One-time standard colour references analysis of hexavalent chromium by 1,5-diphenylcarbazide in environmental water matrices using camera-based approach

ABSTRACT The analysis of hexavalent chromium (Cr (VI)) is an important water quality monitoring for public health safe consumption. Many high-end instruments were accessible to detect total Cr precisely at parts per trillion. Because of high upkeep of instruments, important to recognise on field examination and adjusting to low-volume (green chemistry) was quite challenging. The potential methods of detecting Cr in on-spot field examination were colorimetry. Numerous scientists were carried out by lab or compact spectrophotometer using 1,5-diphenylcarbazide and ability to produce magenta at 540 nm absorbance. The detection limit was achieved as per the WHO regulation levels. Still the requirement of standardisation in the field analysis was not completely suitable for low-volume due to use of standard chemicals. So in our study, the regular method was evaluated with the colour science approach by uniform (Lab, HLab and XYZ) and non-uniform models (RGB) with spectrophotometer and different cameras (mobile and digital still) using its flash light. The study showed Cr (VI) analysis by mobile camera was able to detect at 20 ng/mL. This was far better than recent reported studies with microfluidic and nanotechnology. The analytical relation with ICP-MS and mobile for Cr (VI) was found to be 0.60 (RSE-6.2%) to 0.72 (RSE-6.72%) with adjusted r2 value > 0.98. We have found to be in good agreement of ~70% with 98% confidence level of uniform models than non-uniform. The use of one-time colour standard references of uniform models might be helpful through naked eye judgment and ability to detect at 100 ng/mL based on individual perceptions in field analysis. This will improve in determining the sample collection, site evaluation for remediation and involvement of public in the water monitoring programmes. Further studies have to be conducted in tuning models to detect as precisely like high-end instruments using various NIST SRM standards.


Introduction
The utilisations of hexavalent chromium (Cr (VI)) play a significant role in modern era like plating in finishing & anti-corrosion for metals, paints & pigments, inks, textile dyes, wood preservatives, organic synthesis and leather tanning.It has also major toxic concern due to large release into the environment leading to cancer for public.The mobilisation of Cr (VI) through manmade waste to ecosystem starts from water resources and finally leads to human via drinking/ tap/ potable water's, agriculture water and food.The most stable oxidation state is Cr (III), which is an essential trace element for living organisms.But the second most stable state is Cr (VI) and carcinogenic in nature [1,2].Due to this, the regulation limit is required to control the anthropogenic activities in the environment.Regulators like World Health Organization (WHO) and Bureau of Indian Standards have set the Cr (VI) from 50 to 100 ng/mL for drinking water.But the permissible industrial limit is set at 2 to 5 µg/mL [3,4].Normally at this level, the waste water has to be treated in various ways such as Cr reduction-coagulation-filtration (chemical way) [2,4], electrochemical precipitation treatment (physical and chemical way) [5], biosorption of Cr (VI) using microbes (biological way) [3], and mixtures of graphene oxide, copper oxide & silver oxide-based nanoparticle removal (nanotechnology way) [6].Most expulsion strategy were at present being worked on with MnO 2 nanoparticles and alumina-based catalysts or/else the set up techniques were expensive and not ready to bear the expense by industrialist [7,8].This leads to the likelihood of pollution in the environmental water systems and cause alarm to the public health.
The presence of Cr (VI) in various environmental matrices was reported by many researchers in different parts of the world by natural and manmade.In Brazil, the groundwater was found to be 0.12 μg/mL in aquifer ducts without any anthropogenic contamination [9].This may be due to oxidation of naturally occurring Cr (III) to Cr (VI) by mineral bearing oxides.Hausladen et al. 2018 revealed the MnO 2 (abbreviation in Table S1) bearing oxides were the significant explanation of normal chromium oxidation in California groundwater zones.The concentration was found in the range from 0.002 to 0.294 μg/mL [10].The natural Cr occurs as ore form called chromite.The existence of ore in smaller quantity and high quality is found in the countries like Cuba, India, Pakistan, Yugoslavia, Greece, Brazil, and New Caledonia.Generally, ores in the rocks contain 1000 to 3000 ppm of Cr, but in granite region found in the range from 5 to 200 ppm.The application of Cr in industry has made potential pollution in the soils and water.The reported studies on Cr pollution have affected the environment systems such as agricultural material in Canada (as Cr (III) 10 to 3000 μg/g into the soils), sewage sledge in New York, USA (26 to 48% of Cr (VI)), and coal & fly ash in England (15 to 152 μg/g).In this manner, the wellbeing impacts for public and labourers might prompt sickness like cellular breakdown in the lungs and bronchial asthma.In rare cases, the use of chromic acid mist exposure at 1 to 20 μg/L found to cause precancerous growth in nasal mucosa and skin ulcerations [11].
The mobilisation of Cr in the environment found from soil/sediment to water to living organism via microbes and plants as biosorption and bioaccumulation.This has enhanced the property of remediation research for Cr (VI) reported in microbes such as Enterobacter cloacae, Pseudomonas ambigua, Bacillus sp., Pseudomonas fluorescens, Pseudomonas putida, Escherichia coli, Desulfovibrio vulgaris.These microbes use Cr (VI) reductase enzyme in converting to Cr (III).Similarly in plants such as Brassicaceae family and water hyacinth in recently reported studies might be the accumulator of Cr [12].However, the excess Cr in the soil may hamper the plant metabolism like chlorophyll biosynthesis, photosynthesis and plant defensive systems as they cross the threshold of the toxicity [13].Hence the analysis of Cr is required for continuous monitoring in the areas of water quality, studies on Cr metabolism in microbes & plants, and oxidative-reductive properties in soils & rocks.
The total Cr and Cr (VI) were analysed via high-end instruments like HPLC-ICP-MSdetection limit: 200 pg/mL) [14], ICP-MS (detection limit: 9 pg/mL to 1 ng/mL) [4], ICP-OES (detection limit-4.4ng/mL) [15], AAS (detection limit-0.03ng/ml) [16][17][18] and stripping voltammetry (detection limit: 29.4 ng/mL) (abbreviation in Table S1) [19].These analytical instruments consume in the aspects of chemical separation and extraction, instrument maintenance and time consuming in the laboratory analysis to detect precisely from parts per billion (ppb) to parts per trillion (ppt).However, the WHO regulatory limit is set at 50 ng/mL for drinking water [20].At this level, Lace et al. 2019 have reported the modified EPA method 7196A to detect at 23 ng/mL (based on 3 sigma standard deviation data) in visible spectrophotometer.This method has the ability to bind the Cr (VI) with the 1,5-diphenylcarbazide (DPC) and produces the distinct red to violet colour [21].Many researchers have developed the other portable-based spectrophotometer and camera for analysis in remote regions, where the lab spectrophotometer is not feasible.Kumar et al 2020 has developed custom made mini handheld spectrophotometer with the ability to detect Cr (VI) in citrate-capped silver nanoparticle (detection limit: 26 ng/mL) [22].Researchers have reported the use of nanocomposites as detection for the mefenamic acid, L-cysteine and benzylpenicillin.However, the reported nanocomposites were well proven with the instruments like voltammetry and HPLC for detecting the compounds.This is unsuitable for the on-field analysis and has to be bypassed with the help of colorimetric techniques or portable-based instruments [23][24][25].Santra et al 2018 has developed wireless & portable LED (abbreviation in Table S1) based photometric device for measuring DPC-Cr (VI) analysis (detection limit: 25 ng/mL) [26].Shalaby et al 2020 has reported the use of digital camera with colour-based analysis for DPC method to detect at 10 ng/mL [27].
Using colour-based analysis requires mathematical model to detect the Cr (VI) in water, which is an alternate viable approach for using camera as spectrophotometer.The advantage of using camera in remote water quality monitoring is cheap, easy to use by any people, location tagging, results storage, quick judgment for sample collection, swift analysis, use of polychromatic wavelength, etc.There are 7 billion smartphone users available in this world [28].Thus, the importance of analysis Cr (VI) in camera-based approach is highly efficient on water quality monitoring, where the public can also be involved for health safety.Camera captures the image in the sensors through reflective mode using polychromatic wavelength.These sensors were based on nanostructured coatings such as bismuth sulphide, which is widely used on camera sensors [29].So, we have to analysis the image of colour produced in DPC method.Colours were measured using various mathematical model developed by international commission of illumination (CIE).The models like Lab, Hunter Lab (HLab), Luv, LCH, HSV, HSL, RGB, Yxy, XYZ, Yuv, CMYK, YIQ, YCbCr and HMMD (all the model abbreviations were shown in Table S1) are able to quantify the colour.Among them, Lab and RGB are widely used for human eye colour perception and its colour matching ability with various cameras as reported by many researchers [27,[30][31][32][33].The detection requires the standardisation of camera with Cr (VI) standards, lighting conditions, and analysis of the unknown sample, which requires scientific and technical person.This may become problem with general public, where they are not expert in analysis.Hence it is necessary to prepare a standard one-time colour references like pH paper and ability to capture the image using camera for quantification in environmental water samples as shown in Figure 1.This helps the non-technical /scientific person to judge with human naked eye and supportive for water quality monitoring reporting.So, in our study, we would like to develop the schematic process of one-time standard colour references for mobile camera as shown in Figure 1 for Cr (VI) analysis using DPC method, which can detect up to ppb levels with human naked eye judgment and camera using flash light approaches.

Instrumentation & reagents
In our study, the analytical measurements were performed in three different instruments namely 1) Shimadzu UV-Visible spectrophotometer, Model: 3600 (Shimadzu Pvt.Ltd, Japan), 2) Canon Digital Still Camera, Model: Power shot SX400IS (Canon Pvt.Ltd, India) with 16 megapixels, focal length-4 mm, f-stop-3.4,exposure time-1/60 seconds and ISO speed for camera sensors set for ISO 320 with maximum aperture of 3.53 mm and 3) Samsung Mobile Triple Camera, Model: SM-A505F (Samsung Pvt.Ltd, India) with 25 megapixels, focal length-4 mm, f-stop-1.7,exposure time-1/17 sec and ISO speed for camera sensors set for 100 with maximum aperture of 1.53 mm.All the analytical measurements of real-time water samples were compared with Inductively Coupled Plasma Mass Spectrometry (ICP-MS), Model No: iCAP RQ (Thermo Fisher Scientific Pvt.Ltd, USA).The hexavalent chromium of (Cr (VI)) standard stock at 10 µg/mL was prepared from 99% A.R grade Potassium Dichromate (SD fine Chemicals Pvt Ltd, India), in 18.2 mΩ.cm resistivity of Type 1 Milli Q water (Merck-Millipore Pvt Ltd, India).The total Cr for ICP analysis was prepared from ICP Multielement standard VIII as per the National Institute of Standards & Technology (NIST) Standard Reference Material (SRM) 3112a (Merck Pvt Ltd, Germany).1,5-diphenylcarbazide A.R grade (DPC-0.5%)(Himedia Pvt Ltd, India) was prepared in 99.9% of Acetone spectroscopy grade (Loba Chemie Pvt Ltd, France).0.2 M of sulphuric acid (H 2 SO 4 ) was prepared from 98% assay of concentrated sulphuric acid (Merck Pvt Ltd, India).A known amount of Cr (VI) concentration of 0.15 & 0.55 μg/mL and an unknown amount of plated chromium effluent collected from industry (Local chromium plating industry at Vadodara, Gujarat, India) were spiked in 1 L ground water (tap) matrices collected from our university campus (CHARUSAT, Gujarat, India) to check the analytical capabilities of the colour references-based camera quantification.

Analytical procedure
Our analytical technique was modified (only on volume) from environment protection agency (EPA)-7196A and Lace et al 2019 reported methods [21].Cr (VI) standards (from 0.1 to 1 µg/mL, with aliquot volume from 0.03 to 0.3 mL respectively) were prepared from the stock standard and dissolved in 0.75 mL of 0.2 M H 2 SO 4 in test tube to attain pH of 2.0.Then addition of 0.75 mL of DPC in the respective test tubes was carried out in dark condition and homogeneously mixed the solution with vortex mixture.The solutions were ideally kept for 10 minutes at room temperature.After this, the solutions were analysed in three different analytical instruments such as absorbance spectrum (400 to 700 nm) from spectrophotometer, digital camera and mobile camera for image capturing of the cuvette using its flash light as source.Similarly, the non-spiked and spiked water samples was carried out by the above procedures.

Visible spectrophotometer based colour values
In our study, we have collected the absorbance data of visible spectrophotometer from the sample analysis ranging from 400 to 700 nm wavelength of colour region.Baskaran et al 2021 has reported the data processing of colour values from visible spectrophotometer using absorbance data.The same method was adopted in our study for converting from absorbance data to transmittance data.Later the processed data was applied in software provided by Bruce Justin Lindbloom (http://www.brucelindbloom.com) to obtain the colour values of Lab -Tristimulus values.Lab is a base model for our study which is a colour perception of human eye [30].Afterwards, the Lab was compared with other colour space models such as Red, Green and Blue (RGB) and Hunter based Lab, which was converted by using Colour mine website (http://www.colormine.org) as reported by Shalaby et al 2020 [27].

Digital still camera and mobile camera based colour values
The image capturing of samples was carried out in glass cuvette of 10 mm path length inside a cuboid wooden box having volume of 3,375 cm 3 .The box wall is covered with Whatman filter paper no 1 as background of the image capturing.The distance between the camera and cuvette is 27.7 cm and the light source was used from respective camera flash light (note: not surrounding environment light), while capturing the images.The camera captured in 10 different images and the image pixel resolution was set by cropping the colour portion present in glass cuvette image.This was carefully carried out in Adobe Photoshop CS6, 2012 to avoid the light reflection as well as shadowing effect from the glass cuvette.The cropped images were in the form of square pixels of 71 × 71 for digital still camera and 113 × 113 pixels for mobile camera.Further, the images were processed in image colour summariser web service developed by Martin Krzywinski (Image Colour Summariser Version 0.76, website: http://mkweb.bcgsc.ca/) to obtain the Lab values of the colour as reported by Baskaran et al 2021 [30].Each image was carried out in k-means cluster statistical analysis with parameters was set to 20 clusters and 200-pixel precision.The measured colour data was converted to text format and later in excel readout for colour space calculation.

Colour spacing -delta E (ΔE) analysis
The CIE ΔE 2000 calculation (abbreviation in Table S1) was performed for visible spectrophotometer, digital still camera and mobile camera in Colour mine website (http://www.colormine.org)[31].The calculation measures the distance between the blank and samples as shown in equation (1) The sample(s) and blank (b) of L (luminosity), a (red: green ratio) and b (yellow: blue ratio).The same formulas were applied for Hunter Lab, XYZ and RGB models to study the best analytical linear fit in measuring the Cr (VI) in environment samples.The mean Lab of all blank samples were compared with each blank to obtain their ΔE.Afterwards, their standard deviation (σ) of blank ΔE were considered for detection limit computations.The σ were multiplied with 3 and 10 times to obtain the limit of detection (LOD) and limit of quantification (LOQ) respectively.Similar computations were performed in other colour space models to evaluate its capabilities.

For visible spectrophotometer
The quantification of Cr (VI) in visible spectrophotometer using DPC method was well established at 540 nm absorbance.Since the absorption occurs in green colour spectrum, this method produces a distinct colour of red-violet or otherwise it is called as magenta.By using spectrophotometer, we can measure the absorbance.However, it is required for the standardisation to quantify the Cr (VI) in water samples.In our study, using Beer-Lambert's law relation with transmittance, we have observed and standardised the colour references for visible spectrophotometer as shown in Figures S1 and S2.The same standard solutions were analysed in digital camera and mobile camera with the help of the flash light under dark conditions.The image file of colour reflectance and the conversion from absorbance to transmittance was followed by Baskaran et al 2021 [30].The relative standard deviation (RSD) of Lab and absorbance data of visible spectrophotometer was found to be 1-3% and 1-10% respectively.This has enabled to have various forms of colour space values such as XYZ, HLab and RGB as shown in Table 1 and Supplementary Table S2.
Based on the RGB colour strip of all three analyses, we have investigated it in the colour inspection 3D software by Image J.This is to understand the visual evaluation of Cr (VI) concentrations in various colour space models as shown in supplementary figures (Figure S3 to S5) [34].In visible spectrophotometer, the uniform colour space models such as XYZ, Lab and Yxy (Figure S3a, S3b, & S3d) has shown equal distances of each concentration.Similarly, in non-uniform space models such as HMMD and HSL (Figure S3e and S3f) shows the unequal distance in the concentrations.As the concentration reaches after 0.5 μg/mL, they start to show saturation of the colour.However, this was not visible with the RGB models as shown in Figure S3c.In RGB models, the green value was decreased from blank to 1 μg/mL regularly.We have plotted the chromaticity in 2D which illustrates the considerable linearity increases as concentration shown in Figure 2.However, the absorbance is >1 may saturate the colour due to the deviation of Beer-Lambert's law.In our study, the law was obeyed for the range from 0.1 to 1 μg/mL.The molar absorption coefficient and Sandell's sensitivity was found to be 2.69 × 10 4 /mol.cmand 9.77 × 10 −5 μg/cm 2 .Many researchers have demonstrated the use of Lab and RGB function to quantitate the colour for various analytical measurements.The major reason of using this model is human eye evaluation in mathematical expression.Humans has two visual receptors, which is short, medium and long wavelength sensitive controlled by cones.But the light intensity function is controlled by rods.In combination of rods and cones, the variation of colour is perceived and judged as colour gradient and other mixtures of colours.Hence the human naked eye judgment either, it can quantitate or semi-quantitate depending upon the individuality apart from qualitative.In case of RGB, the primary colour mixtures visualisation as same level as Lab level.This is widely suitable with the electronics and other appliances for colour matching function towards human eye [35][36][37][38].So, in our study, we have considered Lab, HLab, XYZ and RGB for Cr (VI) analysis.In spectrophotometer, the use of light is halogen lamp at visible range, which is D65 source of near daylight region [39].The blank concentration L value was found to be 91 and produced the colour as near white and not exact white.This may be the reason of acetone mixtures with DPC compound has some level of light absorption.This is colourless zone for human eye judgment, but failed to understand the differences of near colourless white region.Due to this property, the colours of various concentration were having white tint in ascending order of concentration.Chromaticity plot shows the Cr (VI) standards (µg/mL) in respective colour coordinates of visible spectrum (400 to 700 nm wavelength) of spectrophotometer with 1-Blank, 2-0.1 µg/mL, 3-0.2 µg/mL, 4-0.3 µg/mL, 5-0.4 µg/mL, 6-0.5 µg/mL, 7-0.6 µg/mL, 8-0.7 µg/mL, 9-0.8 µg/mL, 10-0.9 µg/mL and 11-1 µg/mL.

For digital still camera
The digital still camera was captured the images using its flash light reflection of the coloured solution.This has made the blank concentration L value to be ~100 and produced almost white colour as shown in Table 1.The relative standard deviation (RSD) of Lab was found to be 2-3% in higher concentration and >20% in lower concentration.The effect of light has also hampered the colours in low concentration at 0.1 and 0.2 μg/mL as compared to visible spectrophotometer.The colour strip was evaluated in Image J inspection and found to be in-consistent in all colour space models as shown in Figure S4a-f.Similarly, it was plotted in 2D chromaticity plot observed inconsistent in all concentration as shown in Figure S5.This is due to the xenon-based flash light source of D55 used in digital still camera to capture the colour image of the solution.This made the Cr (VI) concentration colour to be pink region rather than magenta.The colour differences (ΔE) between the visible spectrophotometer and digital still camera at blank and 1 μg/mL was found to be 7 and 10.In this digital still camera, most of the function such as white balance, flash light intensity and ISO is automated due to the non-professional cameras.Using this type of camera with flash light in analysis may rise to false-positive and negative results in Cr (VI) [40].

For mobile camera
The mobile camera has also captured the images using its flash light reflection, which is LED type (abbreviation in Table S1).This has made the blank concentration L value to be ~67 and produced the grey or dark white colour as shown in Table 1.The relative standard deviation (RSD) of Lab was found to be 6-10%.The effect of light intensity didn't hamper in any concentration, but at higher concentration the colour looks to be saturation.This was confirmed with the image strip visualisation in the Image J 3D colour inspection as shown in Figure S6a-f in all colour space models.The 2D chromaticity plot of the mobile camera as shown in Figure 3 is saturating after 0.5 μg/mL concentration of Cr (VI).The LED illuminant was not classified in CIE (abbreviation in Table S1), but they generally considered as warm daylight, which is D55.Researchers has reported the optimisation of LED light in mobile cameras application follows in CIECAM02 model (abbreviation in Table S1).But it can be converted to other tristimulus models Lab, RGB and HLab for colour matching models [41,42].The ΔE between the digital still camera and mobile camera in blank and 1 μg/mL was 23 and 37 respectively.In case of spectrophotometer comparison with mobile camera was 18 and 29 respectively.
Compared to digital still camera, mobile camera is better in distinguish at lower concentration seen till at 0.1 μg/mL.Due to the white balance and ISO was followed in our experiment, the colour was able to detect at low concentrations.However, there is limitation in mobile camera in flash light intensity controls.This can be achieved in the DSLR cameras, where the flash light can be changed depending upon the conditions required for the analysis.Based on the spectrophotometer and mobile camera, the light value from 91 to 60 in blank has to be exploited for understanding and developing the perfect colour references for Cr (VI) analysis.However, the variation of the L value in HLab was 88 and 60 for spectrophotometer and mobile camera.But the digital still camera was 99 as shown in Table S2.The LED (abbreviation in Table S1) based flash light has unique property of capturing colour shades with custom white balance in RGB for cameras as reported by Tung et al 2011 [43].Generally the camera image was capturing in RGB mode than the Lab, but it was converted using Image Colour Summariser Version 0.76 as reported by Baskaran et al 2021 [30].
The standardisation plot (for visible spectrophotometer, digital still camera and mobile camera) of ΔE versus concentration for Lab, HLab, RGB and XYZ was shown in Figure S7a-d.The linearity was highly correlating with the RGB models than the uniform colour space models (XYZ, Lab & HLab).The slope and adjusted R 2 value were shown in Figure S7a-d and mobile camera was able to attain the slope of ~60 ΔE/concentration in all the models.
In case of spectrophotometer, the slope was not same as in RGB, Lab and HLab, which was converted from Lab value.This shows clearly the non-matching of colour in other models, which cannot be evaluated in colour inspection 3D and 2D chromaticity plot.From this study, we have found that preferred performance of colour analysis for Cr (VI) was in this order: mobile camera > visible spectrophotometer> digital still camera.We have also found that RGB model is preferred than the Lab and HLab for mobile camera analysis.However, for the human naked eye judgment, Lab and HLab were preferred to explain the Cr (VI) analysis.This study was found an interesting observation that all the colour models were not perfectly linear to data points, except RGB in visible spectrophotometer.This may be the reason of data were curved in nature and curve fitting analysis was performed in our data.We have found the data was not able to fit the curve and failed in the statistical analysis.The major reason is our data in heteroscedastic in nature and unbiased in the presence/nonpresence of errors as reported by statistical researchers [44,45].The standard deviation of all ΔE analysis with 10 times of repeatability and reproducibility was minimum, which is evidently seen in Figure S7a-d.Hence, we have considered the linear fit analysis and plotted the standardisation of Cr (VI).The standard errors of all models were ranges from 0.7 to 10%.However, all slopes have to be evaluated for detection limit analysis for spiked and non-spiked real-time environmental water samples to analyse the best estimations of Cr (VI) in regardless of the best adjusted R 2 value.

Analytical figures of merit
The best linearity was found with the mobile camera than the visible spectrophotometer as shown in Figure S7b.The detection limit, limit of detection (LOD) and limit of quantification (LOQ) was found ~20, 60, and 200 ng/mL respectively, with colour space model of RGB, Lab and HLab.In case of spectrophotometer, the RGB detection limit is at higher than the Lab and HLab as shown in Figure S7a.This is due to the slope was not in the range of 60 ΔE/concentration, which was found in the mobile camera.But the digital still camera was overestimating the data in detection limit, LOD and LOQ as shown in Figure S7c due to the high illumination in the colour.We have plotted the standardisation plot of absorbance-based spectrophotometer and found to be poorer as shown in Figure S7d, but not poorer than digital still camera in terms of overestimating the parameters.The analysis of Cr (VI) based on colour measurement in camera, spectrophotometer or any other portable device has to be same or nearer slope data in all the colour space models.As it was converted value from base CIE models (abbreviation in Table S1), which has less error.Each model has depended upon the instrument analysis with respect to colour capturing and the light used [46].Hence in our study, the precise quantitative colour references were found to be in mobile camera for the concentration from blank to 0.6 μg/ mL than the spectrophotometer.The analysis in mobile camera and human naked eye judgment is immediate after the DPC reaction with the Cr (VI) for 10 minutes.
Table 2 shows the comparison of the present study of Cr (VI) analysis using DPC method in various instruments reported by other researchers.The detection limit of 20 ng/mL was not achieved with the paper based microfluidic system, nanoparticles and the use of portable spectrophotometer.But achieved in mobile camera as reported by several researchers in terms of detection limit.This shows that the mobile camera has the ease of instrumental analysis, lower cost and environment friendly based on low magnitude of chemical consumptions.However, in our study demonstrated, the use of mobile camera and human naked eye judgment made the removal of chemical standardisations for frequent instrumental analysis at onsite with the help of one-time colour standard references, which is helpful in green chemistry.The detection limit was satisfactory as per WHO regulation limit for on-field analysis [20].This saves the time and also helpful in deciding the collection of samples for monitoring and for any low scale environment remediation, which helps in cost reduction and site/sample selection.However, the study was further evaluated by spiking Cr (VI) in real-time ground water samples for mobile camera & visible spectrophotometer analysis using ΔE/concentration in all the models.Thereby, we have to compare with the ICP-MS.

Comparison of visible spectrophotometer, mobile camera and ICP-MS in Cr (VI) spiked ground water samples
Based on the LOD and LOQ, we have spiked the Cr (VI) at 150, 550 ng/mL in 1 L and unknown concentration of chromium plated (added volume: 100 μL) solution spiked in 0.5 L of ground water samples collected from the university.This is to show the capability of detecting the Cr (VI).Table 3 shows the mobile camera performs better in all the model as mentioned by LOD and LOQ.However, in visible spectrophotometer, the RGB was performed better than the Lab and HLab.The study was observed the absorbance (150 ng/mL) and colour analysis of spectrophotometer not detected in some spiked water samples as well as inconsistent in all the models.This shows their performance of the slope (ΔE/concentration) is not uniform in all the colour models like mobile camera as shown in the Figure S7a & d and makes the detection unreliable.It is not possible to analysis using the linearity correlation R 2 value, standard error, or any other statistical ways.There is no available NIST SRM Cr (VI) standards to analysis using camera.The NIST standards were prepared on the basis of AAS & ICP-MS for precision due to polyatomic interferences generated during analysis.The best way to analysis the performance of ΔE/ concentration is spiking Cr (VI) concentrations in real-time environmental water samples and evaluating the slope.Lace et al. 2019 have reported that the use of DPC with Cr (VI) has better detection in spite of the several interferences such as Mo, V, Hg till 200 µg/mL.In case of Fe (III), Cr (III), Mn, nitrate (NO 3 ), phosphate (PO 4 -100 µg/mL) and Mg the interferences can occur with DPC at 10 µg/mL [21].Thereby in our study, we have compared the real time water samples with ICP-MS analysis, the blank water was found to be 6 ng/mL of total Cr.We found that the V, Hg, Mn and Mo was not detected in ICP-MS but for the Mg and Fe was found to be 7 and 0.8 µg/mL.The real time water sample of our study was found the concentration of NO 3 and PO 4 to be in 7 and 0.05 µg/mL.Similarly, the non-spiked realtime water samples was assessed by the spectrophotometer absorbance analysis and mobile camera found to be not detected as shown in Table 3.This shows that the interferents don't affect DPC, since it is within the reported tolerance limit.The study was observed that the total Cr were higher in ICP-MS analysis in spiked waters.The total Cr analysis were performed as per the NIST SRM 3122a and showing higher estimate.The major reason is polyatomic interferences from argon and matrix generated in plasma such as Argon nitride (ArN + ), argon carbide (ArC + ), argon oxide (ArO + ) can occur with Cr (III)/ (VI).This type of interferences is not possible to remove completely, even with the help of NIST SRM standards [51,52].However, we have compared the mobile camera with available high end ICP-MS by 1:1 linear plot as shown in Figure 4.The mobile camera of all colour space models except RGB was in good agreement with measured concentration in ICP-MS.The value was 0.60 (relative standard error (RSE) −6.2%), 0.65 (RSE -6.55%) & 0.72 (6.72%) with adjusted r 2 value >0.98 in Lab, XYZ, & HLab respectively.Still the precision, inter-comparison laboratories and other unknown interferences related to the mobile camera Cr (VI) analysis with various NIST SRM standards has to be conducted in future studies with the RGB, Lab, HLab and XYZ models.This shows the use of one-time colour standard references in various colour space model (uniform) of mobile camera was useful for the field analysis and the use of LED (abbreviation in Table S1) flash light in specific condition can be achieved.The colour strip of the mobile camera can be used as human naked eye judgment with the mobile camera image capture.These assistances were beneficial for the non-technical/scientific people for on-spot via green chemistry ways.Hence the colour value developed for Cr (VI) in our study might be potential useful in public health water monitoring programmes.

Conclusions
The present study showed the developing the one-time colour standard references for Cr (VI) analysis by DPC method was studied with mobile camera, digital still camera and visible spectrophotometer.The preferred performance of colour analysis for Cr (VI) was in this order: mobile camera > visible spectrophotometer > digital still camera.The requirement of standardisation, low volume of chemical reaction and ease of use instrumental analysis was found with the mobile camera.This was very important on alternate approach for green chemical analysis.The standardisation of mobile camera was able to have the matching of ΔE/concentration values in all the colour space models with the 6-10% RSD.The detection limit, LOD & LOQ were found to be 20, 60 and 200 ng/mL, which is better than the visible spectrophotometer in absorbance mode.However, the LOD and LOQ was satisfied within the WHO regulatory limit and tested with real-time ground water samples by spiking Cr (VI).The analytical relation with ICP-MS and mobile camera for Cr (VI) was found to be 0.60 (RSE-6.2%) to 0.72 (6.72%) with adjusted r 2 value > 0.98.This shows that the ~70% of analysis by mobile camera was valid with uniform colour models (Lab, XYZ, HLab) at 98% confidence level.
In case of human naked eye judgment, the detection limit was 100 ng/mL based on the mobile camera standard colour strips.This is the first time reporting the use of one-time colour standard references in human naked eye judgment like pH paper has some level of quantitative or semi-quantitative for Cr (VI) analysis.This will be helpful for the nontechnical/scientific such as public in environment monitoring.Our study was found that the visible spectrophotometer and digital still camera with their light function has limitation of analysing colour quantitative.This is due to no manual controls for flash light, ISO, and white balance in digital still camera, which is automatic by default.But the mobile camera has the control with ISO and white balance, except flash light.This can be further evaluated in DSLR camera, where the controls have manual.Further studies have to be conducted in tuning the models to detect as precisely like high-end instruments with the help of various NIST SRM standards.The other studies need to be performed in the aspect of inter-comparison studies with other laboratories and public for their difficulties in analysis.

Figure 1 .
Figure 1.Schematic process of of one-time standard colour references for mobile camera Cr (VI) analysis in environment water samples with mobile camera.

Figure 3 .
Figure 3. Chromaticity plot shows the Cr (VI) standards (µg/mL) in respective colour coordinates for Mobile Camera analysis.

Figure 4 .
Figure 4. 1:1 Analytical performance of mobile camera (all colour models) with ICP-MS analysis.

Table 1 .
Hexavalent chromium standard colour references and its data values of visible spectrophotometer, digital still camera and mobile camera.

Table 2 .
Comparison of analytical performance with other reported DPC-based Cr (VI) colorimetric method for onsite.-one time standardised is required to develop the colour standard references chart by performing 10 times of reproducibility and repeatability of the analysis. *

Table 3 .
Analytical performance of Cr (VI) analysis in environment water samples with mobile camera, visible spectrophotometer and ICP-MS.