American Journal of Anal yt ical Chemistry, 2011, 2, 612-618
doi:10.4236/ajac.2011.25069 Published Online September 2011 (http://www.SciRP.org/journal/ajac)
Determination of Melamine in Fresh Milk by
Electrochemistry with Solid Phase Microextraction
at Bismuthyl Chloride Modified Graphite Epoxy
Composite Electrode
Yongchun Zhu1*, Yanjia Zhang1, Jingyi Li2, Yanli Han1, Guobin Dong1, Hongbo Zhang1
1College of Chemistry and Life Science, Shenyang Normal University, Shenyang, China
2Foreign Language Department, Sheny ang Agricult ural University, Shenyang, China
Received April 28, 2011; revised May 3, 2011; accepted June 7, 2011
Abstract
Melamine as an important chemical raw material and a harmful additive in foods has attracted many people’s
attention. In the present paper, The graphite-epoxy composited solid phase electrode was modified with bis-
muth layer by cyclic voltammetric deposition of bismuth from Bi(NO3)3 aqueous solution including 0.10 M
HNO3, and hydrolyzed into micro bismuthyl chloride on-sites. Melamine in fresh milk was extracted with
solid phase micro-extraction on the bismuthyl chloride modified graphite-epoxy composited solid electrode.
The adsorption of melamine on bismuthyl chloride particle surfaces follows a Freundlich adsorption model,
and results in the decrease of the reduction peak current of bismuth in bismuthyl chloride, and determined by
differential pulse voltammetry from fresh milk in a larger concentration range of 10–4  10–12 M with detec-
tion limit of 2.5  10–12 M and relative standard deviation of 2.7%. The method is sensitive, convenient and
was applied in the detection of melamine in fresh milk with relative deviation of 4.2% in content of 0.45
mg/kg melamine in the fresh milk.
Keywords: Melamine, Bismuthyl Chloride Modified Electrode, Adsorption, Differential Pulse Voltammetry,
Solid Phase Micro Extraction
1. Introduction
Melamine (2,4,6-triamino-1,3,5-triazine), as an important
chemical raw material, has been widely used in the pro-
duction of plastics[1-4], paper finishers [5], flame retar-
dant [6], and wrinkle-free textiles[7] in the manufacture
as well as nano materials [8-10]. Recent case of the dis-
covery of melamine as one of the harmful additives in
pet food, animal feed, milk and protein sources including
wheat gluten, rice protein concentrate, and corn gluten
results in an urgent need for rapid detection methods of
melamine in foods [11-13]. The common methods for
detection of melamine have been recommended as GC/
MS [14], LC/MS [15,16], as well as the specific anti-
gen-antibody reaction and immune chromatography
analysis technology [17,18]. All these methods are in-
struments expensive and time consuming. Electrochemi-
cal methods are not favorable for detection of melamine
due to its electrochemical inertness. So to find out a
suitable chemical modified electrode for extraction, pre-
concentration and detection of melamine is the key point
in electrochemical detection of melamine. Melamine as
an aromatic polyamine molecule can interact with ploy-
hydroxides or polyphenol molecules, such as bismuth
oxides, bismuthyl chloride molecules through the hy-
drogen bond formation. Electrochemical deposited bis-
muth electrode has been used conveniently in electro-
chemical analysis [19,20]. The deposited bismuth hy-
droxides or bismuthyl chloride may be favourable in the
extraction and electrochemical detection of melamine.
In the present paper, the bismuthyl chloride film mo-
dified solid phase graphite-epoxy carbon paste electrode
was prepared for the electrochemical solid phase micro
extraction. Melamine from fresh milk was electrochemi-
cally determined, and reported here.

Y. C. ZHU ET AL.
613
2. Experimental Section
2.1. Instrumentals
The electrochemical experiments were carried out on an
Electrochemical Analyzer (model CHI620, USA) with
three-electrode system, a piece of platinum wire as ac-
count electrode, a home-made solid state graphite-epoxy
composited electrode as the working electrode, a KCl
saturated calomel electrode (SCE) as the reference elec-
trode. All potentials reported here were respect to this
reference electrode.
2.2. Reagents
Bismuth nitrate was prepared into 1.0mM stock solution
for the modification of electrode. Potassium chloride was
prepared into 1.0 M stock solution as the electrolyte so-
lution. Disodium hydrogen phosphate and citric acid
were prepared into 0.2 mol/L for pH buffer solution in
the range of 2.2 - 8.0. All chemicals were analytical pure,
and purchased from Shenyang Chemical Co. All solu-
tions were prepared with ultrahigh pure water (18.2 MΩ,
from Milli-QA-10, Millipore Corporation, Billerica,
USA). All electrolyte solution was deaerated with high
purity nitrogen gas to remove oxygen prior to use.
2.3. The Preparation of Basic Electrode
The basic electrode was prepared by mixing Graphite
power (200#, spectral pure), epoxy resin and polyamide
resin into a paste with a weight ratio of 8:2:2, tightly
pressed into a clean glass tube (inner diameter is about 4
mm) with a copper wire at the other end as an electrode
lead, and solidified in air for 72 hours. The prepared
graphite-epoxy composite electrode (GECE) was pol-
ished on sandpapers (400#, 600#, 1000#) successively,
and then polished with glassy paper into mirror surface,
as the basic electrode.
2.4. The Modification of the GECE with
Bismuthyl Chloride
The prepared GECE was set up in 1.0 mM Bismuth ni-
trate solution with 1.0 M nitric acid as electrolyte and pH
buffer solution. Cyclic voltametric experiments were
performed in the potential range of 0.0 - 1.0 V at 0.1 V/s
scan rate in 1000 cycles for the deposition of bismuth on
the electrode surface. After the electrochemical deposi-
tion, the GECE surface was uniformly covered with
bismuth layer. The bismuth modified GECE was washed
with Ultrapure water, set into 0.50 M KCl electrolyte
solution, and perform the CV experiments in the poten-
tial range of –0.2 - 1.6 V at 0.1 V/s scan rate in 100 cy-
cles for the oxidation of bismuth layer. During the oxida-
tion process, the deposited bismuth layer was oxidized
into bismuth ion, hydrolyzed into bismuth hydroxide and
transformed into a bismuthyl chloride film in-situ as the
working electrode used in the following experiments.
3. Results and Discussion
3.1. The Electrochemical Behavior of the
Modified Electrodes
The Bismuth modified GECE electrode in 0.50 M KCl
electrolyte solution gives a CV curve with an irreversible
reduction peak at –0.65 V and an irreversible oxidation
peak at 0.30 V as shown in curve 2 of Figure 1(a). After
the transformation of modified electrode from bismuth
layer to bismuthyl chloride layer, the modified electrode
gives a typical CV curve as shown in curve 1 of Figure
1(b). The irreversible reduction peak shifts to –1.42V
without oxidation peak. The reduction peak potential
difference of 0.768 V between Bi3+ and BiOCl was used
to calculate the solubility product constant of BiOCl as
1.8  10–31 [21]，which indicates the bismuthyl chloride
is very stable. The bismuthyl chloride modified GECE
was set into 0.50M KCl electrolyte solution including 5.0
 10–4 M melamine for 20 min, and then performed cy-
clic voltammetric experiment. A typical CV curve with
the one reduction peak located at –1.44 V was obtained
as shown in curve 2 of Figure 1(b). The further negative
shift of the reduction peak potential indicates the bond-
ing between bismuthyl and melamine is more stable than
that of bismuthyl chloride, which is the basic of the solid
phase extraction.
The cyclic voltammetric experiments were performed
at different scan rates, the obtained reduction peak cur-
rent was plotted against scan rate, and a straight line was
obtained with regression equation of,

1
pc 10.18 684.2.iA vVs


 (1)
The correlation coefficient and standard deviation of
the regression were 0.9968 and 1.885, respectively. This
relationship indicates that the electrochemical reaction is
a surface controlled process, and the reduction reaction
occurs at the electrode surface. The all process of the
electrode reactions can be summarized in the following
equation.
Bi
-3e, -0.3 V
+3e,-0.64 V
Bi
3+
+Cl
-
,2OH
-
BiOCl
ESPME melemine
BiO-melemine +Cl
-
+3e, -1.42 V
Bi
3+
(2)

Y. C. ZHU ET AL.
614
(a)
(b)
Figure 1. (a) CV curves of GECE before the bismuth deposited (1) and after the Bismuth deposited (2) in nitric acid solution.
(b) CV curves of bismuth modified GECE in 0.5 M KCl solution(1) and including 5.0  10–4 M melamine (2). scan rate: 0.05
V/s.
3.2 The Effect of Deposition Amount of Bismuth
The deposition amount of bismuth is very important for
the modification of the GECE surface and extraction of
melamine. In the cyclic voltammetric deposition process,
at the given scan rate (0.10 V/s) and potential range of
–0.20 - –1.6V, the amount of the deposited bismuth and
the effective area of the deposited layer are related to the
number of cycles. In order to obtain a bismuthyl chloride
layer with larger effective area for extraction of mela-
mine, the modified electrode was checked with cyclic
voltammetry in 0.50 M KCl electrode solution (pH 6.0)
after modification. The reduction peak current of bis-
muthyl chloride at –1.42 V plotting against number of
scanning cycles (n) in the deposition step, a one-peak
curve with the peak point located at 2000 cycles was
obtained as shown in Figure 2. The curve can be re-
gressed into a Gaussian function of,

pc 6
2323
155.7exp 1.3396 10
n
iA 








 (3)
The correlation coefficient and standard deviation of
the regression were 0.9812 and 12.9, respectively. This
result indicates the effective area of electrode surface for
the formation of effective bismuthyl chloride layer and
extraction of melamine increases with the increase of
scanning cycles at the beginning of the deposition, but
after about 2000 cycles, but it is reduced after more and
more bismuth deposited from solution. So 2000 cycles
was set as the optimal deposition condition.
Figure 2. The relationshp between reduction peak current
and number of CV cycles. The experimental conditions
were the same as those in Figure 2.

Y. C. ZHU ET AL.
615
3.3 The Effect of Solution pH
Melamine is a weak acid with pka of 5.35 [22], and bis-
muthyl chloride is transferred from bismuth hydroxide,
So solution pH is very important for the formation of
bismuthyl chloride layer and extraction of melamine.
The solution pH was controlled by disodium hydrogen
phosphate-citric acid buffer system in the range of 4.0 -
8.0. The PDV experiments were performed on the bis-
muthyl chloride modified electrode in the solutions with
different pH after setting the electrode in the solution for
4mins and at initial potential of –0.6 V for 100 s. The
obtained reduction peak current at –1.32 V plotted ag-
ainst solution pH is one-peak curve with the maximum
point at pH = 6.0 as shown in Figure 3.
These results indicate that in the case of pH = pKa,
melamine is suitable for the interaction with bismuthyl
chloride by hydrogen bonding. Bismuthyl chloride is
come from bismuth hydroxide, which has a solubility
product constant of 4.0  10–31, pH6.0 is also included
enough concentration of hydroxide ion for the formation
of bismuth hydroxide, and stabilized the bismuthyl chlo-
ride. So the optimal solution pH was chosen as pH = 6.0.
3.4. The Effect of Quiet Time
Quiet time is the period time applying the initial potential
to the electrodes before potential scanning, which can be
used to test the influence of initial potential on extraction
of melamine. The electrodes were set into 0.5 M KCl
solution including 5.0  10–4mM melamine, at initial
potential of –0.6V for different quiet time, and then PDV
experiments were performed in the potential range of
–0.60 V - –1.50 V. The obtained reduction peak current
at –1.32 V plotting against quiet time, a curve was ob-
tained as shown in Figure 4. The first part of the curve in
quiet time range of 0 - 40 s was a straight line with the
regression equation of,
Figure 3. The relationship between reduction peak current
and solution pH.
Figure 4. The relationship between reduction peak current
and quiet time.
pc 117.3 1.608,
R0.9974, SD0.7255
iA ts




(4)
The correlation coefficient and standard deviation of
the regression were 0.9974 and 0.7255, respectively. The
rest part of the curve in quiet time range of 40 - 170 s
was a S-shaped curve, and can be regressed as a Re-
chards model with an equation of,

pc 0.041
120.0
1exp 12.080.082
iA ts
 





(5)
The correlation coefficient and standard deviation of
the regression were 0.9958 and 1.29, respectively, and
the inflection point located at 120 s. This relationship
indicates the process is greatly influenced by initial po-
tential due to the initial state of bismuthyl chloride and
its interaction with melamine. With the increase of quiet
time, the extraction amount of melamine increases so the
reduction peak current decreases. The initial potential
–0.6 V is also the potential for bismuth ion pre-reduced
into bismuth, which can increase the reduction peak cur-
rent without influence on melamine extraction. So in
practice, the optimal quiet time was chosen at the inflec-
tion point of 120 s.
3.5. The Effect of Accumulation Time
Accumulation time is the period time of contacting elec-
trode to melamine aqueous solution before the beginning
of the electrochemical experiment. Accumulation time is
also the extraction time without influence of initial po-
tential. The electrodes were set into 0.50 M KCl solution
including 5.0  10–4 M melamine for different accumula-
tion time, and then PDV experiments at –0.60 V initial
potential and quiet time of 120s were performed in the
potential range of –0.60 V- –1.50 V. The obtained reduc-
tion peak current at –1.32 V was plotted against accu-

616 Y. C. ZHU ET AL.
mulation time, a curve was obtained as shown in Figure
5.
The curve was regressed as a Boltzman function
model with an equation of,
 
pc
accum
147.3
82.0 1expmin 3.20.381 iA t  


 
(6)
The correlation coefficient and standard deviation of
the regression were 0.9995 and 2.25, respectively. This
relation indicates that the melamine molecules were ex-
tracted on bismuthyl chloride surface, and decreased the
reduction peak current of bismuthyl chloride with a in-
flection point at 3.2 min. So in practice, the accumulation
time of 4 min is enough for the reaction to reach the
saturated state.
3.6. The Effect of Melamine Concentration
Under the optimal experimental conditions such as solu-
tion pH 6.0, initial potential of –0.60 V, quiet time of 120
s, accumulation time 4 min, PDV experiments were per-
formed at different concentration of melamine. The ob-
tained reduction peak current at –1.32V sharply decreases
with the increase of melamine concentration known as the
isothermal adsorption curve shown in Figure 6.
Figure 5. The relationship between reduction peak current
and accumulation time.
Figure 6. The relationship between reduction peak current
and melamine concentration.
The curve was regressed as a logarithm model with an
equation of,

pc 88.948.042 logMiA c
 

(7)
The correlation coefficient and standard deviation of
the regression were 0.9925 and 3.09, respectively. The
logarithm of reduction peak current, ipc, against the loga-
rithm of melamine concentration was a linear line with a
regression equation of



pc
log1.9920.0236logM ,
0.989; 0.011
iA c
RSD
 
 (8)
This relation indicates the adsorption of melamine
molecule on the electrode surface follows Freundlich
adsorption model [23] with the adsorption equilibrium
constant of K = 98.2, and n = 42.4, which is a favourable
adsorption process.
3.7. The Standard Curve and Deviation of the
Method
In order to determine the concentration of melamine in
aqueous solution, the Equation (7) was also served as the
standard working curve. The detection limit （in the ratio
of signal to noise 3:1）was calculated as 2.5  10–12 M
with the relative standard deviation of 2.7% in the con-
centration range of 2.5  10–12 – 2.5  10–4 M.
3.8. The Detection of Melamine in Fresh Milk
The Huishan fresh milk (made in Shenyang Huishan
milk Co.) was bought from supermarket. The test solu-
tion was the mixture of 5.0 mL fresh milk, 5 mL diso-
dium hydrogen phosphate-citric acid buffer solution (pH
6.0) and 10.0 mL KCl electrolyte solution. The PDV
experiments were carried out under the optimal condi-
tions: solution pH 6.0, the accumulation time of 4 min,
and initial potential of –0.60 V for 50 s. the experiments
were parallel for 5 times, the average reduction current at
–1.32 V was obtained as 1.626 A with the relative de-
viation of 4.2%. The amount of melamine in the fresh
milk was calculated as 0.45 mg/kg. According to the
national standard, the melamine in fresh milk is less than
1 mg/kg. So the tested fresh milk sample met the re-
quirement standard.
4. Conclusions
As a summary of this work, a graphite-epoxy composite
electrode was modified electrochemically by bismuthyl
chloride in-situ. The bismuthyl chloride layer serves as a
solid phase micro extractant, extracts melamine from
aqueous solution with the help of electrochemistry. The

Y. C. ZHU ET AL.
617
adsorption of melamine on bismuthyl chloride flows
Freundlich adsorption model, and results in the decrease
of reduction current of bismuthyl chloride, and used for
the determination of melamine in the range of 10–4 
10–12 M with detection limit of 2.5  10–12 M and relative
standard deviation of 2.7%.This method offers new way
for the electrochemical detection of electrochemically
inactive species.
5. Acknowledgements
The author would like to acknowledge the financial sup-
ports of the Chinese National Science Foundation
(20875063), Liaoning education minister（2004-c022）
and national key. Laboratory on electroanalytical chem-
istry (2006-06), Shenyang Sciences and technology bu-
reau foundation (2007-GX-32).
6. References
[1] S. V. Levchik and E. D. Weil, “Flame Retardancy of
Thermoplastic Polyesters: a Review of the Recent Lit-
erature,” Polymer International Vol. 54, No. 1, January
2005, pp. 11-35. doi:org/10.1002/pi.1663
[2] Z. Wirpsza, “Achievements in the Use of Melamine as a
Chemical Raw Material. Part I. General Review and
Melamine-Based Coating Resins and Coatings,” Polimery,
Vol. 41, No. 5, June 1996, pp. 265-276.
[3] H. Walch, A. Maier, W. M. Heckl and M. Lackinger,
“Isotopological Supramolecular Networks from Mela-
mine and Fatty Acids,” The Journal of Physical Chemis-
try C, Vol. 113, No. 3, March 2009, pp. 1014-1019.
doi:org/10.1021/jp8078474
[4] M. Henriksson and L. A. Berglund, “Structure and Prop-
erties of Cellulose Nanocomposite Films Containing
Melamine Formaldehyde,” Journal of Applied Polymer
Science, Vol. 106, No. 4, November 2007, pp. 2817-2824.
doi:org/10.1002/app.26946
[5] U. Berger and W.Schnick, “Syntheses, Crystal Structures,
and Vibrational Spectroscopic Properties of MgCN2,
SrCN2, and BaCN2,” Journal of Alloys and Compounds
Vol. 206, No. 2, May 1994,pp. 179-184.
doi:org/10.1016/0925-8388(94)90032-9
[6] M. Groenewolt and M. Antonietti, “Synthesis of g-C3N4
Na-Noparticles in Mesoporous Silica Host Matrices,”
Advanced Materials, Vol. 17, No. 14, July 2005, pp.
1789-1792. doi:org/10.1002/adma.200401756
[7] E. Kroke and M. Schwarz, “Novel Group 14 Nitrides,”
Coordination Chemistry Review, Vol. 248, No.5-6, 2004,
pp. 493-532.
[8] L. Licea-Jiménez and R. W. Rychwalski, “Novel Nano-
composites Using Carbon Nanotubes and Melamine-For-
malde-Hyde,” Materials Science-Poland Vol. 23, No. 4,
January 2005, pp. 909-913.
[9] S. Bajia, R. Sharma and B. Bajia, “Solid-State Micro-
wave Synthesis of Melamine-Formaldehyde Resin,” Solid
State Microwave Synthesis, Vol. 6, No. 1, January 2009,
pp. 120-124.
[10] A. K. Hau, T. H. Kwan and P. K. Li, “Melamine Toxicity
and the Kidney,” Journal of American Society of Ne-
phrology, Vol. 20, No.2, February 2009, 245-250.
doi:org/10.1681/ASN.2008101065
[11] D. Meter and R. J. Hemley, “Low-Compressibility Car-
bon Nitrides,” Science, Vol. 271, No. 5245, January 1996,
pp. 53-55.
[12] S. Ehling, S. Tefera and I. Pho, “Rapid Nondestructive
Screening for Melamine in Dried Milk by Raman Spec-
troscop,” Food Additives and Contaminations, Vol. 24,
No. 12, November 2007, pp. 1319-1325.
[13] S. Mfiligenzi, B. Puschner and L. S. Aston, “Diagnostic
Determination of Melamine and Related Compounds in
Kidney Tissue by Liquid Chromatography/Tandem Mass
Spectrometry,” Journal of Agricultural Food Chemistry,
Vol. 56, No. 17, September 2008, pp. 7593-7599.
doi:org/10.1021/jf801008s
[14] J. J. Litzau, G. E. Mercer and K. J. Mulligan, “GC-MS
Screen for the Presence of Melamine, Ammeline, Am-
melide and Cyanuric Acid,” US FDA Laboratory Infor-
mation Bulletin, No. 4423, October 2008.
http://www.cfsan.fda.gov/~frf/ lib4423.html.
[15] M. Smoker and A. J. Krynitsky, “Interim Method for
Detemination of Melamine and Cyanuric Acid in Food
Using LC/MS-MS,” Version 1.0. US FDA Laboratory
Information Bulletin, No. 4422, October 2008.
http://www.cfsan.fda.gov/~frf/lib4422.html.
[16] S. Turnipseed, C. Casey, C. Nochetto and D. N. Heller,
“Determination of Melamine and Cyanuric Acid Resi-
dues in Infant Formula Using LC-MS/MS,” US FDA
Laboratory Information Bulletin, No. 4421, October 2008.
http://www.cfsan.fda.gov/~frf/lib4421.html.
[17] J. V. Sancho, M. Ibanez and S. Grimalt, “Residue Deter-
mination of Cyromazine and Its Metabolite Melamine In
Chard Samples by Ion-Pair Liquid Chromatography Cou-
pled to Electrospray Tandem Mass Spectrometry,” Anay-
ticall Chimica Acta, Vol. 530, No. 2, February 2005, pp.
237-243.
[18] H. Ogasawara, K. Imaida and H. Ishiwata, “Urinary Blad-
der Carcinogenesis Induced by Melamine in F344 Male
Rats: Correlation between Carcinogenicity and Urolith
formation,” Carcinogenesis, Vol. 16, No. 11, November
1995, pp. 2773-2777.
[19] N. Serrano, N. Martin, J. Diaz-Cruz, C. Arino and M.
Esteban, “Bismuth Film Electrode in Metal Complexation
Studies: Stripping Analysis of the Pb(II)-, Cd(II)-, and
Zn(II)- Binding with Phthalate,” Electroanalysis, Vol. 21,
No. 3-5, February 2009, pp. 431-438.
[20] F. Torma, A. Grun, I. Bitter and K. Toth, “Calixarene/
Nafion-Modified Bismuth-Film Electrodes for Adsorp-
tive Stripping Voltammetric Determination of Lead,”
Electroanalysis, Vol. 21, No. 17-18, September 2009, pp.
1961-1969.
[21] Solubility Products at 25˚C, 11/28/2007
http://www.csudh.edu/Oliver/chemdata/data-ksp.htm.
[22] U. S Food and Drug Administration, “Interim Melamine
and Analogues Safety/Risk Assessment,” 05/22/2009
http://www.fda.gov/ScienceResearch/SpecialToics/PeerRev
iew of Scientific Information and Assessments/ucm 155012.

Y. C. ZHU ET AL.
618
htm.
[23] J. T. Nwabanne and P. K. Igbokwe, “Kinetics and Equi-
librium Modeling of Nickel Adsorption by Cassava
Peel,” Journal of Engineering Applied Science, Vol. 3,
No. 11, November 2008, pp. 829-834.