization
Influence of the sample pH
It is well known that pH of sample solution is one of the important
factors affecting the extraction efficiency of ionizable compounds;
thus, pH of sample solution was optimized separately using OVAT
method. According to the literature, GO can be considered as a material with a negative surface charge due to its acidic groups,[32] so a
cationic analyte can be adsorbed onto it. The ionization plot (Fig. 2)
shows the relationship between pH and the relative quantities of
each species of the compound.[33] Thus, vitamin B12 can be considered as a three-cationic at pH = 1.0 to 5.0, theoretically. In fact, vitamin B12 has different forms, and the percentage of each form is
affected by pH (some cationic and anionic forms). Therefore, at
low acidic pH, vitamin B12 molecules can be adsorbed onto the
negative surface of the GO whereas they cannot be adsorbed on
to the GO at higher pHs due to the repulsion between negative
charge of phosphate group onto vitamin B12 molecule and adsorbent surface. The influence o f acidic pHs (1.0 5.0) on the extraction efficiency was investigated. It was thus deduced that pH = 2
is suitable for extraction (Fig. 3).
Multivariate optimization
In the next step, a multivariate optimization was applied to
optimize the three factors (sample volume, sorbent amount
and extraction time). Multivariate optimization procedures involve simultaneous alteration of all experimental factors studied
according to an experimental design. Experimental designs are
used to plan the experiments in such a way that a maximum extraction recovery is gained from a minimum number of experiments. The best experimental designs for the purpose of
modeling and optimization are the response surface designs.
Having a careful design and experiments analysis, it seeks to relate a response (or output) variable to the levels of a number of
predictor (or input) variables. The input variables are sometimes called independent variables, and the performance measure or quality characteristic is the response. Basically, this
optimization process involves three major steps: performing
the statistically designed experiments, estimating the coefficients in a mathematical model and consequently predicting

18

Figure 1. Characterization of graphene oxide nanosheets: (a) XRD pattern; (b) Raman spectrum and (c) TEM image.

wileyonlinelibrary.com/journal/xrs

Copyright 2014 John Wiley & Sons, Ltd.

X-Ray Spectrom. 2015, 44, 1623

Sample preparation

Figure 2. Ioniziation plot of vitamin B12. Purple circle is positive charge and green circle is negative.

and construction of a second-order polynomial model which

can be expressed as the following equation:
Y b0 bi X i bii X 2i bij X i X j

Figure 3. The effect of pH on the extraction efficiency of vitamin B12.

X-Ray Spectrom. 2015, 44, 1623

where, Y is the response variable (cps), bo is an intercept, bi, bii

and bij are constant regression coefficients of the model and Xi,
Xj (i = 1,3; j = 1,3 and i j) represent the coded level of an independent variable. The values of the coefficients Xi and Xj are related to the effect of these variables on the response.
Coefficients with more than 1 factor term and those with
higher order terms represent interaction terms and quadratic
relationship, respectively. It is worthy to note that the responses in Table 1 were count per second (cps). As usual in
BBD optimization, each factor has three levels in: upper (+1),
central (0) and lower points (
1). For simplicity the three levels
have shown in coded number. For example in sorbent volume
in Table 1:
1 is 1.0 mg, 0 is 2.0 mg and +1 is 3.0 mg. As shown
in Table 1, both of coded and un-coded value of each mentioned parameters in 16 experiment runs have shown. Also in
last column of Table 1, cps of each run is represented.
Analysis of variance (ANOVA) is the most efficient parametric
method available for the analysis of data from experiments. It was
devised originally to test the differences between several different
groups of treatments thus circumventing the problem of making
multiple comparisons between the group means using t-tests.
The F value can be used to determine the significance of individual
parameters in a model. The ANOVA result showed that the F value
of 27.70 is significant, and there is only 0.03% likelihood that large
model F value could occur by chance. Values of p (Prob > F) less

Copyright 2014 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/xrs

19

the response and checking the adequacy of the model. Using

the mathematical model, the levels of the variables giving maximum response can then be calculated. In the present work, in
order to investigate the interaction between variables a Box
Behnken design (BBD) was performed which is a response surface methodology (RSM) based on a highly fractionalized
three-level factorial design. The number of experiments suggested by this is calculated according to the equation 2 K
(K
1) + Co, where K is the number of variables and Co is the
number of center points.[34] In this study, K and Co were set at
3 and 4, respectively, which meant 16 experiments had to be
done. The results of the BBD are presented in Table 1. This design is suitable for exploration of quadratic response surface

(2)

M. Moradi et al.
Table 1. BoxBehnken response surface design and responses value
Run

X1

X2

X3

Sorbent amount (mg)

Sample volume (ml)

Extraction time (min)

cps

1
2
3(C)
4
5(C)
6
7
8
9(C)
10
11
12
13
14(C)
15
16

1
1
0
1
0
0
1
0
0

1

1
0
1
0
0

1

1

1
0
0
0
1
0

1
0
0
0

1
1
0
1
1

0
0
0
1
0
1

1
1
0
1

1

1
0
0

1
0

1.0
3.0
2.0
3.0
2.0
2.0
3.0
2.0
2.0
1.0
1.0
2.0
3.0
2.0
2.0
1.0

50.0
50.0
75.0
75.0
75.0
100.0
75.0
50.0
75.0
75.0
75.0
50.0
100.0
75.0
100.0
100.0

10.0
10.0
10.0
15.0
10.0
15.0
5.0
15.0
10.0
15.0
5.0
5.0
10.0
10.0
5.0
10.0

1.84
3.32
4.02
3.50
3.95
3.21
2.88
2.03
3.91
1.96
2.07
2.11
3.02
4.10
2.55
2.51

(C)Center point

Figure 4. Goodness-of-fit of empirical model.

20

than 0.0500 indicate model terms are significant. It was also shown
that GO amount (p = 0.0002), sample volume (p = 0.0117) and the
interactions between GO amount and sample volume (p = 0.0485)
have a significant effect upon the extraction.
As can be seen, all of the quadratic components of have the significant effect (p < 0.0001) on the extraction efficiency and the related RSM behavior agrees with the literature.[35] Quadratic
component is a non-linear relation between factors and response.
The effect of quadratic components is very complicated, and up
to now the reasons of their effect are not clear.
The response equation fitted the experimental data with R2 of
0.9765, indicating 97.65% of the variability in the response. The
goodness-of-fit of the model to the experimental data shown in
Fig. 4 has an adjusted R2-value of 0.9412. According to Joglekar
and May,[36] R2 should be at least 0.80 for a good fit of a model.
As is observed, the coefficient of determination, R2, was more than
0.80 which means that the obtained equation is adequate for correlating the experimental results. In this figure, red color is related to
higher cps and blue is related to lower cps. Plotting predicted

wileyonlinelibrary.com/journal/xrs

values versus experimental values in Fig. 4 also shows that all of

the predicted values of RSM model are located in close proximity
to the experimental values. In other words, the relation between
the actual responses of each 16 design runs and predicted response
(using Eqn (2)) is presented. It can be seen actual data in each point
is near to predicted one. This supports the hypothesis that the
model equation, Eqn (2), is sufficient to describe the response.
Three-dimensional response surface curves were plotted to
study the interaction between the two factors selected and to determine the optimum conditions for the maximum cps (Fig. 5).
The coded model was used to generate 3-D diagram and a contour
plot of calculated response surface from the interaction between
volume of sorbent amountsample volume, sorbent amount
extraction time and sample volumeextraction time (Fig. 5AC, respectively) as a function of the three variables with the other parameter being at its constant level. The optimal values of variables
are obtained when moving along the major and minor axis of the
ellipse and the response at the center point results in maximum recovery. According to Fig. 5A, the recovery increases first and then
slightly decreases by increasing the amount of sorbent. Compared
to traditional SPE sorbents, nanoparticle sorbents have large surface area and high adsorption efficiency; therefore, satisfactory results can be achieved by using lower amount of GO sorbents. For
investigating the optimum amount of the adsorbent, different
amounts of the adsorbent (1.03.0 mg) were applied to extract of
vitamin B12 from the sample, and 2.3 mg of the GO was sufficient
f or extraction of the vitamin B12.
The effect of sample volume on the extraction efficiency was
studied by extraction of the analyte from 50 to 100 ml of water samples. The most extraction efficiency of the target analyte was obtained when sample volume was 80 ml (Fig. 5A, B). Higher
volumes could not be processed as the amount of the sorbent in
volume unit of the sample decreased due to the high dispersion
of the sorbent in the aqueous media. Therefore, access of analyte
to the sorbent surface became more difficult. Hence, sample volume of 80 ml was selected for subsequent experiments.
The extraction time profiles were studied by varying the extraction time in the range of 515 min. Due to the shorter diffusion
route for GO, extraction of target analyte can be achieved in shorter

Copyright 2014 John Wiley & Sons, Ltd.

X-Ray Spectrom. 2015, 44, 1623

Sample preparation

Figure 5. Response surfaces for: (A) sample volume-sorbent amount, (B) extraction time-sorbent amount and (C) extraction time-sample volume.

time even for larger volumes of samples. The results are presented
in Fig. 5B, C. It is found that the extraction of elements is maximized
when the stirring time is more than 11 min. So stirring of the solution was done for 11 min at subsequent analysis.
Quantitative analysis
According to the overall results of the optimization study, the following experimental conditions were chosen: sorbent amount,
2.3 mg; sample volume, 80 ml and extraction time, 11 min. In order
to proceed with the current evaluation of the proposed extraction
technique, linearity, LOD, enhancement factor (EF) and repeatability
were investigated under optimized conditions with the standard
solutions of the analyte. The performance of the developed procedure is summarized in Table 2. Calibration curve was plotted using
10 spiking levels of vitamin B12 over a range between 10 and
1000 g l
1 and analyzing each level in triplicates. Good linearity
of response (LDR) was observed in the range of 251000 g l
1 with
a correlation of determination more than 0.998. Figure 6 shows
these calibration curves with their equation. It can be seen that it
is linear. The LOD was obtained from the formula CLOD = 3Sb/m,
where Sb is the standard deviation of ten replicates of blank measurements and m is the slope of the calibration curve. Also, EF, calculated as the ratio of slopes of the calibration curve after
preconcentration of the analyte and direct calibration equation,
was obtained 46. Precision of the method (RSD%) was found to
be less than 8.1% based on four-replicate analysis during 1 day at
a concentration of 100 g l
1.
A comparison between the figures of merit of the proposed
method and some published methods for extraction and

determination of vitamin B12 is summarized in Table 3. Clearly,

the proposed method has a good sensitivity and precision with a
suitable dynamic linear range. Also, the obtained LODs for the analyte by the present method with sample volume of 80 ml are comparable with those obtained by other methods. All of the results
reveal that the new developed method is not only a good sample
preconcentration technique, but also an excellent sample cleanup procedure that can be used for trace analysis of vitamin B12 with
XRF. Also, the proposed GO-based SPE has some advantages in
comparison with other applied extraction methods such as SPE, including lower consumption of adsorbent and organic solvents and
shorter extraction times.
Analysis of real samples
In order to evaluate the applicability of the developed extraction
method to analysis of the vitamin B12 in the real samples with complex matrices, three selected pharmaceutical formulations were extracted and analyzed using the proposed method under the
optimum conditions. Sample preparation for the real samples was
performed according to Section 2.4. The results of extraction of vitamin B12 from multivitamin tablet, effervescent tablet and injection samples are tabulated in Table 4. Results showed that
vitamin B12 was present in the analyzed multivitamin and

Table 2. Figures of merit of the proposed extraction method for

vitamin B12
20.0
0.998
25.01000
46
8.1
Figure 6. Calibration curve of vitamin B12.

X-Ray Spectrom. 2015, 44, 1623

Copyright 2014 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/xrs

21

Limit of detection (microgram/l)

2
R
Linear dynamic range (microgram/l)
Enrichment factor
RSD (n =4)

M. Moradi et al.
Table 3. Comparison of the proposed method with other methods developed for extraction and determination of vitamin B12
Matrix

1

Method

LOD (g l )

Urine
Multivitamin tablet
Food products

LPME/HPLC-UV
SPE/AASb
Immunoaffinity
/HPLC-UV
HPLC-UV
SPE/XRF

Food products
Pharmaceutical
formulations

1

Linear range (g l )

RSD (%)

Ref

90.0
2.7
3.0

4008000
2.7300
10200

4.5
4.0
3.2

[37]

500
20.0

100010 000
25.01000

8.1

[39]

[38]
[17]

This work

Liquid phase microextraction.

Atomic absorption spectroscopy.

Table 4. Determination of vitamin B12 in different real samples using

proposed method
Samples

Added
(g/l) a

Multivitamintablet

Effervescent
tablet
Injection

100

100

100

Determined
amount

Labeled Recovery RSD

amount
(%)
(%)

6.6b

6.0b

4.5

4.9b

310c

5.0b

333c

106

93

98

6.3
7.5
6.9
5.0
3.8

for in sample handling of metal-organic compounds for their detection by XRF. The results further demonstrated that the proposed GO-based extraction method has good precision,
linearity, and accuracy over the investigated concentration
range. Also, the proposed method offers a simple, fast, sensitive
and selective method for extraction and determination of vitamin B12. Furthermore, there is a possibility of extraction of vitamin B12 from large volumes of sample and therefore sample
diluting and decrease of matrix effects are possible.
Acknowledgements
The authors are grateful for the financial support of this work from
the Iran National Science Foundation (INSF) (Tehran, Iran).

The added amount to diluted sample.

g per multivitamin tablet or effervescent tablet.
c
The concentration of B12 in injection.
b

References
effervescent tablets with concentration of 6.6 and 4.9 g per each
tablet, respectively. Also, the method was tested for extraction of vitamin from an injection sample containing mixture of three vitamins of B1, B6 and B12 (containing 333 mg l
1 of vitamin B12). For
this purpose, 200 l of the injection sample was diluted to 80 ml
with deionized water. According to Table 4, the results obtained
by the current method agreed well with the labeled values of vitamin B12 for multivitamin tablet, effervescent tablet and injection
sample. The RSD (%) were in the ranges of 3.87.5%. As we know,
recovery (R%) in extraction methods represents the amount of
extracted analyte from sample to extraction phase. To determination of R%, we can use add-found manner, so that a known concentration of analyte was added into sample and after extraction using
proposed technique, the amount of extracted analyte was
determined. If the recovery of added amount is in the range of
90110%, the proposed method is validated. As can be seen in
Table 4, recovery values of real samples are in the range of 93 to
106%; thus, proposed extraction method prior to XRF analysis is
suitable method for vitamin B12 determination.

Conclusions

22

Liquid samples usually provide a high X-ray scatter background

resulting in poor signal-to-noise ratio. To overcome this problem, many different chemical and physical preconcentration
techniques have been proposed prior to XRF analysis. The
extraction of vitamin B12, which contains a cobalt atom in the
molecule, from a liquid onto a graphene oxide is a new idea

wileyonlinelibrary.com/journal/xrs

[1] S. S. Kumar, R. S. Chouhan, M. S. Thakur. Anal. Biochem. 2010, 398, 139.

[2] M. A. Bernard, P. A. Nakonezny, T. M. Kashner. J. Am. Geriatr. Soc. 1998,
46, 1199.
[3] Y. Yamini, E. Tahmasebi, L. Ranjbar. Biol. Trace Elem. Res. 2012, 147, 378.
[4] R. C. Oh, L. D. Brown. Am. Fam. Physician 2003, 67, 979.
[5] B. P. Kelleher, J. M. Scott, S. D. O. Brain. Med. Lab. Sci. 1990, 47, 90.
[6] P. J. Casey, K. R. Speckman, F. Ebert, W. E. Hobbs. J. Assoc. Off. Anal.
Chem. 1982, 65, 85.
[7] K. Muhammed, D. Briggs, G. Jones. Food Chem. 1993, 48, 431.
[8] M. L. Marszall, A. Lebiedzinska, W. Czamowski, P. Szefer. J. Chromatogr.
A 2005, 1094, 91.
[9] A. Zafra-Gomez, A. Garballo, J. C. Morales, L. E. Garcia-Ayuso. J. Agric.
Food Chem. 2006, 54, 4531.
[10] P. Vinas, N. Campillo, I. Lopez-Garcia, M. Hernandez-Cordoba. Anal.
Chim. Acta 1996, 318, 319.
[11] H. B. Li, F. Chen, Y. Jiang. J. Chromatogr. A 2000, 891, 243.
[12] C. Pakinm, M. Bergaentzl, D. Aoud-Werner, C. Hasselmann.
J. Chromatogr. A 2005, 1081, 182.
[13] E. G. Yanes, N. J. Miller-Ihli. Spectrochim. Acta B 2004, 59, 891.
[14] A. Makarov, J. Szpunar. J. Anal. At. Spectrom. 1999, 14, 1323.
[15] H. Chassaigne, R. Lobinski. Anal. Chim. Acta 1998, 359, 227.
[16] G. Yaling, G. Fei, G. Sumangala, C. Andrew, S. Qinghai, G. Mingruo. Food
Chem. 2008, 110, 769.
[17] O. Heudi, T. Kilinic, P. Fontannaz, E. Marley. J. Chromatogr. A 2006,
1101, 63.
[18] E. Margu, I. Queralt, M. Hidalgo. Trend Anal. Chem. 2009, 28, 362.
[19] R. A. Silva, A. N. Masi. X-ray Spectrom. 2007, 36, 279.
[20] E. Margu, B. Zawisza, R. Sitko. Trend Anal. Chem. 2014, 53, 73.
[21] V. Cojocaru, S. Spiridon. J. Radional. Nucl. Chem. 1993, 170, 259.
[22] M. S. Carvalho, M. de Lourdes, F. Domingues, J. L. Mantovano,
E. Q. S. Filho. Spectrochim. Acta B 1998, 53, 1945.
[23] W. Abe, S. Isaka, Y. Koike, K. Nakano, K. Fujita, T. Nakamura. X-Ray
Spectrom. 2006, 35, 184.
[24] T. Sato, T. Nakamura. Hydrometallurgy 1980, 6, 3.

Copyright 2014 John Wiley & Sons, Ltd.

X-Ray Spectrom. 2015, 44, 1623

Sample preparation
[25] E. Margu, C. Fonts, K. Van Meel, M. Hidalgo, I. Queralt. Spectroscopy
Europe 2008, 20, 11.
[26] Y. Yamini, N. Amiri, M. Karimi. X-Ray Spectrom. 2009, 38, 474.
[27] B. Zawisza, R. Sitko, E. Malicka, E. Talik. Anal. Methods 2013, 5, 6425.
[28] E. Margu, R. Van Grieken, C. Fonts, M. Hidalgo, I. Queralt. Appl. Spectro.
Rev. 2010, 45, 179.
[29] W. S. Hummers Jr., R. E. Offeman. J. Am. Chem. Soc. 1958, 80, 1339.
[30] M. J. Allen, V. C. Tung, R. B. Kaner. Chem. Rev. 2009, 110, 132.
[31] J. W. Lee, A. S. Hall, J. D. Kim, T. E. Mallouk. Chem. Mater. 2012, 24, 1158.

[32] Q. Tu, L. Pang, Y. Chen, Y. Zhang, R. Zhang, B. Lu, J. Wang. Analyst 2014,
139, 105.
[33] http://www.chemicalize.org/structure/. Retrieved 18 January 2014.
[34] D. Bingol, M. Kulcu. Analyst 2011, 136, 4036.
[35] J. Xia, B. Xiang, W. Zhang. Anal. Chim. Acta 2008, 625, 28.
[36] A. M. Joglekar, A. T. May. Cereal Food World 1987, 32, 857.
[37] P. Berton, R. P. Monasterio, R. G. Wuilloud. Talanta 2010, 97, 521.
[38] M. R. Hadjmohammadi, V. Sharifi. J. Food Drug Anal 2007, 15, 285.
[39] L. S. S. Kumar, M. S. Thakur. Anal. Biochem. 2011, 418, 238.

23

X-Ray Spectrom. 2015, 44, 1623

Copyright 2014 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/xrs