International Dairy Journal 44 (2015) 1e5

Contents lists available at ScienceDirect

International Dairy Journal

journal homepage: www.elsevier.com/locate/idairyj

Novel method for the determination of tetracycline antibiotics in

bovine milk based on digital-image-based colorimetry
Rungrat Urapen a, Prinya Masawat b, *
a
Department of Chemistry and Centre of Excellence for Innovation in Chemistry, Faculty of Science, Mahidol University, Bangkok, Thailand
b
Naresuan University, Faculty of Science, Department of Chemistry, 65000 Phitsanulok, Thailand

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

Article history: Solid-phase extraction was used to increase the intrinsic colour of tetracycline (TC), so that colorimetry
Received 10 September 2014 could be employed as an analytical method for the detection of TC in milk. Digital-image-based color-
Received in revised form imetry, a measurement based on capturing colour of solutions by a webcam camera, was coupled with an
17 November 2014
articial neural network learning model for quantitative chemical analysis. The red, green, and blue
Accepted 1 December 2014
Available online 19 December 2014
values were used for data processing and were calibrated using values of standard solutions (from a
database). The results were obtained directly as TC concentration. The colour intensity of the extracts was
tested for accuracy using samples spiked with three TC concentrations. The average percentage re-
coveries were favourable in all samples and the relative accuracy of the method was between 91.9% and
105.0%. The method was validated by comparing the results with measurements performed using
ultra-violet/visible spectrophotometry. There were no signicant difference observed between the two
methods.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Tetracycline antibiotics (oxytetracycline, OTC, tetracycline, TC,
chlortetracycline, CTC, and doxycycline, DTC) are broad-spectrum
antibiotics that are active against both Gram-positive and
Gram-negative bacteria. In dairy farming, TCs are widely used as
feed supplements to prevent or treat mastitis and metritis in cows
(Jian, James, & Jack, 2012). Due to their cost-effectiveness, good oral
absorption, and broad-spectrum activity, TCs are widely used as
veterinary drugs. The rate of metabolism of TCs has been estimated
to be 25e75% in dairy cows, and a signicant amount of TCs is
transferred into bovine milk; consequently, they can affect human
consumers. The adverse effects on consumers include liver damage,
possible allergic reactions, yellowing of teeth, and gastrointestinal
disturbance (Jose, Johanan, Karina, Israel, & Jose, 2010). To prevent
any health problems in consumers, the European Union (EU) and
Food and Agriculture Organization of the United Nations (FAO)/
World Health Organization (WHO) have established a maximum
residue limit (MRL) in milk of 100 mg L
1 for TC, OTC, and CTC (EC,
* Corresponding author. Tel.: 66 55963439.
E-mail address: prinyam@nu.ac.th (P. Masawat).
1999; WHO, 1990). The US Food and Drug Administration (FDA) has
set the tolerance of TC, OTC, and CTC in milk of 300 mg L
1 (FDA,
2005).
The residues of TCs in milk samples have been determined using
various techniques such as immunoassays, high performance liquid
chromatography (HPLC), capillary electrophoresis, ow
injection analysis, electrochemistry, uorescence spectrometry,
and mass spectrometry. Electrochemistry was found to be a
suitable technique, which could be performed without sample
pre-concentration, when combined with immunosensors that
exhibit important advantages such as specicity, and low limit of
detection in TC analysis (Felipe et al., 2013, 2014). Nevertheless,
there are some disadvantages, including a complicated and a time-
consuming procedure. Thus, HPLC is still useful for sensitive, ac-
curate and highly specic TC analysis (Zhao, Zhang, & Gan, 2004).
HPLC equipment and various components are often expensive,
leading to high purchase and maintenance costs. In addition, this
technique requires large amounts of solvents that may be toxic, and
it is time-consuming.
Consequently, digital-image-based colorimetry (DIC) was
chosen instead of HPLC for the analysis of tetracycline residues in
milk samples. DIC is a technique that relies on the solution colour

http://dx.doi.org/10.1016/j.idairyj.2014.12.002
0958-6946/ 2014 Elsevier Ltd. All rights reserved.
2 R. Urapen, P. Masawat / International Dairy Journal 44 (2015) 1e5
photographed by a webcam camera with charge-coupled devices
(CCDs) sensor (Ste
fani et al., 2013), a digital camera with comple-
mentary metal-oxide semiconductor (CMOS) sensor (Aree &
Niamh, 2011), and a mobile phone with optical sensor (Sarun,
Kosom, & Yuttana, 2014).
Although there are various methods of image data processing
obtained from DIC, an articial neural network (ANN) was chosen
in this work. An ANN learning model is a computational model for
information processing based on the structure and functions of
biological neural networks. Information processing in the ANN
happens through neural network changes, or learning, in a sense,
based on the ow of information from input to output through the
network structure of ANN. The ANN method is widely used in
prediction and classication based on learning from examples
(Durairaj & Thamilselvan, 2013).
Compared with HPLC, DIC-ANN offers many advantages,
including decreased time for analysis (high speed), decreased sol-
vent use, low costs, and no need for a standard curve in every
analysis. Nevertheless, both HPLC and DIC-ANN need sample
preparation, such as solid-phase extraction (SPE), before analysis.
SPE has been applied for isolation of TCs from complex matrices
(Hendrik & Herman, 2007; Luciana, Alejandro, & He
ctor, 2010;
Wang, Yang, & Cheng, 2007).
The objective of this study was to develop the SPE process for
preconcentration, clean-up and increasing colour of TC for milk
samples before analysis by DIC-ANN, to provide a simple and fast
method for routine regulatory analysis of TC residues in bovine
milk. Finally, the results from DIC-ANN were compared with those
from UV-visible spectrophotometry.
2. Materials and methods
2.1. Chemicals and reagents
A standard stock solution (1000 mg mL
1) of TC (Merck, Darm-
stadt, Germany) was prepared by dissolving TC hydrochloride in
methanol (RCL Labscan, Bangkok, Thailand), and stored in dark
glass bottles. The working standard solutions were prepared
weekly by dilution of the stock solution with 5.0 M hydrochloric
acid in methanol to produce calibration standards of concentration
1.0, 3.0, 5.0, 7.0, and 9.0 mg mL
1.
The trichloroacetic acid (TCA; Fisher Scientic UK Ltd., Lough-
borough, Leicestershire, UK) solution was prepared in deionised
water (ELGA Ltd., High Wycombe, Buckinghamshire, England). This
solution was used for sample preparation to remove the proteins in
the bovine milk samples.
2.2. Milk sample preparation
Milk samples were randomly collected from supermarkets in
Phitsanulok, Thailand. The samples were classied into 3 types;
UHT milk (ultra-high temperature milk or ultra-heat treated milk)
(three different brands), pasteurised milk (three different brands),
and sterilized milk (one brand).
Ten boxes of each brand of each type of milk were mixed
together and a 5.0 mL sample was transferred into a 15.0 mL
centrifuge tube. Five millilitres of 10.0% (v/v) trichloroacetic acid
were added to the milk sample followed by shaking for 5 min and
centrifugation at (2150 g) for 10 min. The supernatant was ltered
through a 0.45 mm nylon membrane and passed through a
SPE cartridge (VertiPak C18, 6.0 mL, 500 mg Vertical Chromatog-
raphy Ltd., Nonthaburi, Thailand) which was conditioned by
washing sequentially with 4.0 mL of methanol, 4.0 mL of water, and
4.0 mL of 10.0% (v/v) trichloroacetic acid. The sample extract was
passed through the C18 SPE tube at an average ow rate of
1.0 mL min
1. The column was dried while drawing full vacuum
after washing with 4.0 mL of water. The TC residue was gravity-
eluted with methanol, and the extract was evaporated under a
ow of nitrogen to 2.5 mL and measured with an UV-visible
spectrophotometer and DIC-ANN.
2.3. Digital-image-based colorimetry
DIC using a CMOS camera as a detector [Microsoft (Thailand)
Ltd., Bangkok, Thailand] coupled with back propagation neural
network (BPNN; Nontawat, Yuthapong, & Prinya, 2013) was applied
for the analysis of TC residue in milk samples. This method was
based on the red, green and blue (RGB) values of different colour
intensities of sample and standard solutions. The coloured solu-
tions were photographed with a webcam camera at a resolution of
1280  720 pixels. The brightness of the illuminating white LEDs
was 300 lux, which is the optimum condition for photography. The
colour pictures of TC standard solutions were used as the learning
database and were used for testing the samples. The concentrations
of sample solutions are reported in units of mg mL
1.
2.4. Ultra-violet/visible spectrophotometry
All absorption spectral recordings and absorbance measure-
ments were performed on a double beam V-650 UV-visible spec-
trophotometer (Jasco, Tokyo, Japan) by monitoring the absorbance
changes in the wavelength region 200e600 nm. Matched sets of
10-mm cells quartz cuvette were used throughout. The optimum
wavelength for TC determination was established as 432 nm. A
yellow soluble TC was present in all samples. Sample solutions
were measured at 432 nm for all TC concentrations.
3. Results and discussion
3.1. Comparison of solvent for forming colour of tetracyclines
In general, tetracycline analysis by spectrophotometry needs
colour-forming reagents such as p-N,N-dimethylphenylenediamine
and sodium meta periodate (Elisha, Mohammad, Deepthi, Ben, &
Beulah, 2011), chloramine-T (Jose
, Fl

avio, Mayara, Helena, &
Leonardo, 2010), or metal ions (Sultan, Al-Zamil, & Al-Alarfaj,
1988). The complexation between metal ions and TC was used as
a basic reaction for the determination of TC by spectrophotometry.
However, this method uses chemicals as reagents, which can be
environmentally hazardous and add costs to the analysis. Thus, in
this study of the application of DIC to detect the TC in milk, the
intrinsic colour of TC was used. Solvents used to dissolve TC include
hydrochloric acid in water (Chico, Meca, Companyo
, Prat, &
Granados, 2008), hydrochloric acid in methanol (Zhao et al.,
2004), methanol (Luciana et al., 2010), 1.0% (v/v) acetic acid
(Naoto, 2003), distilled water (Wang et al., 2007), deionised water
(Tsai et al., 2010), and ultrapure water (Hongliang et al., 2013).
Hence, 10 different solvents were studied in this work to select the
optimal one giving suitable colour of TC that could be taken by the
webcam. The solvents were mineral water, deionised water,
methanol, 0.1, 1.0, 3.0 and 5.0 M hydrochloric acid in deionised
water, 5.0 M hydrochloric acid in methanol, and 1.0 and 10.0% (v/v)
acetic acid. The TC standard of 25.0 mg mL
1 was dissolved in those
solvents and then detected with digital-image-based colorimeter
and UV-visible spectrophotometer.
Mineral water (Pranutpong & Wirat, 2012) was used as a green
reagent for colour formation with TC because mineral water
contains metal ions such as calcium(II), magnesium(II), copper(II),
sodium(I), and iron(III). Mineral water could be used as complexing
agent; however, the colour of the TC complex is not appropriate to
 R. Urapen, P. Masawat / International Dairy Journal 44 (2015) 1e5 3

determination of TC residues in milk, isolation of TC from milk

using SPE is required. Milk is a complex biological matrix;
therefore, if milk is not processed by SPE, the analysis results ob-
tained may be erroneous and inaccurate (Fig. 2). The conditioning
and eluting steps are important; therefore, they were studied
further.
The conditioning step was studied in 2 series; the rst series
used methanol and deionised water and the second series used
methanol, deionised water, and trichloroacetic acid (4 mL of each
solvent). Both series were investigated with UV-visible spectro-
photometer and digital-image-based colorimeter (Table 1). The
second series was chosen as it provides higher percentage of
recovery.
The eluting step used methanol at different volumes; 3.0, 4.0,
Fig. 1. The spectra of 25.0 mg mL
1 tetracycline standard dissolved in 5.0 M hydro-
5.0, 6.0, 7.0 and 8.0 mL (Table 2). A volume of 3.0 mL of methanol
chloric acid in methanol (closed circles) or deionised water (closed triangles).
was selected as eluent because, at this volume, the spectra of blank
sample and spiked sample provide the highest absorbance and the
percentage of recovery is the highest. Moreover, the digital images
of the extracted solutions when eluted with 3.0 mL methanol are
more dark-yellow and brighter than those obtained at the other
volumes.

3.3. Quantitative analysis of tetracycline

3.3.1. Quantitation by ultra-violet/visible spectrophotometry

The calibration curve was linear over the concentration range of
1.0e9.0 mg mL
1, with a correlation coefcient (R2) > 0.9995. The
percentage recoveries of TC from milk sample spiked at concen-
trations of 0.5, 1.0, and 2.0 mg mL
1 were determined, and it was
found that the average percentage recovery was favourable. The
limit of detection (LOD) and limit of quantication (LOQ) for TC
were 0.03 and 0.1 mg mL
1, respectively.
Fig. 2. The spectra of blank ultra-heat-treated milk sample, which was processed
(closed circles) or not processed (closed triangles) by solid-phase extraction. 3.3.2. Quantitation by DIC-ANN
The DIC-ANN for assay of TC uses its intrinsic colour. The lowest
concentration for which an image could be acquired was
be detected with DIC. General solvents like deionised water and
1.0 mg mL
1. Hence, the LOQ of DIC-ANN (Nontawat et al., 2013) for
methanol provide a colourless solution of TC. Furthermore, other
TC was 1.0 mg mL
1. Because the LOQ was higher than the MRL
solvents such as 0.1 M hydrochloric acid (Jacques et al., 1998) and
1.0% (v/v) acetic acid also give colourless solution of TC. Therefore,
Table 2
we chose to increase the concentration of hydrochloric acid and
The concentration of tetracycline (TC) from the extracted blank ultra-heat-treated
acetic acid. (UHT) milk samples and UHT milk spiked with 2.0 mg mL
1 TC and percentage re-
It was found that 5.0 M hydrochloric acid in deionised water and covery obtained at different volume of eluent (methanol) in solid-phase extraction.a
in methanol result in darker-yellow TC solution than the other
Volume of Concentration of TC (mg mL
1) Recovery (%)
solvent systems. The 5.0 M hydrochloric acid in methanol was methanol (mL)
Blank milk Milk 2.0 mg mL
1 TC
chosen because the spectrum (Fig. 1) shows higher absorbance than
sample
with 5.0 M HCl in DI water (Fig. 1). Moreover, methanol is also used
in the elution step of SPE. 3 5.6 7.5 94.4 0.13
4 4.3 5.9 63.2 0.24
5 4.1 6.0 77.6 0.09
3.2. Solid phase extraction conditions 6 3.8 5.3 61.6 0.42
7 3.5 4.8 53.1 0.31
8 3.2 5.1 76.5 0.46
SPE is used for separation the compounds of interest from im-
a
The results were obtained using an UV-visible spectrophotometer; values are
purities, as well as for the preconcentration of samples. SPE consists
means standard deviation (n 3).
of 4 steps: conditioning, loading, washing, and eluting. For the

Table 1
The concentration of tetracycline (TC) residues in the extracted blank ultra-heat-treated (UHT) milk samples and UHT milk spiked with 1.0 mg mL
1 TC and percentage recovery
obtained by different conditioning steps of solid-phase extraction.a

Conditioning step UV-visible spectrophotometer Digital-image-based colorimeter

Blank milk (mg mL
1) Spiked milk (mg mL
1) Recovery (%) Blank milk (mg mL
1) Spiked milk (mg mL
1) Recovery (%)

Series 1 6.3 0.23 6.7 0.27 47.3 6.5 4.02 7.0 0.83 50.0
Series 2 6.3 0.18 7.3 0.05 99.6 6.0 0.84 7.0 0.42 100.0
a
The results obtained with an UV-visible spectrophotometer or digital-image-based colorimeter were compared; values are means standard deviation (n 3). Series 1,
methanol and deionised water; Series 2, methanol, deionised water and trichloroacetic acid (4 mL of each solvent).
4 R. Urapen, P. Masawat / International Dairy Journal 44 (2015) 1e5

Table 3
The concentration (mg mL
1) and relative accuracy (%) of tetracycline determination in the extracted blank milk sample and milk spiked with tetracycline at different
concentrations.a

Milk samples Blank milk Milk 0.5 mg mL
1 TC Milk 1.0 mg mL
1 TC Milk 2.0 mg mL
1 TC

UV-VIS DIC RA UV-VIS DIC RA UV-VIS DIC RA UV-VIS DIC RA

UHT milk brand 1 6.0 0.10 6.0 0.42 99.8 6.5 0.04 6.5 4.02 100.5 6.9 0.06 7.0 0.00 100.9 7.8 0.20 7.5 0.85 96.3
UHT milk brand 2 6.4 0.06 6.5 4.46 101.9 6.9 0.04 7.0 0.42 102.2 7.2 0.05 7.0 0.83 97.5 8.3 0.05 8.0 1.27 96.7
UHT milk brand 3 6.8 0.03 6.5 4.02 95.9 7.3 0.08 7.5 0.85 103.3 7.7 0.10 8.0 0.42 103.5 8.8 0.13 9.0 0.00 102.6
Pasteurised milk brand 1 6.1 0.03 6.0 0.42 99.2 6.5 0.03 6.5 0.03 105.0 6.9 0.07 7.0 0.42 101.5 7.9 0.16 8.0 0.42 101.4
Pasteurised milk brand 2 5.7 0.06 5.5 0.85 96.3 6.2 0.06 6.2 0.42 91.9 6.7 0.05 6.5 0.83 97.0 7.5 0.19 7.5 0.00 99.6
Pasteurised milk brand 3 6.6 0.06 6.5 0.42 99.2 7.0 0.04 7.0 0.00 99.7 7.6 0.06 7.5 0.42 99.2 8.5 0.18 8.5 0.00 99.7
Sterilised milk brand 1 6.5 0.19 6.0 0.42 92.0 7.0 0.03 7.0 0.42 99.7 7.5 0.03 7.5 0.85 100.3 8.2 0.11 8.0 0.42 97.3
a
Samples were analysed with digital-image-based colorimeter (DIC) and UV-visible spectrophotometer (UV-VIS); values are means standard deviation (n 3).
Abbreviations are: RA, relative accuracy; TC, tetracycline; UHT milk, ultra-heat treated-milk.

Commission, Ministry of Education and Naresuan University (in the

year 2013) are gratefully acknowledged. Special thanks to Depart-
ment of Chemistry, Faculty of Science, Naresuan University for
laboratory facilities and consumable materials. The authors thank
Dr. Filip Kielar for helpful comments and corrections of the
manuscript.

References

Fig. 3. The relationship between total RGB values and tetracycline concentration
spiked in blank ultra-heat-treated milk sample before (closed triangles) and after
(closed circles) solid-phase extraction.
(0.1 mg mL
1), SPE was chosen for pre-concentration and increasing
the colour of the extracted TC. It was found that enrichment factor
(EF), dened as the ratio of the slopes of the linear section of the
calibration curves before and after pre-concentration (Zhoalun,
1993), was 71 (Fig. 3) and the colour intensity of the extracted
milk sample was visibly increased after SPE. Consequently, DIC-
ANN, combined with SPE, could be utilised for the determination
of TC residues in milk at a concentration lower than the LOQ of the
proposed technique.
The TC standard solutions in the concentration range of
1.0e9.0 mg mL
1 were photographed with the laboratory-made
digital-image-based colorimeter. A series of TC standard solution
images were saved as the calibration set for testing the unknown
extracted milk samples. It was found that the average percentage
recoveries were in the acceptable range, with mean square error
(Ipredict, 2012) less than 0.2% and relative accuracy (Saowanee,
Duangkamol, & Paisal, 2013) in the range of 91.9e105.0%
(Table 3) for three brands of milk samples.
4. Conclusions
The SPE-DIC-ANN method developed in this study provides a
rapid, sensitive, reliable, precise, reagent-saving, inexpensive, and
simple procedure for the determination of tetracycline residues in
bovine milk. Furthermore, the digital-image-based quantication
using database of TC standard solutions shows sufcient precision
and accuracy. Compared with UV-visible spectrophotometry, the
calibration was measured only once and could be used at all future
measurements, as long as the photographic conditions are
identical.
Acknowledgements
Financial support from the Centre of Excellence for Innovation in
Chemistry (PERCH-CIC), Ofce of the Higher Education
Aree, C., & Niamh, N. D. (2011). Rapid and semi-quantitative presumptive tests for
opiate drugs. Talanta, 86, 284e292.
Chico, J., Meca, S., Companyo
, R., Prat, M. D., & Granados, M. (2008). Restricted
access materials for sample clean-up in the analysis of trace levels of
tetracyclines by liquid chromatography application to food and environmental
analysis. Journal of Chromatography A, 1181, 1e8.
Durairaj, M., & Thamilselvan, P. (2013). Applications of articial neural network for
IVF data analysis and prediction. Journal of Engineering, Computers and Applied
Sciences, 2, 11e15.
EC. (1999). Commission regulation (EU) no 508/1999 of 4 March 1999 amending
annexes I to IV to Council regulation (EEC) no 2377/90 laying down a
community procedure for the establishment of maximum residue limits of
veterinary medicinal products in foodstuffs of animal origin. Ofcial Journal of
the European Communities, 60, 22e23.
Elisha, D. T., Mohammad, T., Deepthi, G. K., Ben, M. C., & Beulah, R. C. (2011).
Spectrophotometric determination of tetracyclines using p-N,N-dimethylphe-
nylenediamine and sodium metaperiodate. Rasayan Journal of Chemistry, 4,
539e543.
FDA. (2005). MI-05-5 milk safety branch (HFS-626) of tolerance and/or safe levels of
animal drug residues in milk. Washington, DC, USA: U.S. Food and Drug
Administration. Available at http://www.fda.gov/food/guidanceregulation/
guidancedocumentsregulatoryin- formation/milk/ucm077350.htm Accessed
13.09.12.
Felipe, C., Susana, C., Mara, G., Daniel, G. P., Angle, J. R., Pilar, M. M., et al. (2013).
Integrated disposable electrochemical immunosensors for the simultaneous
determination of sulfonamide and tetracycline antibiotics residues in milk.
Biosensors and Bioelectronics, 50, 100e105.
Felipe, C., Victor, R. M., Susana, C., Maria, G., Rebeca, M. T., Angle, J. R., et al. (2014).
Rapid screening of multiple antibiotic residues in milk using disposable
amperometric magnetosensors. Analytica Chimica Acta, 820, 32e38.
Hendrik, D. R., & Herman, D. R. (2007). Determination of tetracycline antibiotics
in cow's milk by liquid chromatography/tandem mass spectrometry. Rapid
Communications in Mass Spectrometry, 21, 1511e1520.
Hongliang, T., Chanjiao, M., Yonghai, S., Fugang, X., Shouhui, C., & Li, W. (2013).
Determination of tetracycline in milk by using nucleotide/lanthanide coordi-
nation polymer-based ternary complex. Biosensors and Bioelectronics, 50,
447e452.
Ipredict. (2012). Time-series forecasting error statistics. Available at http://www.
ipredict.it/ErrorStatistics.aspx Accessed 20.02.12.
Jacques, F. M. N., Gerard, L., Jan, S., Harry, V. E., Henk, K., & Henk, S. (1998). Testing of
raw milk for tetracycline residues. Journal of Dairy Science, 81, 2341e2345.
Jian, W., James, D. M., & Jack, F. K. (2012). Chemical analysis of antibiotic residues in
food. New Jersey, NJ, USA: John Wiley & Sons.
Jose, A. R., Johana, E., Karina, A. A., Israel, S. I., & Jose, M. M. (2010). Determination of
tetracyclines in milk samples by magnetic solid phase extraction ow injection
analysis. Microchimica Acta, 171, 407e413.

, L. R., Fla
Jose
vio, C. B. F., Mayara, S. R., Helena, R. P., & Leonardo, P. (2010). A simple
spectrophotometric method for the determination of tetracycline and doxycy-
cline in pharmaceutical formulations using chloramine-T. Ecl
etica Qumica, 35,
139e146.
Luciana, V. C., Alejandro, C. O., & He
ctor, C. G. (2010). Development of a novel
strategy for preconcentration of antibiotic residues in milk and their quanti-
tation by capillary electrophoresis. Talanta, 82, 213e221.
 R. Urapen, P. Masawat / International Dairy Journal 44 (2015) 1e5
Naoto, F. (2003). Isolation of tetracyclines in milk using a solid-phase extracting
column and water eluent. Talanta, 59, 155e159.
Nontawat, B., Yuthapong, U., & Prinya, M. (2013). Design and fabrication of articial
neural network-digital image-based colorimeter for protein assay in natural
rubber latex and medical latex gloves. Microchemical Journal, 106, 270e275.
Pranutpong, K., & Wirat, L. (2012). Determination of tetracycline using green
reagent. In The 4th Annual Northeast Pharmacy Research Conference of 2012
Pharmacy Profession in Harmony (pp. 201e205). Thailand: Faculty of Phar-
maceutical Sciences, Khon Kaen University. PP-PS-026-KKU.
Saowanee, K., Duangkamol, V., & Paisal, J. (2013). Evaluation method for laboratory
work. Bangkok, Thailand: National Institute of Metrology. Available at http://
www.nimt.or.th/nimt/lab/index.lab.php?DeptId31&LabId42&menuarticle.
detail&ArticleId516. Accessed 14.03.12.
Sarun, S., Kosom, C., & Yuttana, I. (2014). Mobile-platform based colorimeter for
monitoring chlorine concentration in water. Sensors and Actuators B, 191,
561e566.

fani, I. E. A., Marcelo, B. L., Inaka
Ste ~, S. B., Wellington, S. L., Luciano, F. A.,
Ma
rio, C. U. A., et al. (2013). A digital image-based ow-batch analyzer for
determining Al(III) and Cr(VI) in water. Microchemical Journal, 109, 106e111.
5
Sultan, S. M., Al-Zamil, I. Z., & Al-Alarfaj, N. A. (1988). Complexometric spectro-
photometric assay of tetracyclines in drug formulations. Talanta, 35, 375e378.
Tsai, W., Huang, T., Chen, H., Huang, J., Hsue, M., Chuang, H., et al. (2010).
Determination of tetracyclines in surface water and milk by the magnesium
hydroxide coprecipitation method. Journal of Chromatography A, 1217, 415e418.
Wang, S., Yang, P., & Cheng, Y. (2007). Analysis of tetracycline residues in bovine
milk by CE-MS with eld-amplied sample stacking. Electrophoresis, 28,
4173e4179.
WHO. (1990). WHO technical report series no. 799. Evaluation of certain veterinary
drug residues in food (36th report of the Joint FAO/WHO Expert Committee on Food
Additives). Geneva, Switzerland: World Health Organization.
Zhao, F., Zhang, X., & Gan, Y. (2004). Determination of tetracyclines in ovine milk by
high-performance liquid chromatography with a coulometric electrode array
system. Journal of Chromatography A, 1055, 109e114.
Zhoalun, F. (1993). Flow injection separation and preconcentration (p. 12). Berlin,
Germany: Weinheim.