Introduction

Treating native chitin with acid hydrolysis or disintegrating chitin under weakly acidic conditions generates chitin nanowhiskers (ChNWs) [1] and chitin nanofibers (ChNFs) [2], which are fibrous nanochitin materials that are biodegradable, exhibit high crystallinity and have a lower density than that of other metal/inorganic materials. In several recent publications, nanochitin materials were used to create various materials with novel functionalities, typically as fillers for nanocomposite materials [1, 2]. ChNWs and ChNFs possess considerable amino groups on their surfaces, and the quantify of surface amino groups (SAGs) significantly alters various properties of ChNWs/ChNFs, e.g., their suspension viscosity, liquid crystal formation, and ionic adsorption capability and/or binding of various guest molecules [1, 2]. Therefore, to successfully control the properties of ChNWs/ChNFs, it is imperative to quantitatively measure SAGs. A conventional method that has been widely used for the quantification of SAGs is conductometric titration [3], although it involves several drawbacks; for example, this method requires a relatively large amount (up to several hundred milligrams) of sample and lengthy measurement times (typically one or two hours for a single measurement), and a nonversatile automatic titrator is necessary for automation. These drawbacks strongly necessitate the development of a new, rapid, and facile method for quantifying SAGs.

For the quantification of SAGs on various solid supports, previous publications have examined the use of guest molecules, which adsorb on SAGs by 1:1 stoichiometry or through amino-labeling reagents that quantitatively bind to SAGs [4,5,6,7,8,9,10,11,12,13]. These labeling reagents are added in excess to SAG-containing solids, and after sufficient reactions with SAGs, the unreacted labeling reagents are quantified by several methods, including spectrophotometry. Subtracting the excess labeling reagents from the initially added amounts provides the amounts of SAGs. Excess labeling reagents can be directly measured or converted into detectable (quantitatively measurable) compounds by quantitative reactions with subsequently added reagents. If other molecules are generated and liberated after quantitative reactions with SAGs, they can be used for SAG quantification because the amounts of these molecules and SAG are likely identical. In addition, some guest molecules (e.g., dyes) that undergo adsorptive interactions at known stoichiometric ratios can be useful, as they indicate the amounts of SAGs present after the number of remaining molecules is subtracted from the number of initially added molecules. Some candidates for the quantification of SAGs on ChNWs/ChNFs are amino-labeling reagents that react quantitatively with SAGs or anionic dyes that interact only ionically with positive SAGs.

The authors’ research group has already established such a method to quantify surface anionic groups of cellulose nanowhiskers (CNWs) and cellulose nanofibers (CNFs), i.e., analogs of ChNWs and ChNFs, via the adsorption of toluidine blue O (TBO) [12, 13]. TBO, a cationic dye, is adsorbed onto various anionic functional groups (e.g., carboxyl or sulfate groups) at a 1:1 stoichiometry. The method involves adding excess TBO against the surface anionic groups on CNWs/CNFs and then quantitively determining the remaining TBO via spectrophotometry after TBO:anionic group interaction (adsorption). Subtracting the amount of TBO remaining from the initial addition indicates the amount of surface anionic groups. The procedure is rapid (e.g., 50–100 measurements can be performed per day) and facile (the procedure involves shaking, centrifugation, dilution, and spectrophotometry), uses trace amounts of sample ( < 2.5 mg per measurements), is performed with a versatile UV/Vis spectrophotometer and provides a precision similar to that of titration. Similarly, this methodology is expected to help researchers estimate the SAG content of ChNWs/ChNFs. As stated above, many previous publications have reported the successful quantification of SAGs on various solid supports using amino-labeling reagents, including 2,4,6-trinitrobenzensulfonic acid sodium salt (TNBS) [4, 5], o-phthalaldehyde (OPA) [6, 7], ninhydrin [8, 9], and anionic dyes, including Reactive Red 4 (RR4) [10] and Acid Orange 7 (AO7), [11] suggesting successful quantification of the SAGs of ChNWs/ChNFs. The present study examined the usability of these reagents and dyes for measuring the content of SAG in ChNWs/ChNFs, tracing the methodology applied for the surface group quantification of nanocelluloses [12, 13].

Materials and methods

Acid hydrolysis of commercial chitin powders using 3 M hydrochloric acid was performed to prepare ChNWs [14, 15]. In addition to the preparation, the same chitin powders were treated with 50% sodium hydroxide solution to generate deacetylated chitin powders, followed by a similar hydrolysis step to yield two samples of surface-deacetylated ChNWs (DaChNWs) with a higher SAG content [15]. The normal (undeacetylated) ChNWs and DaChNWs were mixed at different ratios to yield various mixed ChNWs (mChNWs) with different levels of SAG. Two ChNF samples with different levels of deacetylation (corresponding to different SAGs), HL-01 and S-HL-02, were kindly donated by Marine Nano-fiber Co., Ltd. (Tottori, Japan). The SAG contents of these ChNWs/ChNFs were determined by conductometric titration using 0.01 M NaOHaq., and the SAG content values determined by titration, Stit, were compared with the values obtained by the following methods using labeling reagents and dyes.

The reactions between the employed labeling reagents (i.e., TNBS, OPA, and ninhydrin) and dyes (i.e., RR4 and AO7) are depicted in Fig. 1. Briefly, all the labeling reagents or dyes were added to aqueous suspensions of ChNWs/ChNFs in excess quantities compared to amount of SAGs, followed by a sufficient incubation time for reactions/adsorptions and subsequent centrifugation. When TNBS, RR4, and AO7 were used, the excess amounts of these reagents that remained in the supernatant were quantified by spectrophotometry and were subtracted from the amount of the initial addition to determine the SAG content values, STNBS, SRR4, and SAO7, with these reagents, respectively (Fig. 1a, d). When OPA was used, an excess amount of glycine was added prior to centrifugation to completely convert the remaining OPA into a yellow-colored adduct. The concentration of this adduct was determined by spectrophotometry and subtracted from the initial OPA molar amount, providing the SAG value SOPA (Fig. 1b). The reaction of ninhydrin with SAGs generated Ruhemann’s purple, a purple-colored product. The amounts of this product linearly correlated with the SAG content and were directly converted into SAG content values, Snin, using a calibration curve. Details of all the methods used are described in the Supporting Information.

Fig. 1
figure 1

Reactions between the SAGs of ChNWs/ChNFs and various reagents. a Binding of TNBS, b binding of OPA mediated by 2-mercaptoethanol, c generation of Ruhemann’s purple after the addition of ninhydrin, and d 1:1 electrostatic adsorption of a dye (only the case using AO7 is depicted). Conditions: i, 0.5-M borate buffer; ii, 0.1-M borate buffer; iii, lithium acetate buffer/ethanol mixture; and iv, adsorption at pH = 1–3 (see Supporting Information for detailed procedures)

Full size image

Results and discussion

Comparisons of STNBS, SOPA, Snin, SRR4, and SAO7 (i.e., the SAG content values obtained by the described labeling reagents or dyes) with Stit (i.e., the values obtained by conductometric titration) are shown in Figs. 2–4. The actual values are summarized in Table S2 in the Supporting Information, together with their standard deviation values; note that these deviations are not shown in Figs. 2–4 for clarity, although the deviations of most obtained values were less than 5%; thus, the reproducibility was sufficiently (see Table S2). As clearly indicated in Fig. 2, the values were significantly underestimated when the amino-labeling reagents TNBS, OPA, and ninhydrin were applied, as these values were considerably lower than those of Stit. However, for the ChNWs and ChNFs measurements, good proportionality was observed between the STNBS and Stit and between the Snin and Stit. Therefore, STNBS and Snin can be converted into Stit by using simple equations. In contrast, this linear trend was not found for SOPA, and its conversion into Stit was disabled via certain equations, rendering the use of the OPA inappropriate for SAG quantification. All the ChNW samples had similar SOPA values (98–150 µmol/g; see Table S2 for details), whereas the two ChNFs had higher SOPA levels (708 and 1262 µmol/g) but lower values than those of the corresponding Stit. These results imply that the reactivity of the SAGs of ChNWs/ChNFs is lower, especially those in crystalline regions; ChNWs are the residues after acid hydrolysis and are composed of crystalline regions of chitin. While the SAGs in the crystalline and amorphous regions of chitin show relatively low reactivity for OPA (see below for details), the former may be less reactive than the latter, probably due to low mobility and/or steric reasons. Figure 3 shows the relationships between SRR4 obtained at pH 1–3 and Stit, indicating significant underestimations and poor linearity at all pH values. In contrast, similar examinations using AO7 indicated that SOPA and Stit corresponded well at pH = 2 and 3, while this correlation was not observed at pH = 1, as shown in Fig. 4a. Unfortunately, the desirable quantitative property observed for ChNWs at pH = 2 and 3 was not observed for ChNFs measurements, as shown in Fig. 4b, indicating that the SAG content regions of ChNFs are considerably underestimated even at pH = 2 and 3.

Fig. 2
figure 2

Comparisons of STNBS, SOPA, and Snin with Stit. a Over the range of SAGs including only ChNWs, b over the range including ChNWs and ChNFs. Open circles, ChNWs; open squares, DaChNWs70; filled squares, DaChNWs100; filled circles, various mChNWs; crosses, two ChNFs (also for Figs. 3 and 4). The red, blue, and black symbols indicate the STNBS, SOPA and Snin results, respectively. Note that fitted linear lines were created to fit the symbols in each plot only so that the lines in a and b are not identical (also for Figs. 3 and 4)

Full size image
Fig. 3
figure 3

Comparisons of SRR4 with Stit. a Over the range of SAGs including only ChNWs, b over the range including ChNWs and ChNFs. The red, blue, and orange symbols indicate the results measured at pH = 1, 2, and 3, respectively

Full size image
Fig. 4
figure 4

Comparisons of SAO7 and Stit. a Over the range of SAGs including only ChNWs; b Over the range including ChNWs and ChNFs. The red, blue, and orange symbols indicate the results measured at pH = 1, 2, and 3, respectively

Full size image

The above observed results suggest that the reactivity of SAGs on ChNWs/ChNFs is lower than that of conventional aliphatic amines; for example, quantitative amino-labeling reagents and cationic dyes, some of which have been successfully applied for amino group quantification of solid supports [4,5,6,7, 9] or chitin/chitosan [8, 10], react or interact with the SAGs of ChNWs/ChNFs only under limited reaction conditions. The observed low reactivities may result from several factors. One is the low mobility of these SAGs due to anchoring on solid ChNWs/ChNFs, which results in a lower reactivity than that of aliphatic amines. This hypothesis is well supported by the authors’ separate examinations; namely, SAG quantification of Lewatit VPOC1065, an amino-containing ion exchange resin, via TNBS reaction and nonaqueous titration (see the Supporting Information) also yielded lower STNBS values (1077 ± 87 µmol/g) than that of Stit (2881 ± 0.05 µmol/g). However, our deviations were not fully explained by previous successful measurements of SAGs on many solid supports using TNBS [4, 5], OPA [6, 7], and ninhydrin [9]. Another possibility is the steric hindrance of bulky labeling reagents precedingly binding to SAGs, which hinders subsequent binding of the labeling reagent molecule to another SAG. However, this hypothesis seems implausible in the case of ninhydrin, which liberates Ruhemann’s purple product and does not persist on solid surfaces; therefore, this compound does not display a significant steric hinderance effect [9]. A previous study reported that the reactivity of ninhydrin with amino groups of chitosan was lower than that of d-glucosamine, and the reactivity tended to decrease as the molecular mass of chitosan increased [8]; this phenomenon could also explain the low reactivity of ninhydrin. The position of amino groups, namely, the C-2 position of d-glucosamine in ChNWs/ChNFs, can also explain the low reactivity since some previous studies on the derivatization of cellulose indicated that the reactivity of C-2 hydroxyl groups was lower (or lowest in some cases) than that of other (e.g., C-6) hydroxyl groups [16, 17]. However, this explanation also seems insufficient because the amino groups of d-glucosamine and those of chitosan exhibit different reactivities, which have been reported above [8]. The relatively low pKa values (6.3) [3] of the SAGs on ChNWs/ChNFs may contribute to the lower reactivity; reaction conditions for TNBS and OPA, i.e., under pH = 9–10, are close to the pKa values of common aliphatic amines (typically 9–10) but considerably higher than those of the SAGs of ChNWs/ChNFs (6.3) [3]. This difference might lower the reactivity of the latter compounds, although it cannot explain the low reactivity of ninhydrin at pH = 5.2.

The differences observed between the adsorption behaviors of the two employed dyes, RR4 and AO7, are also intriguing, as well as their changes with pH. These dyes, which are both anionic, should electrostatically interact with cationic (i.e., protonated) SAGs of ChNWs/ChNFs. Therefore, the authors employed reaction conditions of pH = 1–3, which are sufficiently lower than the abovementioned pKa value (6.3) of SAGs of ChNWs/ChNFs, for full protonation. The states of the two dyes at these pH values, i.e., whether the dyes act as monovalent anions or not, are important for 1:1 adsorption to the SAGs. A single RR4 molecule contains four sulfonate groups, some of which might be dissociated at pH = 1–3 since the pKa for RR4 is reported to be 4.4 [18]. Therefore, these experimental pH values do not guarantee an average charge of −1 per RR4 molecule; some RR4 molecules might contain multiple dissociated sulfonated groups. These RR4 molecules might interact with several SAGs simultaneously, resulting in poor quantitative properties. In contrast, AO7 is a monosulfonate, and the pKa of the sulfonate is 1.0 (corresponding to the lower value of the two pKa values reported [19]). Therefore, under the experimental conditions of pH = 2 and 3, almost all sulfonate groups dissociated to form anions, whereas the SAGs of ChNWs/ChNFs were protonated, providing an ideal 1:1 electrostatic interaction. However, almost half of the sulfonate groups on AO7 were undissociated at pH = 1, resulting in poor quantification. The speculated mechanisms, which sufficiently explain the ideal adsorption of AO7 on ChNWs, fail to explain why the ChNFs exhibit poor reactivity. Although the essential reasons remain unclear, one possible explanation is the presence of liberated water-soluble chitosan molecules in ChNF samples, especially in S-HL-02, which has a high degree of deacetylation. These soluble chitosan molecules (which are possibly generated during deacetylation by harsh treatments with strong alkalinity) may be solubilized in the supernatant of centrifuged samples together with adsorbed AO7, resulting in an increase in the absorbance of the supernatant and a decrease in the apparent SAG content. Another explanation may be related to the harsh alkali deacetylation treatment, as the regions of chitin molecules with loose assemblies increase inside ChNFs; these regions also contain some amino groups that are accessible to (small) titrant alkali molecules but inaccessible to (large) AO7 molecules, leading to underestimation of AO7 (this remains a speculation and still does not successfully explain the similar degree of crystallinity exhibited by the two ChNFs in X-ray analyses, as described by the manufacturer). The differences between the accessibility of a titrant alkali, that of binding reagents (TNBS, OPA, and ninhydrin) and that of dyes (RR4 and AO7), to amino groups inside ChNWs/ChNFs may explain the underestimation of S values attained by spectrophotometry compared with Stit. A previous study reported similar results for the low reactivity of TNBS and OPA with SAGs on solid surfaces [7].

Conclusions

Three different amino-labeling reagents (TNBS, OPA, and ninhydrin) and two anionic dyes (RR4 and AO7) were examined to quantify the SAGs of ChNWs/ChNFs. From the obtained results, the authors reached the following conclusions.

  1. a.

    For the evaluation of ChNWs only, the use of AO7 at pH = 2 or 3 yields SAG content values, SAO7, which are directly comparable with the titrated values, Stit, as follows:

    $$\begin{array}{cc}{S}_{{{{{{\rm{tit}}}}}}}={S}_{{{{{{\rm{AO}}}}}}7} & ({{{{{\rm{at\; pH}}}}}}=2{{{{{\rm{or}}}}}}3,{S}_{{{{{{\rm{tit}}}}}}} < 800{{{{{\rm{\mu mol}}}}}}/{{{{{\rm{g}}}}}})\end{array}$$
    (1)
  2. b.

    For the evaluation of ChNWs and ChNFs within a consistent measurement system, the use of TNBS or ninhydrin is recommended. The obtained values, STNBS and Snin, can be converted into values comparable to Stit according to Eqs. (2) and (3):

    $${S}_{{{{{{\rm{tit}}}}}}}=5.50\cdot {S}_{{{{{{\rm{TNBS}}}}}}} ({S}_{{{{{{\rm{tit}}}}}}} < 3,000{{{{{\rm{\mu mol}}}}}}/{{{{{\rm{g}}}}}})$$
    (2)
    $${S}_{{{{{{\rm{tit}}}}}}}=3.45\cdot {S}_{{{{{{\rm{nin}}}}}}} ({S}_{{{{{{\rm{tit}}}}}}} < 3,000{{{{{\rm{\mu mol}}}}}}/{{{{{\rm{g}}}}}})$$
    (3)