Immobilized Sorption-Colorimetric Microprobes for Chemical Analysis

Abstract

Here, we propose a concept of immobilized sorption-colorimetric microprobes for preconcentration and sensing of colored analytical forms. Individual particles of sorbents distributed over a small area of 1 mm × 1 mm and attached to an easy-to-made strip with an adhesive layer were considered as the probes producing colorimetric responses through preconcentration of colored analytical forms. These responses were then directly recorded with a scanner at 1200 dpi, separated from a background, and processed to obtain information about the amounts of analytes. The food dyes Fast Green FCF, and Ponceau-4R were used as the proof-of-concept colored analytes. The microprobes based on silica modified with quaternary ammonium bases and on alumina were studied. Some features of the analytes’ adsorption by the probes and their scanometric sensing were found. It was shown that the proposed method is applicable for the determination of 1–7 mg L⁻¹ of the dyes.

1. Introduction

Miniaturization of devices for sorption preconcentration is one of the steadily developing areas of analytical chemistry. Minicolumns for extraction of analytes from aqueous and gaseous phases have been developed and successfully used. The use of such columns made it possible to reduce the volume of analyzed samples and to automate the analysis by injecting the entire volume of eluate directly into the instrument. Thus, flow atomic absorption methods for the determination of metals [1] or sorption-chromatographic methods for the determination of organic compounds [2,3] are known. These methods involve the quantitative extraction of analytes when the process is carried out in a dynamic mode [4].

Another method of the miniaturized solid-phase extraction involves implementation in a static mode (microsorption), when only 1–4% of the amounts of analytes is extracted by a sorbent, such as a fiber fixed in a glass tube, followed by thermal desorption directly in a gas chromatograph. All of these methods require sophisticated equipment, and many of them are time-consuming.

Methods of sorption preconcentration that exclude a desorption stage have been developed when the determination of the microcomponent is carried out directly on the surface of a sorbent. In this case, spectral methods of analysis are mainly used for detection. Recently, miniature devices for carrying out solid-phase extraction of analytes with the aim of their subsequent determination on the sorbent surface have attracted the attention of researchers. A procedure has been developed for the determination of antibiotics in natural waters and milk, based on sorption preconcentration of antibiotics on a small amount (~0.3 mg) of europium hydroxide precipitate due to the complex formation between Eu(III) and an antibiotic. The determination was carried out using digital colorimetry (smartphone) by sensitized fluorescence of europium on its hydroxide [5]. In a similar context, it should also be mentioned such a method of preconcentration as dispersive liquid–liquid microextraction, which uses a minimum amount of an extractant (up to several tens of microliters). Using this approach, a method has been developed for the determination of penicillins in the form of associates with methylene blue, based on measuring the colorimetric characteristics of the extracts using a smartphone [6].

The studies of sorption processes, even on a single particle of sorbent, have been reported [7,8,9,10]. For example, a microprobe based on a single crystal formed by Tb³⁺ ions and 1,3,5-benzenetribenzoic acid ligands was used for the selective determination of phosphate ions [7]. The use of such microprobes will make it possible, for example, to reduce the time of analysis and extract analytes from a small sample volume, which is especially important for the analysis of biological objects. When extracting colored compounds, it will be possible to use colorimetry with an ordinary laboratory scanner as a recording device.

Modern digital imaging devices (scanners and digital photo- and video cameras) allow one to easily and quickly register the colorimetric characteristics of one or more objects simultaneously. They have been applied in various fields of science, including analytical chemistry [11,12,13,14]. In most cases, these devices allow for capturing excellent-quality images with high resolution, which corresponds to a good locality of the colorimetric signal measurement. This opens up the possibility of a simple and affordable study of small samples, which is in line with the sharp today’s trend toward the miniaturization of chemical analysis and its transition from the bench to the hand [15]. Examples of the use of household color-recording devices have been described, e.g., for the determination of drugs [16], ascorbic acid using a smartphone [17], and copper in alcohol using a digital camera [18].

In this paper, a concept of immobilized sorption-colorimetric microprobes for the preconcentration and sensing of colored analytical forms is proposed. Individual sorbent particles or their arrays fixed on an adhesive substrate were considered as the probes (see Section 2.1). The contact of the probes with the analyzed solution results in the extraction of colored analytical forms and the appearance of a colorimetric response, which can be measured directly on the surface of the solid phase (see Section 2.2). The food dyes Fast Green FCF and Ponceau 4R were chosen as the analytes.

2. Materials and Methods

2.1. Reagents and Instruments

Solutions of Ponceau 4R (Sigma-ALDRICH, Burlington, MA, USA, ≥75%), Fast Green FCF (Sigma-ALDRICH, ≥85%), c = 0.1 g L⁻¹, and 1 mol L⁻¹ hydrochloric acid were used in work. Sorbents: silica chemically modified with quaternary ammonium bases (CMS-QAB, specific surface area 500 m² g⁻¹, particle size 50 µm, pore diameter 6 nm), and γ-alumina (specific surface area 140 m² g⁻¹, particle size 100–200 µm, pore diameter 8 nm) were used as the sorption-colorimetric microprobes. An Epson Perfection 4180 Photo office scanner was used to take images.

To operate microprobes, we used a simple homemade device holder, as shown in Figure 1. The device consisted of a transparent polymer plate with an adhesive zone prepared by gluing a transparent double-sided tape GPT020F (3M Company, St. Paul, MN, USA).

A sorbent was applied to the center of the adhesive zone using a pattern with a cut-out square hole of 1 mm × 1 mm or 2 mm × 2 mm with gentle pressure on it to attach it to the adhesive tape. After that, the excess sorbent was removed, and the pattern was removed from the adhesive zone. As a result, an area remained on the adhesive layer containing the separate immobilized sorption-colorimetric microprobes or their dense array, depending on the applied amount of the sorbent.

2.2. Methods

A typical analytical procedure using the immobilized sorption colorimetry microprobes is schematically shown in Figure 2. An aliquot of the dye solution and 0.5 mL of 1 mol L⁻¹ hydrochloric acid solution were added into 15 mL-test tubes and diluted with distilled water up to 10 mL. The solution was mixed and transferred into a beaker. The holder with microprobes was immersed in the solution and kept for 30 min. The device was then removed from the beaker, washed with water, and dried in air. In the case of a small sample volume analysis, the solution was dropped onto a horizontally located holder with the immobilized microprobes and held for 30 min. Next, the analyzed solution was washed off with distilled water, and the device was dried in air. Blank microprobes were prepared using the same procedure without the addition of the dyes.

The holders with microprobes were scanned at a resolution of 1200 dpi; digital image files in TIF format were saved to a computer and analyzed using the standard Histogram procedure using Adobe Photoshop software in the RGB color space. A square area of 1 mm × 1 mm or 2 mm × 2 mm was selected as a region of interest (ROI). The channel intensities were found from the histograms of their distribution, visually dividing the maximum in the histogram in half and considering the position of the secant as a desired value.

3. Results and Discussion

The concept of immobilized sorption-colorimetric microprobes proposed in this work implies the use of individual sorption elements (for example, individual particles of sorbent) or their small arrays for the extraction of colored analytical forms with subsequent registration of the colorimetric response directly on the surface of the microprobes. For the convenience of handling microprobes, an easy-to-made device was proposed (Figure 1), which uses an adhesive layer of double-sided adhesive tape to fix them on the plate. The number of applied microprobes can be controlled by using patterns with holes of different sizes and different sorbent filling densities. It was established that the microprobes are firmly held on the surface of the adhesive layer, and they are not washed off during their use, washing, and drying. The devices can be easily placed on the scanner glass and scanned at high resolution in both reflection and transmission modes.

Various sorbents can be used to create the microprobes. The main requirements for them are high efficiency and sorption rate in relation to the studied analytical form, providing sensitivity and rapid determination, as well as the absence of intrinsic color, which reduces background noise during the colorimetric measurement. In this work, silica chemically modified with quaternary ammonium bases (CMS-QAB), and γ-alumina were chosen for the preparation of the microprobes. These sorbents have a rigid, wide-pore matrix structure, which provides a high mass transfer rate. The presence of anion-exchange centers guarantees the efficiency of extraction of anionic dyes, which were the model analytes in this study. The sorption microprobes were used to determine the anionic food dyes Fast Green FCF (FG) and Ponceau-4R (P-4R). Previously [19], we determined the distribution coefficients in the Henry region for the sorption isotherms on γ-alumina (5 × 10² cm³ g⁻¹ for FG and 1 × 10³ cm³ g⁻¹ for P-4R). When using CMS-QAB, these values were determined as 3 × 10³ cm³ g⁻¹, which guarantees high extraction efficiency for the dyes.

To quantify color, the RGB color system was used, formed by three primary colors—red (R), green (G), and blue (B) [20]. The sensitivity of one or another channel to a change in the content of the analytical form depends on its color. For the green dye FG, the most sensitive channel is R; for P-4R, it is B.

3.1. Peculiarities of Recording the Colorimetric Signal of Microprobes

To optimize the conditions for recording the signal of microprobes, it is necessary to set the scanner resolution, at which the channel intensity values do not depend on it. The chosen resolution should also ensure the measurement of individual sorbent granules but should not lead to a strong slowdown in scanning and a significant increase in the file size. Figure 3 shows the change in channel intensities with increasing the scanning resolution. At low resolutions, individual microprobes cannot be recognized in the scanned image. Instead, a number of pixels possessing average intensity values between the microprobes and the background can be observed. Increasing resolution results in appearing images of the individual microprobes. Simultaneously their signals can be more effectively separated from the background. It can also be seen that the almost constant values of channel intensities are observed starting from 1200 dpi. At lower resolutions, the errors of the response measurement increase remarkably because of the small number of pixels in the image. At higher resolutions, a significant increase in the file size is noted without a remarkable increase in the image quality. Therefore, in further experiments, the scanning was carried out at the resolution of 1200 dpi.

The effect of the scanning mode (transmission or reflection) on the quality of microprobes images and their response in the presence of an analyte was studied. The results are represented in Figure 4. In this figure, the left part of the histogram represents R, G, and B channel intensities for the blank microprobes recorded in the transmission and reflection mode. It can be seen that the reflection mode provides higher intensity values (brighter image). After the adsorption of FG (the right part of the histogram), all channel intensities decrease, and the R channel provides the highest change, i.e., the highest analytical response. Obviously, the reflection mode ensures the greatest difference between the blank and the probe in the R channel.

It can also be concluded from the figure that the image sharpness is much higher when scanning is performed in the transmission mode. Therefore, this mode can be recommended for studying the faceting and structure of sorbent particles. However, it can be seen from the presented histograms that the channel intensities and their change during dye sorption are smaller in the transmission mode than those in the reflection one.

From an analytical point of view, it is important that a sensory system provides the maximum response during dye sorption. Therefore, in further studies, the reflection mode was considered the optimal one.

Changing the size of the microprobe zone from 1 mm × 1 mm to 2 mm × 2 mm has practically no effect on the value of the analytical response. The 2 mm × 2 mm zone was measured at various points on the image using a 1 mm × 1 mm region of interest (ROI). It was shown that the value of relative standard deviation (RSD) for five measurements at different points of the sample did not exceed 7%. In order to miniaturize the device, we used the size of the microprobe zone of 1 mm × 1 mm in all further experiments.

When measuring macroscopic amounts of sorbents, the colorimetric characteristics of the substrate have practically no effect on the recorded signal. In contrast to this variant, the layer of microprobes immobilized on an adhesive substrate is characterized by a small thickness and greater inhomogeneity. The immobilized sorption-colorimetric microprobes can be distributed unevenly and incompactly on the adhesive substrate. This makes it necessary to separate their signal from the signal associated with the background. For this purpose, it was proposed to carry out a colorimetric analysis using the histograms of the channel intensity distribution provided by the standard application of Adobe Photoshop. This possibility is illustrated in Figure 5a. In this figure, a square area marked with a dotted line is a region of interest (ROI), i.e., an area selected on the image that is used for calculating the average intensities of R, G, and B channels. Adobe Photoshop provides a histogram of intensity distribution for each of the three color channels. Examples of such distributions for R channels built for different ROI are shown in the top parts of each image in Figure 5.

The choice of an area completely filled with the sorbent as an ROI gives a histogram with a maximum R channel intensity distribution at 145. A gradual decrease in the proportion of the sorbent and an increase in the proportion of the background within the ROI leads to two peaks appearing on the histogram. The left peak is still associated with the “useful” signal from the microprobes, while the right one is due to the background contribution. It can be seen from the presented images that the microprobe signal can be extracted with sufficient accuracy even when the microprobe proportion in the ROI is only about 10%. The standard deviation of the recorded “useful” signal is 1.5, which is 2% of the difference between the background and the microprobe signals. In this case, the background contribution can be effectively separated and does not interfere with the “useful” signal.

An important issue is also the possibility of reducing the size of the microprobes array down to single particles of the sorbent. This capability was demonstrated by reducing the size of ROI (Figure 5b). It can be seen from the figure that a decrease in ROI size from 1 mm × 1 mm to 85 μm × 85 μm leads to a depletion of the histogram and its narrowing. However, the microprobe signal can still be determined even at an ROI size of about a single sorbent particle, and its standard deviation is only two units.

3.2. Prospects for the Analytical Application of Microprobes

For the practical use of microprobes in chemical analysis, it is necessary to establish: the time to reach sorption equilibrium and the dependence of the analytical response on the concentration of an analyte in solution. The study of recovery depending on the volume of an analyzed sample is of particular interest to assess the possibility of using microprobes for small sample volumes.

It was found that 30 min is sufficient to reach equilibrium. Calibration plots of channel intensities of the microprobes based on CMS-QAB and alumina on the concentration of dyes FG and P-4R were plotted (Figure 6). It was established that for CMS-QAB in the range of 0–0.01 mg mL⁻¹ FG, the graph is well described (R² = 0.996) by an exponent R = y₀ + Ae^−c/t, where R is the corresponding channel intensity, c is the concentration of FG, y₀, A and t are parameters of the exponential fitting.

The sensitivity coefficient (slope) in the initial section, which can be calculated as the ratio of the parameters A/t, is 2.0 × 10⁴. The plot for P-4R is also well described by the exponent for channel B (R² = 0.995) in the range of 0–0.01 mg mL⁻¹ with a sensitivity coefficient of 1.9 × 10⁴.

When alumina was used as a sorbent, the dyes were extracted from 0.05 mol L⁻¹ HCl [21]. The calibration curve for P-4R on alumina does not differ much from the one on CMS-QAB. On the contrary, the R channel intensity sharply decreases with increasing the concentration of FG, which limits the determination range to 0.007 mg mL⁻¹. The sensitivity coefficient at the initial section of the calibration curve (4.2 × 10⁴) is twice as high as compared to CMS-QAB. A similar effect was noted earlier when using diffuse reflectance spectroscopy [22]. It can be associated with a change in the conformation of FG molecules on the surface of the studied sorbents.

The estimated limits of detection lay below 0.001 mg mL⁻¹, which indicates good prospects in sensing low concentrations of the dyes. Also, it can be supposed that the microprobes will exhibit good sensitivity in detecting other colored analytical forms that are firmly adsorbed on the chosen solid phase.

In spite of the good sensitivity of the proposed method, it is worth noting that it possesses some selectivity limitations, which are inherent to molecular spectroscopy and colorimetry. They are associated with interferences from other colored compounds and analytical forms. These problems can be normally solved by choosing proper sample preparation, analytical reagents, or mathematical treatment of the colorimetric responses. As an example, it can be supposed that P-4R will not interfere with the determination of FG within the calibration range since the intensity of the R channel, which is an analytical response for FG, is almost unchanged for different concentrations of P-4R. While it is difficult to determine P-4R in the presence of FG since the intensity of the B channel changes with changing concentration of FG. In this case, it is possible to determine P-4R by the difference, taking into account the contribution to the intensity of channel B from FG.

To assess the possibility of using microprobes to determine small-volume samples, the dependence of the analytical response on the sample volume was studied (Figure 7). To perform this, the solution of FG was applied to the microprobe by dropping. It was established that the colorimetric response, which is a decrease in the R channel intensity, is observed when applying even as low as 10 μL of the analyzed solution, and it remains constant with further increase in the volume. It can be explained by the small number of microprobes within the detection zone, resulting in a small amount of the analyte adsorbed. The quantity of adsorbed analyte is normally proportional to its equilibrium concentration in solution. The small amount of adsorbed dye means small changes in its concentration in the solution. Therefore, the colorimetric response does not depend on the volume of the solution.

These results indicate the possibility of using the sorption-colorimetric microprobes to analyze samples in a wide range of volumes, up to several tens of microliters. It is promising from the point of view of minimizing the reagent consumption, as well as solving problems when the available sample volume is limited.

4. Conclusions

A concept of immobilized sorption-colorimetric microprobes for the determination of colored analytical forms has been proposed and tested by the example of the food dyes Fast Green FCF and Ponceau-4R. A design of an easy-to-made microprobe holder carrying a 1 mm × 1 mm sensing zone has been developed. Alumina and silica, chemically modified with quaternary ammonium bases, have shown good prospects as microprobe sorbents. The conditions for recording the colorimetric analytical response using a scanner have been optimized: channels—R and B for FG and P-4R, respectively, resolution—1200 dpi, and registration mode—reflection.

It has been established that the calibration curves for both dyes on CMS-QAB and for P-4R on alumina in the range of 0–0.01 mg mL⁻¹ can be fitted with an exponential equation y = y₀ + Ae^−c/t. In the case of FG on alumina, the R channel intensity drops sharply with increasing concentration, which limits the determination range by 0.007 mg mL⁻¹. The sensitivity coefficient at the initial section (4.2 × 10⁴) is twice as high compared to CMS-QAB. It has been shown that for samples of small volume (0.01–0.2 mL), the analytical response does not depend on the sample volume.

The proposed immobilized sorption-colorimetric microprobes are easy to prepare, convenient to use, cheap, and suitable for the analysis of small sample volumes, which opens broad prospects for their practical application in the determination of colored analytical forms.