Synthesis of pH indicator from saccharides and agroforestry wastes with Brønsted acids via a thermo-chemical strategy

2.1. Materials

Glucose (AR), fructose (AR), 5-hydroxymethylfurfural (HMF, 98 wt%), levulinic acid (99 wt%), xylose (99 wt%), sucrose (AR), molasses, cellobiose (AR), chitosan (deacetylation degree: 85%), starch (AR), microcrystalline cellulose (MCC, 99 wt%, granule diameter = 20‒100 μm, polymerization degree = 214), ethanol (99 wt%), and H-form Amberlyst-15 resin were purchased from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Lignin was obtained from Shanghai Aladdin Biochemical Technology Co., Ltd. Sulfuric acid (H₂SO₄, 98 wt%), phosphoric acid (85 wt%, H₃PO₄), sodium hydroxide (NaOH, 99 wt%), propylene glycol (PG, 99 wt%), and ethylene glycol (EG, 99 wt%) were obtained from Tianjin Guangfu Fine Chemical Research Institute (Tianjin, China). Hβ zeolite with a Si/Al ratio of 13 was obtained from Nankai Catalyst Co., Ltd (Tianjin, China). Banana leaf, banyan sawdust, reed, cornstalk, rice straw, and sugarcane bagasse were obtained from farm fields of Liuzhou City. Medical cotton was provided by Huachen Medical Apparatus Co., Ltd (Heze, China). Printing paper (A4) and express box paper were purchased online from Shanghai M&G Stationery Inc. Deionized water was obtained from a Millipore generator (USA).

2.2. Synthesis of acid-base indicator

Typically, 0.1 g of organic substrate was thoroughly mixed with 0.9-1.9 g of 10 wt% H₂SO₄ (aq) in a glass evaporating dish (15 cm in diameter). This evaporating dish was heated at 130°C in an oven; 20-50 min later, the water was completely evaporated, and the solid was incompletely carbonized. Noticeably, undesirable, yellow, water-soluble byproducts could be eliminated through thorough heating and drying. The solid was withdrawn for cooling in air; 5 min later, 5 mL of water was added, and the solid in the water was shredded with a glass rod. After stirring for 1 min, the suspension was filtered with a 0.45-μm-pore-diameter membrane, and the pH of the filtrate was adjusted to 3 with saturated NaOH (aq) for the measurement of absorbance at 430 nm with a UV-Vis spectrometer (UV3600, Shimadzu, Japan). The absorbance value was measured at least 3 times until the relative error was lower than 5%. The pH was monitored with a digital pH meter (HQ30D, Hach, USA). The residual solid was rinsed with water thrice, and it was dried at 120°C in the oven. The dried solid was a black or brown carbonaceous byproduct. Five drops of saturated NaOH (aq) were added to this acidic filtrate to adjust the pH to 11. If the value of absorbance was larger than 1 (overrange), the sample was diluted with water, and the absorbance of the dilute sample at pH 11 was taken to calculate the real absorbance value Eq. (1). Real absorbance is an important index to the performance of a reaction system.

2.3. Isolation of products

To reduce the experimental errors, larger amounts of glucose and sulfuric acid (ratio unaltered) were used in the reaction-isolation experiment. 1 g of glucose was mixed with 19 g of 10 wt% H₂SO₄ (aq). The reaction took place at 130°C for 50 min; 50 mL of water was added to dissolve the target substance. After filtration, NaOH (aq) was added until the solution was green. The solution was transferred to a chromatography column (MCC, 35 cm long, 4 cm in diameter). When the solution permeated the bed, a large quantity of water was injected and pressed into the column with an air pump. The green band gradually migrated to the bottom and was collected in a vial. The pH of the collected solution was adjusted to 7 with 10 wt% H₂SO₄, and it was dried at 60°C in a vacuum oven. Crystals were precipitated. The precipitate was shredded and mixed with ethanol. After filtration, the filtrate was dried at 70°C, and a yellow, gum-like indicator product was obtained. The mass yield of the indicator was calculated by Eq. (2).

After the reaction and filtration, a carbonaceous byproduct was obtained, and its mass yield was also calculated by Eq. (2). The potential of this byproduct as a solid-acid catalyst was evaluated in the dehydration–acetalization reactions of diols, with Hβ zeolite as the reference. 0.2 g of catalyst, 2 g of PG, and 8 g of EG were mixed at 180°C in a flask equipped with a condenser. 2 h later, the reaction system was filtered, and the filtrate was analyzed by gas chromatography (Agilent GC6890N, HP-INNOWAX column, FID detector). The conversions of diols were calculated by following the method documented in this literature [11]. This experiment was conducted twice, and the mean conversion values were taken as the real values.

2.4. Identification and characterization of products

The structure of indicator molecules was probed by ¹H and ¹³C NMR (AVANCEIII, 400 M, Bruker, Germany; Solvent: D₂O) and LC-MS (Agilent 1290II-6460, USA; C₁₈ column, mobile phase: water). The contents of C and H elements in the indicator product were measured by elemental analysis (EA, Unicube, Elementar, Germany). The functional groups of the indicator and carbonaceous byproduct were determined by FTIR (Nicolet IS5, Thermo Fisher, USA; resolution: 0.5 cm⁻¹). The structure of carbonaceous byproduct was probed with an X’Pert PRO X-ray diffractometer (40 kV, 40 mA; PANalytical B.V., Netherlands) equipped with a CuKα source, and the scan was performed from 5 to 80° (2θ) at a rate of 8° min^-1. After the reaction, the components in the filtrate were detected by LC-MS (C₁₈ column, Agilent 1290II-6460). The carbonaceous byproduct was digested in concentrated HNO₃ and H₂O₂ with a microwave for the determination of its sulfur content by ICP–OES (Agilent 725-ES, USA) at 182.03 nm. In addition, the number of total acid sites was measured via titration with 1 wt% NaOH.

3.1. Acid-base indicating abilities of products obtained from varied saccharides and agroforestry wastes

After the acid-mediated reactions of most of the sugars and agroforestry wastes, a pH-sensitive compound was generated. As shown in Table 1, the reactions of glucose catalyzed by H₂SO₄, H₃PO₄, and H-form zeolite gave products with visible-light absorbance (Eq. (1)) in alkaline medium (pH 11), and the absorbance (430 nm) was very low in highly acidic medium (pH 3), showing a pH-sensitive property. Correspondingly, the solution of products was bright green at pH 11 and colorless at pH 3. The indicator was also generated over solid-acid resin, but the polymer skeleton of this resin was slightly decomposed at 130°C, resulting in the unexpectedly high absorbance due to its decomposition products. If fructose and 5-hydroxymethylfurfural (HMF) were taken as the substrate, the absorbance at pH 11 was higher, implying that the target product (bright green indicator) might be produced after the isomerization reaction of glucose and the dehydration reaction of fructose [12]. The levulinic acid was converted into a different product, which was yellow in both acidic and alkaline systems, and thereby its pH-indicating performance was poor. Although HMF can be converted into levulinic acid over Brønsted acid catalysts [13], this conversion did not contribute to the synthesis of the target product. This target product was also successfully generated from the saccharides (sucrose, cellobiose, starch, and cellulose) that could be transformed into glucose and HMF over acid catalysts [14]. The complex composition of molasses led to the high absorbance value at pH 3, so the color change from acidic to basic medium was not obvious. This target product could not be generated from other saccharides (xylose and chitosan) and lignin (component of agroforestry waste), because glucose or HMF could not be produced in the reactions of these substrates. This target product was also produced in the reactions with real agroforestry wastes and waste paper, which contained cellulose and sugar components.

Glucose, fructose, and HMF were preferable substrates in the synthesis of the indicator, because the corresponding absorbance was low at pH 3 and high at pH 11. Considering its relatively low price, availability, and close relevance to cellulose and agroforestry wastes, glucose was taken as the substrate in the following study to improve the absorbance value at pH 11 (Table 2). At low temperatures (110–120°C), the absorbance at pH 3 was very high, indicating that levulinic acid was selectively produced and converted into the undesirable, yellow colorant. Levulinic acid is the hydration product of HMF [15], and the systems still contained copious water at the end of reactions at low temperatures, implying that the existence of water may be the reason for this yellow colorant. If the reaction temperature was elevated to 130°C, the water could be completely evaporated, and the absorbance at pH 3 was drastically reduced, indicating that this yellow colorant would be transformed into a solid or gas in a dry status. Therefore, the heating and drying procedure is very important to eliminate the undesirable colorant and to purify the target pH indicator. At 140–150°C, this undesirable colorant was eliminated (low absorbance at pH 3), but the absorbance at pH 11 was lower at a higher temperature, indicating that the indicator was thermally decomposable. Because the absorbance at pH 11 was much higher than that at pH 3, the content of the target indicator was higher than that of the undesirable colorant, and thereby the heat-resisting ability of the former was stronger than the latter. With the increase of H₂SO₄ dosage, the absorbance at pH 11 was continuously elevated, indicating that a Brønsted acid catalyst is indispensable for the target reaction. The absorbance value reached a maximum after the 50-min reaction. After isolation, the mass yields (Eq. (2) of pH indicator and carbonaceous byproduct reached 3.7 and 55%, respectively. Gas and water-soluble byproducts accounted for 41.3 wt% of the total products. The precipitate in ethanol was identified to be Na₂SO₄ by XRD (PDF 00-002-0838, Figure S1). Glucose and its hydrates were not detected (PDF 00-009-0623 and 00-002-0224), indicating that glucose was completely transformed during the thermo-chemical reaction. With longer reaction time (1 h), the absorbance was obviously decreased, indicating that this indicator was thermally unstable. Nevertheless, it was more stable than the undesirable colorant.

The color changes of this glucose-derived indicator under varied pH conditions have been illustrated in Figure 1. After the reaction, the pH of the filtrate was adjusted to 1 with H₂SO₄. Then, NaOH was continuously added, and the pH value was gradually increased. At pH 4.4, the color turned bright green (more obvious to the naked eye). In the range of 5–6, the color change was obvious. The color was more intense at higher pH in the range of 4.4–11, and it did not vary at pH ≥ 11. The pH values of 4.4–11 can be indicated by comparing with the color strengths of these photographs. Then, H₂SO₄ (aq) was added to the alkaline sample (pH 13), and the color gradually faded along with the decrease in pH. Additionally, these colors were consistent with those at the same pH before the addition of H₂SO₄, proving that the color changes of this pH indicator were stable and reversible.

Close

Figure 1.

Digital photographs of this pH indicator in different acidic and alkaline environments.

Export to PPT

3.2. Determination of the molecular structures of products

3.2.1. Nuclear magnetic resonance (NMR)

The indicator product was isolated from the reaction products via elution, column chromatography, precipitation, and drying. It was a pale-yellow, sticky, gum-like substance. This product was dissolved in D₂O for NMR measurements. As depicted in Figure 2(a), the ¹H NMR spectrum consists of four peaks, including two triplets. The weak peaks at 1.4 ppm may be ascribed to CH₂ groups in the reaction byproducts. According to the spectra of levulinic acid [16], acetylacetone (SDBS No. 1030), and 1-hexene (SDBS No. 275), these four peaks were assigned to alkyl and alkene groups, and the former were adjacent to OH and C=O groups. Due to the exchange between the active H atom in OH and the D atom in D₂O, the signal of OH was not detected. The intrinsic peaks of benzene ring (SDBS No. 898), glucose and its isomers [17], fructose and its isomers [18], ethanol (SDBS No. 1300), HMF (SDBS No. 5478), and levulinic acid [16] were not observed in this spectrum, indicating that the refined indicator was hardly contaminated by the substrate, intermediates, and byproducts. Figure 2(b) shows that the ¹³C NMR spectrum consists of four peaks. According to the spectra of acetylacetone (SDBS No. 1030), ethanol (SDBS No. 1300), and 1-hexene (SDBS No. 275), these four peaks correspond to (C=C)–C=O and alkyl groups in the vicinity of OH and C=O groups, consistent with the ¹H NMR spectrum (Figure 2a). It is noteworthy that the integral area of a peak is not proportional to the number of corresponding carbon atoms. The intensity of a ¹³C NMR peak depends on the number of H atoms adjacent to this carbon atom. Based on the information reflected by ¹H and ¹³C NMR spectra, the structure of the pH indicator was obtained, as illustrated in the inset of Figure 2(b). The relative peak area (RPA) in Figure 2(a) was calculated via counting the pixels in the Photoshop software (version 14.0), and the relative proton numbers were proportional to their RPA values. The numbers of protons in those four groups follow the ratio of 2: 2: 2: 2.84, consistent with the proposed molecular structure. The sensitivity of ¹³C NMR to a quaternary carbon atom is very low, and therefore, the carbon atom in the C=C group was not obvious (Figure 2b). It has a highly conjugated moiety (O=C–C=C–C=O) and the molecular weight of 142. The C/H molar ratio in this product was determined to be 0.68, consistent with the theoretical value (0.7). In addition, the reaction products were analyzed by LC-MS, and a component with short retention time (0.33 min, C₁₈ column) agreed with this structure (Figure S2a). The three unsaturated C atoms are not linked with H atoms (Figure S2b), so they could not be detected by ¹³C NMR. The results of reactions with different substrates proved that HMF was an important intermediate (Table 1), and thereby HMF underwent a ring-opening reaction to give this pH indicator.

Figure S2

Download

Close

Figure 2.

(a) ¹H and (b) ¹³C NMR spectra of the pH-indicating product. Inset of B: structure, name, and formula of this product. (Solvent: D₂O. RPA: relative peak area).

Export to PPT

3.2.2. Fourier transform infrared spectroscopy (FTIR)

Figure 3(a) shows the infrared spectrum of the pH indicator with reference to that of glucose. The indicator contained abundant OH groups, consistent with the molecular structure based on NMR (inset of Figure 2b). Typical twin bands took place at 1552 and 1580 cm¹, and the right band was more intense. Both of them correspond to conjugated C=C–C=O groups, i. e., α,β-unsaturated ketone [19]. The substrate glucose and intermediate HMF molecules only contain a CH₂ group. After the reaction, the abundance of CH₂ and CH₃ was obviously increased, implying that new alkyl groups were generated in the pH indicator. The band at 1160 cm^-1 was assigned to the vibration of a C–O bond in a CH₂OH group [20], consistent with the proposed molecular geometry (inset of Figure 2b). A very weak shoulder of COOH at 1753 cm⁻¹ was detected [21], implying that a very small proportion of CH₂OH groups were oxidized. Figure 3(b) shows the spectrum of the carbonaceous byproduct. Compared to the spectrum of glucose, the abundance of OH was dramatically reduced, implying the occurrence of a dehydration reaction during the heating and drying procedure [22]. Based on the XRD pattern (inset of Figure 3b), the skeleton of this carbonaceous byproduct was polyfuran and amorphous carbon, formed by the polymerization of HMF [23]. Therefore, C=C–C=O groups were also detected. The bands at 1200 and 1025 cm^-1 were assigned to SO₃H groups [24]. The number of acid sites on this carbonaceous byproduct was measured to be 1.04 mmol g^-1 by titration with NaOH (aq). However, its sulfur content was determined to be merely 0.63 mmol g^-1 by ICP–OES. Therefore, SO₃H groups accounted for 61% of total acid sites, which also include other acidic groups such as COOH. This material was employed as a solid-acid catalyst in the dehydration–acetalization reactions of diols, which can only be catalyzed by strong Brønsted acids [25]. As shown in Table 3, the conversions of PG and EG were close to the values over a typical zeolite catalyst (0% conversion without a catalyst), proving that this carbonaceous byproduct (mass yield = 55%) can serve as a strong solid acid in the catalysis field.

Close

Figure 3.

(a) FTIR spectra of refined pH indicator and (b) carbonaceous byproduct. Inset of B: XRD pattern of carbonaceous byproduct.

Export to PPT

Table 3. Catalytic activity of the carbonaceous byproduct and zeolite.

Catalyst	PG conversion (%)	EG conversion (%)
None	0	0
Carbonaceous byproduct	17.2	7.9
Hβ zeolite	23.7	8.4

3.2.3. Ultraviolet-visible (UV-Vis) spectroscopy

The UV-Vis spectra of pH indicators obtained from glucose and its derivatives, polysaccharides, and real agroforestry wastes (Table 1) have been presented in Figures 4(a-c), respectively. The UV absorbance was high at 208–215 and 242–246 nm for all the indicators obtained from pure reagents (Figures 4a, b) and corn stalk. Taking glucose and HMF for example, these spectra of indicators were distinct from those of corresponding substrates (Figure S3), indicating that the substrates were completely transformed during the reactions. The band at 242–246 nm corresponds to O=C–C=C groups, which have been confirmed by ¹³NMR and FTIR (Figures 2 and 3). In contrast, the absorbance at about 270 nm was higher for the indicators obtained from rice straws and banana leaves. This difference is ascribed to the non-saccharide components in the straws and leaves. For those indicators obtained from pure reagents (Figures 4a, b), the UV absorption was more intense at a higher wavelength (246 nm) at pH 11 compared to pH 3, reflecting the effect of pH on the molecular structure. The UV-absorption behavior was also reflected by the intense fluorescence during Raman measurements (laser wavelengths = 532 and 785 nm). The fluorescence was too strong to conduct Raman measurements. More interestingly, UV absorption took place at 353 nm at pH 3 in the spectra of indicators produced from glucose, HMF, sucrose, and MCC, and these bands were transferred to the visible-light region (430 nm) at pH 11. This red shift is similar to that from 242 to 246 nm: absorption wavelength was larger in alkaline medium compared to highly acidic medium. Therefore, the pH change led to the change in molecular structure, followed by the changes in the light absorption behavior and color of the solution of the indicator.

Figure S3

Download

Close

Figure 4.

UV-Vis spectra of products generated from possible intermediates (a), polysaccharides (b), and real agroforestry wastes (c). The symbol * stands for the product of this substrate.

Export to PPT

As described in Table 1, the 430-nm absorbance at pH 3 was very low, while that at pH 11 was high for these indicators. Correspondingly, the color at pH 11 was bright green instead of yellow (produced from levulinic acid) or brown (produced from lignin). Therefore, the absorption bands at 430 nm are the reason for the bright green color shown in Figure 1. Similar color changes (from colorless to bright green with the increase of pH) were also observed for the products obtained from agroforestry wastes, but the colors at pH 11 were much lighter than those produced from glucose (Figure 1). Correspondingly, the absorbance values (Table 1) were relatively low, and the bands at 242–246 and 430 nm at pH 11 were very weak (Figures 4c). Real agroforestry wastes exhibited much poorer performance in the production of pH indicator. On the one hand, the absorbance was much lower. On the other hand, the indicator products contained a larger quantity of impurities.

3.3. Reaction network

The reaction network has been illustrated in Figure 5. The experimental results (Table 1) demonstrate that these reactions are catalyzed by Brønsted acids. Glucose, fructose, and HMF are indispensable intermediates, while the reaction of levulinic acid gave undesirable yellow colorants instead of the target pH indicator. HMF can be generated from fructose via a dehydration reaction, and fructose can be generated from glucose via an isomerization reaction [12]. Glucose is the hydrolysis product of cellulose [26]. The polymerization of glucose and HMF gives the carbonaceous byproduct or humin [27]. If H₂SO₄ is selected as the catalyst, the sulfonation reaction takes place, and the carbonaceous byproduct will be functionalized to be a solid acid. The side CHO group of HMF is reduced into CH₃ via hydrodeoxygenation reaction with H₂, and this reaction can occur without a metal catalyst [28]. H₂ is generated from the reaction between steam and carbon, which is generated from carbonized saccharides, especially in the presence of concentrated H₂SO₄ with strong dehydration–carbonization ability. This (more carbon and H₂) is a reason for the high absorbance and high yield of the indicator with H₂SO₄ as the catalyst (Table 1). Alternatively, levulinic acid is inevitably generated via the acid-catalyzed rehydration reaction of HMF [15], resulting in the formation of yellow colorants. These yellow colorants are unstable at high temperatures and can be eliminated through heating in a dry state. The reducing product of HMF undergoes a ring-opening reaction to give an α,δ-dicarbonyl compound [29]. This compound reacts with formaldehyde via acid-mediated aldol condensation reaction [30], and formaldehyde is generated from the thermal decomposition of sugars and HMF [31]. In the final step, the C₇ diketone molecule is rearranged to give the target indicator, which is a highly conjugated, polar compound. The rearrangement reactions of ketones are catalyzed by Brønsted acid catalysts [32], which play a crucial role in most of the catalytic reactions for the synthesis of pH indicators.

Close

Figure 5.

Plausible proton-mediated reaction pathway.

Export to PPT

3.4. Structure-dependent color change process

This compound is a highly conjugated α,β-unsaturated diketone, which can be transformed into enol via keto–enol isomerization in aqueous solution even at room temperature [33]. However, compared to the conjugated structures of many pH indicators and pigments [34], the degree of conjugation is not enough for color changes. Therefore, it is possible that these 3-π-bond compounds (existing in highly acidic media) are transparent in the visible-light region. After the addition of bases, the acidic medium is changed into a basic medium, and the OH group is deprotonated accordingly (Figure 6). The negatively charged O atom has electron-pushing ability, so the electron clouds of C–H bonds are pushed toward the p-orbital of the adjacent C=C bond to form a σ-π hyperconjugation interaction, enhancing the delocalization of electrons of the entire conjugated system. As a result, the red shift of the UV absorption band takes place, resulting in the visible-light absorption (Figure 4). It has been reported that the σ-π hyperconjugation interaction can lead to the red shift of UV absorption bands [35].

Close

Figure 6.

Possible chromic mechanism of the pH indicator.

Export to PPT