H₃BO₃, Na₂B₄O₇⋅10H₂O, Na₂B₈O₇⋅10H₂O were obtained from Shanghai Su Yi Chemical Reagent Co., Ltd., China. Craft paper (70 cm × 50 cm) was purchased from Pingxiang County Qunfeng Archive Supplies Fittings Factory, China.

H₃BO₃, Na₂B₄O₇⋅10H₂O, Na₂B₈O₇⋅10H₂O were dissolved in the H₂O with stirring at the mass ratio of 2:3:16. Then, the craft paper was soaked into the above slat solution with ultrasonication for 5 min at 30°C. Moreover, the craft paper was dried for 3 h. Finally, the craft paper is dried in a microwave dryer for 8 min.

To characterize flame-retardant properties of the untreated and treated samples, vertical burning test were carried out according to national standard of China (GB/T 14656–2009), and results are listed in Tables 1 and 2. The flame-retardant performance test results of cardboard in the standard shall be up to the following requirements. 1) Average after-flame time ≤ 5 s; 2) Average after-glow time ≤ 60 s; 3) average char-length 115 mm or less. See the Table 1, the result shows that after flame time and after glow time of untreated craft paper were 22.5 and 165.1 s, respectively. Meanwhile, the char length was 210 mm, which indicates that the untreated craft paper had a high flammability. In contrast, the after-flame time and after-glow time of treated craft paper were only 0.9 s and 28.3 s, respectively, affirming good flame retardancy. To illustrate, Fig. 1 compares the images of the untreated and treated sample at the different burning time. When exposed to the flame, the untreated sample was immediately ignited and quickly burnt out until extinguish at 22 s. In contrast, after ignition, the treated craft paper self-extinguishes in a very short time (∼1 s), showing a significantly enhanced resistance to the fire propagation. Thus, the boron-based flame retardants greatly reduced the inflammability of their modified craft paper, because treated sample is easy to carbonize during the pyrolysis.

The decomposition behaviors of materials are closely related to their flammability. Here TG measurements were conducted to exam the decomposition of the paper samples in nitrogen and air atmosphere. The recorded characteristic temperatures, T_d,5%, T_onset, T_max, D_max and char yield are compared in Fig. 2 and Table 3. In nitrogen, T_d,5% of untreated and treated samples is 70 and 75°C, respectively, due largely to the liberation of absorbed water, see Fig. 2a. The initial decomposition temperature (T_onset) of untreated samples is 325°C, owing likely to the decomposition of the cellulose amorphous region, producing a small amount of water and carbon dioxide. The major decomposition is found at 325–425°C (the mass loss up to 70%), with the maximum (T_max) at 355°C. This finding lies in that the cellulose crystalline region is depolymerized and produces many volatiles. The formation of the chars mainly occurred over 360°C, during which process the paper fibers are gradually carbonized by removing some small molecules. Finally, for untreated paper sample only 18.18% residues at 750°C were found.

The T_onset and T_max of the treated sample were 295 and 335°C, respectively, and are shifted to lower temperature regions compared with the untreated samples. As revealed in Table 3, D_max (rate at T_max) of treated sample are lower than that the untreated sample. This phenomenon indicates that boron compounds decreased the thermal decomposition of fibers, which might be caused that boron compounds promoted the formation of carbonized layer on the surface fibers [43–47]. Meanwhile, the heat would be absorbed by vaporization of their crystal water during the decomposition of the borates, thereby playing a dual function: dilution and cooling [48–50]. With further increased temperature, the thermal decomposition products B₂O₃ from boric acid, borax and disodium octaborate will form glassy melt, covering the sample surface and promoting the charring process. Furthermore, the glassy matter also insulating air and combustibles preventing further pyrolysis (combustion). The char residue of treated is as high as 39.01% at 750°C, which almost 2 times of the untreated.

Figs. 2c, 2d compared decomposition of the untreated and treated samples in an air atmosphere, with three apparent thermal decompositions observed. The treated samples show a higher degradation temperature than untreated samples dose during this decomposition stage (250–400°C). With the increased temperature, the treated samples display reduced degradation and improved thermal stability during the second and third degradation stages. As a result, the residual of the treated samples at 750°C increases by 70% compared with the untreated samples. Such thermal degradation behaviors of the treated samples can be attributed to the earlier decomposition of boron-based flame retardants that could promote the thermal degradation of the paper fiber to form more stable char residual to protect the underlying fiber, and thus improve the thermal stability.

To further investigate the flame retardant mechanisms of boron-based flame retardants on the craft paper, the morphologies of the untreated and treated samples are examined with SEM. Fig. 3 displays the SEM images of the char residues of the untreated and treated paper samples after the vertical burning test. The untreated craft paper fibers are seriously damaged after burning with the fiber residues being thin and broken, and their char residues are cracked with cavities. In contrast, the treated craft paper fibers retain their original morphology (Fig. 3d) with many glass cages formed and evenly coated the surface. So, it is advantageous for boron-based flame retardants to endow strong charring ability and physical barrier against fire with the treated paper. Moreover, formation of protective layer effectively inhibited the depolymerization of the cellulose fiber. This finding agrees with the other reports [51–55]. Furthermore, the boron-containing layer could effectively the underlying fiber from further combustion, producing the strong the self-extinguishing action of the treated sample (Fig. 1).

Elemental composition of char residues of the untreated and treated samples was analyzed by using XPS spectra (Fig. 4). The obtained atomic percentage of char was listed in Table 4. Fig. 4 shows that the treated sample has the additional signals at 191.44 eV(B1 s) which confirms the presence of borate oxide. The treated sample shows B atom of as high as 20.7%. The formed B₂O₃ from thermal decomposition of boric acid, borax and disodium octaborate could improve the barrier property to hold back the heat, air and flammable gases.

Introducing flame retardants will bring negative influence on mechanical performances of materials sometimes. For kraft paper, tensile properties are an important consideration for applications. As compared in Fig. 5, the tensile strength and breaking elongation of the untreated and treated paper samples are 9.41 MPa and 5.87%, respectively. Compared with the untreated sample, the tensile strength and breaking elongation of the treated sample reduced by only 2.5% and 3.9%, respectively. Therefore boron-based flame retardants does not much affect the mechanical properties of the treated paper, because only the small loading needed to achieve desired flame retardancy.

pH value is a very important factor to indicate the aging process of papers, because acidic compounds will form during this process and will accelerate the further degradation. In this study, to examine change of pH values of the paper samples, a large piece of craft paper was cut into equal halves, with one half treated by boron-based flame retardants and the other untreated, and dried at room temperature naturally. The pH values are measured on the any five points of sample and the results are shown in the Fig. 6. Compared with the untreated paper, pH value of treated the craft paper increases markedly, implicating that the treated paper can provide moderate alkaline environment to retard paper aging, which is benefit to the long-term preservation of paper-based archives.