Microchemical Journal 149 (2019) 104072

Contents lists available at ScienceDirect

Microchemical Journal
journal homepage: www.elsevier.com/locate/microc

A smartphone-based colorimetry after dispersive liquid–liquid T

microextraction for rapid quantification of calcium in water and food
samples
Bo Penga, , Juanjuan Zhoua, Jiamei Xua, Mimi Fana, Yongjun Maa, Min Zhoua, Tingting Lib,
⁎

Shengguo Zhaoc
a
Key Laboratory of Bioelectrochemistry & Environmental Analysis of Gansu Province, Key Laboratory of Eco-Environment-Related Polymer Materials, Ministry of
Education, College of Chemistry and Chemical Engineering, Northwest Normal University, Lanzhou 730070, China
b
Food Test Center of Jiuquan, Jiuquan 735000, China
c
Gansu Entry-Exit Inspection and Quarantine Bureau, Lanzhou 730020, China

ARTICLE INFO ABSTRACT

Keywords: A simple method on dispersive liquid–liquid microextraction combining with smartphone colorimetric analysis
Calcium was developed to determine trace calcium. The colored complex of calcium and glyoxal bis(2-Hydroxyanil)
Microextraction (GBHA) can be extracted into a mixture solvent of chloroform and 1,2-dichloromethane in the presence of
RGB tetradecyldimethylbenzylammonium chloride (zephiramine). The R, G and B value of organic phase after DLLME
Smartphone
was caught by Android app Color Grab. The quantitative relationship between the color intensity and calcium
Colorimetry
content was constructed based on G channel as the analytical signal. The calibration curve was linear in the
range of 0.06–1.5 μg mL−1 of calcium ion with a limit of detection (LOD) of 0.017 μg mL−1 and limit of
quantification (LOQ) of 0.056 μg mL−1. The proposed method was successfully applied to analyze calcium in
water and food samples and validated with FAAS method and CRM material.

1. Introduction chromomeric reagents such as glyoxal bis(2-hydroxyanil) (GBHA)

[5,6], bromopyrogallol red [7], calmagite [8], methylthymol blue
Calcium is one of the most important elements in human body and (MTB) [9], Eriochrome Black T (EBT) [10], calmagite [8], etc. The
widely exists in environmental water bodies. Calcium is closely related UV–Vis method is widely employed in many small factories, routine
to body function and metabolism. At present, many people don't have works or base laboratory in China because the colorimetric method has
enough calcium intakes, especially for the elderly and adolescents. advantages including low cost, fast reaction, enough accuracy and re-
Excessive intake of calcium in human body can lead to excessive blood liability suitable for resource limited settings. However, most of these
calcium and endanger human health. The content of calcium in water is conventional colorimetric analyses are still limited due to lack of sta-
one of the regular items of industrial water, such as precipitation, bility, adequate sensitivity and selectivity or suffering complex ma-
groundwater, and oilfield water injection and so on. Different mature nipulation. Therefore, an accurate, simple and cost-effective routine
instrumental techniques have been developed for the analysis of cal- detecting method for trace calcium ion is also important.
cium ion, such as atomic absorption spectrometry [1], atomic emission Many household color-recording devices, such as office scanners
spectrometry [2], X-ray fluorescence [3], inductively coupled plasma [11], digital cameras [12], smartphones [13] have been introduced for
mass spectrometry (ICP-MS) [4], etc.. Although these instrumental rapid quantitative chemical analysis [14], chemical and biosensing
analyses are accepted for very good accuracy and accuracy in chal- [15], bioanalytical diagnosis [13], food diagnostics [16,17] and water
lenging detection scenes, most of these them need expensive purchase analysis [18,19]. Smartphone-based spectrometer and colorimetry have
and maintenance cost, time-consuming manipulation steps and special been gaining continuous attention, popularization and fast application
training for measurement accuracy, precision and sensitivity. due to their low cost, fast response, simple analysis process [20], real-
The UV–Vis spectrophotometric determination of calcium is gen- time communicate on [21], field detection [22], and compatibility with
erally based on the complexation reaction between calcium and some data acquisition and processing for “lab-on-a-chip” systems [23]

⁎
Corresponding author.
E-mail address: pengbo.nwnu@hotmail.com (B. Peng).

https://doi.org/10.1016/j.microc.2019.104072
Received 31 May 2019; Received in revised form 4 July 2019; Accepted 4 July 2019
Available online 04 July 2019
0026-265X/ © 2019 Elsevier B.V. All rights reserved.
B. Peng, et al. Microchemical Journal 149 (2019) 104072

especially in resource-limited settings [15]. The digital colorimetry reproducible way is crucial to the study, especially for RGB color space-
based on digital camera has been applied to calcium water hardness based models whose value of intensity are highly susceptible to ambient
analysis [6] using GBHA as an indicator in which Factorial Analysis was light source and other factors [33]. A preliminary experiment in the
employed to explore RGB color intensity data. However, GBHA‑calcium present study has shown the above size and layout was the optimal to
complex will easily breakdown exposure to air [24–26], especially capture color by directly using the bottom of centrifuge tube as mea-
during transition and placement in the spectrometer. The chelates suring place in stu, no sample dilution and transfer needed. The in-
change to be stable if they exist in an organic solvent such as chloro- tensity of the color product inside centrifuge tube is picked using Color
form [26]. Consequently the underlying organic layer in matched round Grab under automatic focusing and automatic shooting mode. Ambient
spectrophotometer cells instead of rectangular cells was ever employed light does not affect color capturing because of using the closed box
without removal of the aqueous layer [27]. [34–36]. The closed box in the experiments is made of inexpensive and
In this paper, a method on DLLME combining with Android smart- flexible polystyrene (EPS) with the white internal mist surface without
phone colorimetric analysis in-stu centrifuge tube was developed to reflection effect in order to obtain uniform internal illumination. All
determine trace calcium. The free Android app Color Grab [28] was measurements were uniformly carried out at room temperature.
used to pick RGB data of the colored solvent layer in a homemade Calcium quantifications in real samples were carried out using a Z-
closed box for quantitative analysis. No extra image exporting, im- 2000 (Hitachi, Japan) atomic absorption spectrophotometer equipped
porting and processing with computer software [29–31] was required. with standard air–acetylene flame, single element hollow cathode
The current method has been successfully applied to the determination lamps and Zeeman background correction. The flame composition was
of calcium in real samples with simple operation, high accuracy and always the same for all analytes according to the manufacturer's re-
lower detection limit. The smartphone colorimetry box is also readily to commendations. Operating conditions used were: air flow rate,
deploy on site. The aim of this work is to develop a fast and potential 10 L min−1; acetylene flow rate, 1.8 L min−1; burner height, 5.5 mm;
portable method for calcium analysis alternative to traditional re- wavelengths, 422.7 nm; slit widths, 0.2 nm; lamp currents, 10 mA; so-
ference methods at very low cost. lution aspiration rate, 4.5 mL min−1; measurement time, 3 s; and the
number of replicates, 3.

2. Material and methods 2.2. Chemicals and reagents

2.1. Instruments All Chemicals and reagents used in this work were of analytical-re-
agent grade unless otherwise stated. Ultra-pure water (resistivity of
A pH-meter (PHS-3C, Jinke, Shanghai, China) was used to measure 18.2 MΩ cm) was used throughout the work. Stock standard solution of
pH. A hotplate used for heating during sample digestion. A vortex Ca2+ (1000 μg mL−1) was supplied by National Non-ferrous Metal &
agitator was used to vigorously mix the solution in centrifuge tube. A Electronic Material Analytical and Testing Center (Beijing, China). The
centrifuge (80–1, Guohua, Shanghai, China) with 15 mL calibrated working solution of Ca2+ (2.5 × 10−4 mol L−1) was freshly prepared
centrifuge tube for accelerating the process of the phase separation. and diluted from the stock solution. Glyoxal bis (2-Hydroxyanil) (GBHA)
A Xiaomi 4G smartphone was used. Parameter: Android 4.4, CPU was purchased from Beichen Fangzheng reagent factory of Tianjin
Snapdragon 625 OctaCore 2.0 GHz, GPU Adreno 506, Screen (5.0 in., (Tianjin, China). GBHA working solution (1.25 × 10−3 mol L−1) was
1920 × 1080 (FHD), LCD IPS, Colors 16 M), Camera Rear 13.0 MP, prepared in absolute ethanol every day and stored in the dark at 4 °C
3 GB RAM, Internal storage 32 GB, Lithium Polymer 4100 mA h Battery. until used. Zephiramine was purchased from Sigma Aldrich cargo Co.,
A homemade smartphone colorimetric box (7.0 cm × 10.0 cm × 5 cm) Ltd. (Shanghai, China), and 5 × 10−4 mol L−1 zephiramine aqueous
with a low wattage LED light was designed to catch the color of organic solution was used. Chloroform, absolute alcohol, 1, 2-dichloromethane,
phase at the bottom of centrifuge tube without removal of the aqueous sodium borate(stored in plastic bottles), sodium hydroxide, calcium
layer and sampling transfer, or dilution either (Fig. 1). carbonate, sodium chloride, nitric acid (70%), hydrochloric acid (37%),
A 3 W 6500 K white LED light is embedded in a polystyrene block hydrogen peroxide (30%) were purchased from Tianjin Kaitong
(2.0 cm × 3 cm × 5 cm) to obtain uniform scattered light on top of Chemical Reagent Co., Ltd. (Tianjin, China). All glassware containers
sample. The light block is put beside the centrifuge tube as the light were previously immersed with 2 mol L−1 nitric acid for 24 h and sub-
source, 1 cm from the tube, 9 cm from the bottom of tube. The top light sequently washed with Ultra-pure water.
layout can efficiently eliminate any effect from reflected light occurring
in glass tubes, and avoid consequent lensing effect resulting from 2.3. Dispersive liquid–liquid microextraction procedure
bottom light [32]. The condition to capture and measure colors in a
In DMLLE procedure, 0.5 mL 1.25 × 10−3 mol L−1 glyoxal bis (2-
Hydroxyanil) (GBHA), 1.5 mL absolute alcohol, 0.5 mL Ca2+ solution
and 3.0 mL of borax buffer solution were added into a 15 mL centrifuge
tube and diluted to 10.0 mL. Afterward, 1.4 mL 5 × 10−4 mol L−1
zephiramine solution was added and 300 μL of mixed solvent chloro-
form and 1, 2-dichloromethane (1:3, v/v) was rapidly injected into the
solution by a micro syringe to induce the formation of cloudy solution.
By following this strict sequence of addition, it has been possible to
obtain satisfactory extraction efficiency and reproducibility [5]. The
mixture solution was vigorously agitated using a vortex agitator for
1 min at 3000 rpm until the droplets were finely dispersed and a cloudy
solution further formed. Then the turbid mixture was centrifuged at
4000 rpm for 5 min. The dispersed fine droplets of the extraction phase
were settled at the bottom of centrifuge tube giving an intense orange
red color. Then color value RGB was caught by using a smartphone with
Color Grab. The reagent blank solution was also run and recorded under
the same procedure and conditions. A schematic diagram of DLLME-
Fig. 1. Smartphone colorimetric box. smartphone colorimetric procedure is roundly shown in Fig. 2.

2
B. Peng, et al. Microchemical Journal 149 (2019) 104072

Fig. 2. Schematic diagram of DLLME-smartphone colorimetry procedure.

2.4. Sample collection and pre-treatment According to preliminary statistical analysis, linear relationships be-
tween calcium and color variables deduced from the G basic color, such
Tap water samples were collected from public water supplies in our as log (G), G/R, or G/Itot ratios [6] and other mathematical models
lab. The bottle mineral water and milk powders were purchased from [34]. ∆EL⁎a⁎b⁎ and ∆ELuv based on other color spaces [39] were also
local market (Lanzhou, China). The certified reference materials (CRM) studied and compared. In all cases, results were not better than using
water sample was purchased from the Standard Sample Research the single G channel. The G channel showed a good linear response with
Institute of the Ministry of Environmental Protection (Beijing, China). satisfactory coefficient of determination because the green color is just
Calcium and zinc gluconates oral liquid samples were purchased from the complementary color of calcium-GBHA complex. Therefore, the
Hubei Midday Pharmaceutical Ltd. (Hubei, China). Calcium‑zinc oral individual G channel value was deduced and employed as the vertical
liquid samples were purchased from Harbin Pharmaceutical Group ordinate (analytical response) for the further experiments. To build
Sixth Pharmaceutical Factory (Harbin, China). Limestone was pur- calibration curve in this study, horizontal ordinate is the concentration
chased from Tianjin Kaixin Chemical Industry Co., Ltd. (Tianjin, China). of calcium and the vertical ordinate (instrumental response) is the re-
All water samples were filtered through the ultrafiltration membrane lative intensity of G channel which is normalized as follow:
with aperture 2–4 μm before analyzing.
Limestone sample (0.05 g) was put into 100 mL conical flask and I = (IG blank IG)/255 (1)
dissolved with appropriate amount of 2 mol L−1 hydrochloric acid. The
conical flask was heated on a hotplate below 110 °C for further diges- where IG represents the G channel color value of sample solution, IG
tion in fume hood until completion of sample decomposition. Repeat blank represents the G channel color value of reagent blank solution. The
the procedure if the solution was not transparent. Finally, the tem- intensity with normalization has proved to be linearly related to the
perature was slightly elevated until almost dry the digested remnants calcium concentration. The linear relationship between calcium con-
for eliminating most of the residual acid. The solution was finally centration and the relative intensity I is constructed under the optimum
transferred into a 50 mL volumetric flask and made up to the mark with conditions.
water. It is worth noting that the confirmation of an appropriate and rea-
Milk powder sample (0.4 g) was laid into 100 mL conical flask, sonable instrumental response is probably the key problem when ap-
10 mL concentrated nitric acid was added to wet up the sample for plying smartphone colorimetry to a certain color Hue series. Different
about 12 h at room temperature. It was followed by addition of 10 mL studies revealed that the conversion of the device-dependent non-uni-
concentrated nitric acid and adding dropwise 30% hydrogen peroxide form RGB signals into the corresponding device-in-dependent, uniform
and heating on a hotplate below 130 °C for further digestion in fume color spaces signals [19,38–40] can greatly enhance the calibration
hood until completion of sample decomposition. Repeat the procedure linearity and sensitivity. However, no significant superiority was ob-
if the solution was not transparent. Finally, the temperature was served when employing other color spaces instead of RGB color space in
slightly elevated until almost dry the digested for eliminating most of current study. It can be considered that the homemade closed box could
the residual acid. The solution was finally transferred into a 50 mL provide a uniform and stable measuring place with constant conditions.
volumetric flask and made up to the mark with water (pH = 2.0–4.0). The differences in smartphone hardware and quantitative models re-
All digested samples were filtered through the ultrafiltration membrane sulting from different color space were weakened. Therefore, the
with aperture 2–4 μm and stored in the refrigerator until the time of the comparison of analytical signals derived from same color space or de-
analysis. Reagent blank solutions were prepared similarly to evaluate rived from various color spaces should be carried out carefully if the
some possible contamination. determination and calculation depend on image color quantification
using camera or smartphone.
2.5. Data processing

Color Grab is a powerful, easy operation of free color acquisition 2.6. Statistical analysis
software that one touch shot can obtain RGB, HSV, Lab, CMYK and
other more than a dozen color space values with self-defined Region of All experiments were carried out three times and the data was ex-
Interest (ROI). The raw RGB values (IR, IG, and IB) were directly read pressed as mean ± SD (standard deviation). The student's t-test was
out by the App on smartphone for further quantitative analysis [37,38] used for the statistical evaluation of the results. Treatment of data and
without additional computational digital image transformation and statistical analysis (average value and recovery) were performed using
process. Calibration curves for calcium quantification employing cur- datasheets prepared in Microsoft Excel 2013 and P-values of < 0.05
rent method were first built to evaluate the channels, R, G, and B. were considered statistically significant.

3
B. Peng, et al. Microchemical Journal 149 (2019) 104072

3. Results and discussion DLLME. The volume optimization result of zephiramine was shown in
Fig. S3.
3.1. Effect of type, ratio and amount of extraction solvent The color intensity of the organic phase reached the highest when
volume of zephiramine was 1.4 mL. Therefore, 1.4 mL of zephiramine
Chloroform, 1, 2-dichloromethane and carbon tetrachloride are was selected as the most effective volume. The experiment showed that
three kinds of most commonly-used extractants in DLLME. It was found choosing extra 1.5 mL of alcohol as a disperser solvent on the basis of
that the nature of the organic extractants tremendously influenced the using zephiramine could improve the extraction efficiency. Therefore,
extractability of the calcium-GBHA complex. The phase ratio and the the combination use of 1.5 mL ethanol and 1.4 mL zephiramine solution
total volume of organic extractant would also seriously affect enrich- was employed throughout the experiment.
ment factor and sensitivity. Therefore, single or combinations of these
solvents were investigated in the study. 3.3. Effect of pH
The result indicated that the extraction efficiency and color stability
were greatly improved in the presence of 1, 2-dichloromethane as a The chromogenic reaction of calcium and GBHA is usually carried
synergistic solvent. It was very difficult to reach the maximum extrac- out under alkaline conditions, thus pH has main influence on the ex-
tion efficiency and color intensity using any of these solvents alone. traction efficiency. The result plotted in Fig. S4 showed that the max-
However, the extract color faded fairly quickly after removal of the imum color intensity was at pH = 12.5–13.0 with borax sodium hy-
aqueous layer [5]. Therefore, the combination of chloroform and 1, 2- droxide buffer. Thus buffer solutions with pH = 12.8 were finally
dichloromethane mixed solvent was chosen. chosen.
Different volume ratios of chloroform to 1,2-dichloromethane
mixed solvent (1:6–5:1) were subjected to the same DLLME procedure.
3.4. Effect of GBHA
The DLLME extraction efficiency was found to be highly dependent
upon the volume ratio of chloroform to 1,2-dichloromethane. When the
The reaction of GBHA with calcium in alkaline solution to give a
volume ratio of chloroform to 1, 2-dichloromethane was 1: 3, the color
bright red-colored chelate has served as a classically sensitive and se-
intensity of the organic phase reached the maximum (Fig. S1). Thus the
lective chromomeric reagent for calcium [26,40] detection. As shown in
mixed solvent with a volume ratio of 1:3 was maintained throughout in
Fig. S5, the color intensity increased with the increasing of GBHA
the experiments.
concentration and reached maximum when GBHA volume was at
The volume of mixed solvent (200–550 μL) was determined. The
0.5 mL, and then the intensity became slightly reduced with increasing
result in Fig. S2 indicated the DLLME was incomplete if the total vo-
volume. GBHA was prepared in ethanol due to its limited solubility and
lume of organic phase was lower than 280 μL. While the color intensity
stability in water. The total amount of ethanol would contribute to the
began to decrease as the volume was higher than 320 μL due to dilution
dispersive effect in the DLLME. Thus this is the second sources of
and decreasing in enrichment factor. Thus 300 μL of mixed solvent of
ethanol introduced (the other one is in Section 3.2). The optimum
chloroform and 1, 2-dichloromethane was selected as the optimum for
GBHA volume was chosen as 0.5 mL in all studies.
balancing both high enrichment factor and lower detection limit.

3.5. Effect of salt

3.2. Effect of type and volume of disperser solvent

The influence of ionic strength was evaluated by adding different

The injection of the disperser solvent along with mixed solvent
amounts of NaCl (0–1.0 mol L−1). The results indicated that the ex-
followed by vortex agitating made a stable cloudy solution.
traction efficiency had not obvious changes with increasing NaCl ad-
Consequently, the extraction equilibrium was achieved greatly by in-
dition compared with no salt added. Thus, salt was not employed in all
creasing the contact area between the organic phase and aqueous
experiments.
phase. However, the calcium-GBHA chelate was non-extractable even
though GBHA had a superior selectivity towards calcium [40]. It was
important to control exactly the volumes and types of the reactive so- 3.6. Effect of extraction time and centrifugation time
lutions, including extraction solvent and disperser agent in DLLME.
Several common hydrophilic dispersive solvents such as acetonitrile, The extraction time in DLLME procedure refers to the interval time
ethanol, methanol and acetone were tested at first with 300 μL ex- between injections the dispersant and the extractant into the aqueous
tracting solvent (chloroform and 1, 2-dichloromethane). However no sample and starting to centrifuge. According to the obtained results, the
efficient extraction phenomenon was observed with these dispersive equilibrium state was obtained quickly within 1 min with vortex agi-
solvents. tating in the presence of zephiramine. The variations of longer vortex
Then different surfactants or emulsifiers such as TritonX-114, Octyl agitating time were not significant. The time-consuming step was the
phenol polyoxyethylene (10) ether (OP-10), methyl three oxygen am- centrifuging of sample solution. Fig. S6 indicated that the sample
monium chloride, Aliquat336 (N235), polyvinyl alcohol, twelve alkyl- mixture was centrifuged for 5 min at 4000 rpm to reach a clear phase
benzalkanoate, polyvinylpyrrolidone, n-chloro sixteen alkyl pyrrole, separation.
zephiramine were also tried along with disperser solvent. The results
showed that only zephiramine [41], a kind of quaternary ammonium 3.7. The color stability of organic phase
base, could prompt the DLLME extraction efficiency in the presence of
ethanol. The similar active synergistic effect coming from zephiramine The red complex of calcium with GBHA remains stable only 6 h
was also found in other UV–vis detection scenes [42]. Practically, the when it is extracted with 1,2-dichloroethane [40], while it fades easily
color of the organic phase keeps very deep and stable at the bottom of once the solution is exposed to air in basic media within a few minutes
the centrifuge tube. This phenomenon could be a kind of “self-syner- [5,26]. In current study, the color of chelates in organic phase could
gism” induced by the presence of zephiramine molecule. Zephiramine keep constant up to 20 h because the complexes in organic phase were
has an positive effect of changing a coordination unsaturated chelate to sealed at the bottom of the centrifuge tube after phase separation.
extractable coordination-saturated chelate by replacing water mole- DLLME procedure provides a more favorable environment to stabilize
cules with the same reagent as the ligand GBHA [41]. No active effect the colored analyte in the organic phase, avoiding air oxidation and
on the DLLME was found using other surfactants or emulsifiers. Thus facilitating the color-picking in stu. All the measurement 30 min of
zephiramine played an indispensable role in the current process of standing time was selected.

4
B. Peng, et al. Microchemical Journal 149 (2019) 104072

3.8. Interferences confidence level. It turns out that the developed method can be suc-
cessfully apply to quantitative analysis of Ca2+ in real samples with
The potential interference from other species on the determination accurate results.
of 1 μg mL−1 Ca2+ was researched. Solutions containing Ca2+ and
various amounts of interfering ions were treated according to the re- 5. Conclusion
commended procedure. The tolerance limit of various ions was defined
as the largest amount making a variation of less than ± 5%. GHBA A DLLME procedure combined with smartphone colorimetric de-
chelates calcium at a high pH value so interferences are generally termination based on RGB color space in stu was proven to be an effi-
limited [5]. The result showed that up to 2000-fold amounts of K+, cient, fast, reproducible and sensitive methodology to analyze trace
MoO42−, NH4+ did not interfere, up to 500-fold amounts of SCN− and amount of calcium in water and food simples. The current method was
thiourea did not interfere, 200-fold amounts of SO42−, Ba2+, Mn2+, optimized and validated showing low LOQ, good repeatability and
Cr3+, Fe2+, and PO43− ions did not interfere. 50-fold amounts of some linearity. The developed method has lower sensitivity and detection
transition metal ions Ni2+, Fe3+, Cu2+, Sr2+, Co2+, Zn2+ and Mg2+ limit in comparing with digital camera method [6]. Android app Color
had positive interference. EDTA could be used as masking agent for Grab facilitates color catching in real time. DLLME technique combing
Sr2+, Co2+, Zn2+. Thiourea was used to mask Cu2+, Hg2+, Ni2+, Fe2+. one touch color picking in-stu greatly enhances the sensitivity and
Both triethanolamine and tartrate could be used to mask Fe3+, Al3+. convenience of analysis. The analytical method using smartphone App
However, excessively large amount of interfering ions should be sepa- is interesting and attractive because the whole procedures require re-
rated at first by controlling pH. latively inexpensive equipment, simple to operate, with satisfactory
accuracy, sensitivity. Most chromogenic reactions suitable for extrac-
3.9. Analytical features tion spectrophotometry could be easily simplified as current metho-
dology accompanying improvement of sensitivity. The developed
Under the optimized conditions, the analytical parameters were method is prospect to be applicable and provides employment of usual
determined. The calibration curve was linear in the range of laboratory equipment as a screening method for trace level ions.
0.06–1.5 μg mL−1 of calcium ion with a correlation coefficient (R2) of
0.9998 (Regression equation, I = 0.4896C - 0.0188). The LOD and LOQ Declaration of Competing Interest
of the method were calculated from the ratio of 3 and 10 times of
standard deviation of the measurement blanks (n = 10) to the slope of The authors declare that there are no conflicts of interest.
the linear section of calibration curve was 0.017 μg L−1 and
0.056 μg L−1. The precision of the method for 0.5 μg mL−1 calcium Acknowledgement
detection was calculated in terms of relative standard deviation (RSD)
was 2.71% for the intra-day precision (n = 9) and 3.93% for inter-day The authors would like to thank the National Natural Science
(n = 9). Foundation of China (21767026, 21862019) for financial support.
The same concentration of calcium was measured by using 10 dif-
ferent centrifuge tubes to evaluate the repeatability resulting from the Appendix A. Supplementary data
shape difference of centrifuge tubes, the precision was calculated in
terms of RSD was 1.07%. The same centrifuge tube was continuously Supplementary data to this article can be found online at https://
determined by rotation angle 36 degrees of each rotation. The precision doi.org/10.1016/j.microc.2019.104072.
of 10 times determinations was calculated in terms of RSD was 1.01%.
The precision experiments suggest that the color measurement is not References
affected by a slight displacement of the centrifuge tube when each in-
serting. It can be also found the subtle difference of the bottom of [1] F.M. Fortunato, M.A. Bechlin, J.A.G. Neto, G.L. Donati, B.T. Jones, Internal stan-
centrifugal tubes in shape or glass thickness does not affect the color dard addition calibration: determination of calcium and magnesium by atomic
absorption spectrometry, Microchem. J. 122 (2015) 63–69.
catching either. Such small deviation indicates that the disturbance [2] N. Ozbek, S. Akman, Method development for the determination of calcium, copper,
caused by the irregularity of glass material processing and the aniso- magnesium, manganese, iron, potassium, phosphorus and zinc in different types of
tropy of centrifuge tube can be negligible in color-picking. The sample breads by microwave induced plasma-atomic emission spectrometry, Food Chem.
200 (2016) 245–248.
insert/injection in current method has satisfactory reproducibility due [3] D.V. Babos, V.C. Costa, M.A. Sperança, E.R. Pereira-Filho, Direct determination of
to the well-designed dark box with a constant illumination conditions calcium and phosphorus in mineral supplements for cattle by wavelength dispersive
which are the key problem when adopting RGB color space as a X-ray fluorescence (WD-XRF), Microchem. J. 137 (2018) 272–276.
[4] B. Han, M. Ge, H. Zhao, Y. Yan, J. Zeng, T. Zhang, W. Zhou, J. Zhang, J. Wang,
quantitative model.
C. Zhang, Determination of serum calcium levels by 42Ca isotope dilution in-
ductively coupled plasma mass spectrometry, Clin. Chem. Lab. Med. 56 (2017)
4. Application 51–58.
[5] C.W. Milligan, F. Lindstrom, Colorimetric determination of calcium using reagents
of the glyoxal bis(2-hydroxyanil) class, Anal. Chem. 44 (1972) 1822–1829.
The practical samples were subjected to the DLLME procedure. The [6] A. Lopez-Molinero, V. Tejedor Cubero, R. Domingo Irigoyen, D. Sipiera Piazuelo,
suggested method was validated by comparing determination results Feasibility of digital image colorimetry—application for water calcium hardness
with Flame Atomic Absorption Spectrophotometry (FAAS) method. All determination, Talanta 103 (2013) 236–244.
[7] M. Benamor, N. Aguerssif, Simultaneous determination of calcium and magnesium
values of t value calculated according to the paired t-test are smaller by derivative spectrophotometry in pharmaceutical products, Spectrochim. Acta A
than t95%, 4 = 2.78 (t-test for a confidence level of 95%, repeated 5 Mol. Biomol. Spectrosc. 69 (2008) 676–681.
times). The data revealed that the determination between FAAS and the [8] Z. Rasouli, R. Ghavami, Simultaneously detection of calcium and magnesium in
various samples by calmagite and chemometrics data processing, Spectrochim. Acta
proposed method did not present statistically significant differences. A Mol. Biomol. Spectrosc. 169 (2016) 72–81.
The recovery experiments were conducted and all samples were spiked [9] M.A. Gil Libarona, F.A. Iñón, On the use of a small synthetic calibration set for the
with Ca2+ standard solutions in three levels. The recoveries were found simultaneous spectrophotometric multivariate determination of Ca2+ and Mg2+
in groundwater: chemical and spectral considerations, Anal. Chim. Acta 536 (2005)
ranging from 96.0%–103%. The determination results were showed in 159–169.
Table S1. The proposed method was also applied in the determination [10] A.Y. Shishov, L.S. Nikolaeva, L.N. Moskvin, A.V. Bulatov, Fully automated spec-
of calcium in water quality sample Certified Reference Material trophotometric procedure for simultaneous determination of calcium and magne-
sium in biodiesel, Talanta 135 (2015) 133–137.
(GSB07–1192-2000). The results were given in Table S2. Student's t-test [11] Z. Gorocs, A. Ozcan, Biomedical imaging and sensing using flatbed scanners, Lab
data showed there was no statistically significant difference at a 95%

5
B. Peng, et al. Microchemical Journal 149 (2019) 104072

Chip 14 (2014) 3248–3257. hydroxyanil), Anal. Biochem. 21 (1967) 320–324.

[12] O. Liñero, M. Cidad, G. Arana, C. Nguyen, A. de Diego, The use of a standard digital [28] Color Grab. Loomatix Ltd. URL http://www.loomatix.com/, (Accessed 30 June
camera as an inexpensive, portable, fast and non-destructive analytical tool to 2019).
measure colour: estimation of the ripening stage of tomatoes (Solanum lyco- [29] S. Sateanchok, S. Wangkarn, C. Saenjum, K. Grudpan, A cost-effective assay for
persicum) as a case study, Microchem. J. 134 (2017) 284–288. antioxidant using simple cotton thread combining paper based device with mobile
[13] S. Kanchi, M.I. Sabela, P.S. Mdluli, Inamuddin, K. Bisetty, Smartphone based phone detection, Talanta 177 (2018) 171–175.
bioanalytical and diagnosis applications: a review, Biosens. Bioelectron. 102 (2018) [30] K. Thongprajukaew, A. Choodum, B. Sa-E, U. Hayee, Smart phone: a popular device
136–149. supports amylase activity assay in fisheries research, Food Chem. 163 (2014)
[14] V.V. Apyari, M.V. Gorbunova, A.I. Isachenko, S.G. Dmitrienko, Y.A. Zolotov, Use of 87–91.
household color-recording devices in quantitative chemical analysis, J. Anal. Chem. [31] M. Vaher, M. Borissova, A. Seiman, T. Aid, H. Kolde, J. Kazarjan, M. Kaljurand,
72 (2017) 1127–1137. Automatic spot preparation and image processing of paper microzone-based assays
[15] K.E. McCracken, J.-Y. Yoon, Recent approaches for optical smartphone sensing in for analysis of bioactive compounds in plant extracts, Food Chem. 143 (2014)
resource-limited settings: a brief review, Anal. Methods 8 (2016) 6591–6601. 465–471.
[16] G. Rateni, P. Dario, F. Cavallo, Smartphone-based food diagnostic technologies: a [32] N. Maleki, A. Safavi, F. Sedaghatpour, Single-step calibration, prediction and real
review, Sensors 17 (2017) 1453. samples data acquisition for artificial neural network using a CCD camera, Talanta
[17] Y. Chen, G. Fu, Y. Zilberman, W. Ruan, S.K. Ameri, Y.S. Zhang, E. Miller, 64 (2004) 830–835.
S.R. Sonkusale, Low cost smart phone diagnostics for food using paper-based col- [33] A. Lopez-Molinero, D. Liñan, D. Sipiera, R. Falcon, Chemometric interpretation of
orimetric sensor arrays, Food Control 82 (2017) 227–232. digital image colorimetry. Application for titanium determination in plastics,
[18] M.A. Hossain, J. Canning, S. Ast, P.J. Rutledge, A. Jamalipour, Early warning Microchem. J. 96 (96) (2010) 380–385.
smartphone diagnostics for water security and analysis using real-time pH mapping, [34] A. Choodum, K. Parabun, N. Klawach, N.N. Daeid, P. Kanatharana,
Photonic Sensors 5 (2015) 289–297. W. Wongniramaikul, Real time quantitative colourimetric test for methampheta-
[19] I. Hussain, M. Das, K.U. Ahamad, P. Nath, Water salinity detection using a smart- mine detection using digital and mobile phone technology, Forensic Sci. Int. 235
phone, Sensors Actuators B Chem. 239 (2017) 1042–1050. (2014) 8–13.
[20] M.M.A. Zeinhom, Y. Wang, Y. Song, M.-J. Zhu, Y. Lin, D. Du, A portable smart- [35] W. Wongniramaikul, W. Limsakul, A. Choodum, A biodegradable colorimetric film
phone device for rapid and sensitive detection of E. coli O157:H7 in yoghurt and for rapid low-cost field determination of formaldehyde contamination by digital
egg, Biosens. Bioelectron. 99 (2018) 479–485. image colorimetry, Food Chem. 249 (2018) 154–161.
[21] M.E. Solmaz, A.Y. Mutlu, G. Alankus, V. Kılıç, A. Bayram, N. Horzum, Quantifying [36] A. Choodum, P. Kanatharana, W. Wongniramaikul, N. NicDaeid, Poly vinyl alcohol
colorimetric tests using a smartphone app based on machine learning classifiers, cryogel as a selective test kit for pre and post blast trinitrotoluene, Sensors
Sensors Actuators B Chem., 255, Part 2 (2018) 1967–1973. Actuators B Chem. 222 (2016) 654–662.
[22] E. Haque, B.J. Mailloux, D. de Wolff, S. Gilioli, C. Kelly, E. Ahmed, C. Small, [37] A. Choodum, P. Kanatharana, W. Wongniramaikul, N. NicDaeid, Rapid quantitative
K.M. Ahmed, A. van Geen, B.C. Bostick, Quantitative drinking water arsenic con- colourimetric tests for trinitrotoluene (TNT) in soil, Forensic Sci. Int. 222 (2012)
centrations in field environments using mobile phone photometry of field kits, Sci. 340–345.
Total Environ. 618 (2018) 579–585. [38] D. Bueno, L.F. Valdez, J.M. Gutierrez Salgado, J.L. Marty, R. Munoz, Colorimetric
[23] V. Kılıç, N. Horzum, M.E. Solmaz, From Sophisticated Analysis to Colorimetric analysis of Ochratoxin A in beverage samples, Sensors (Basel) 16 (2016).
Determination Smartphone Spectrometers and Colorimetry, https://www. [39] A.A. Mohamed, A.A. Shalaby, Digital imaging devices as sensors for iron determi-
intechopen.com/online-first/from-sophisticated-analysis-to-colorimetric- nation, Food Chem. 274 (2019) 360–367.
determination-smartphone-spectrometers-and-colorimetry , Accessed date: 30 June [40] M. Nishimura, S. Noriki, Extraction and spectrophotometric determination of cal-
2019. cium with glyoxal-bis(2-hydroxyanil), Bunseki kagaku 21 (1972) 640–643.
[24] L. Singer, W.D. Armstrong, L.M. Colman, Microdetermination of calcium in biolo- [41] M. Nishimura, S. Noriki, S. Muramoto, Effects of quaternary ammonium bases on
gical materials, Anal. Biochem. 9 (1964) 217–223. valence-saturated but coordination-unsaturated chelates: part I. Extraction of che-
[25] F. Lindstrom, C.W. Milligan, Mechanism of color fading in the direct spectro- lates of glyoxal bis-(2-hydroxyanil) and o-(salicylideneamino) phenol, Anal. Chim.
photometric method for Ca using glyoxal bis(2-hydroxyanil), Anal. Chem. 39 Acta 70 (1974) 121–126.
(1967) 132–133. [42] A. Niazi, N. Khorshidi, P. Ghaemmaghami, Microwave-assisted of dispersive li-
[26] F. Lindstrom, C.W. Milligan, Derivatives of Glyoxal Bis(2-hydroxyanil) direct cal- quid–liquid microextraction and spectrophotometric determination of uranium
cium reagents, Anal. Chem. 36 (1964) 1334–1338. after optimization based on Box–Behnken design and chemometrics methods,
[27] J.F. Scaife, Microdetermination of calcium in thymus tissue using glyoxal bis(2- Spectrochim. Acta A Mol. Biomol. Spectrosc. 135 (2015) 69–75.