3.1 Bleaching performance
The Response Surface Methodology (RSM) based on Central Composite Design (CCD) was used to evaluate and optimize the hot peroxide bleaching process for newly synthesized TBUCB. The Analysis of Variance for WI is given in Table 4.
Coefficients of determination (correlation coefficient) (R2), ranging from 0% to 100%, indicates the percentage of variation in the response that is explained by the model. The higher the R2 value, the better the model fits the response data. As can be seen from Table 4, the model is adequate to provide an accurate prediction of response function since R2 > 0.9 and lack of fit P value >0.05. R2 is found to be as 99.23%, indicating that 99.23% of the response variability can be explained by the model. R2and Adj- R2values are higher than 0.9, providing a high trend between the experimental and the predicted values.
Concentration of activator (A), molar ratio of activator:H2O2 (B) and temperature (D) are found to have significant influence on WI due to the higher F-ratio (178.94, 187.78, 869.35, respectively) and the smaller significance level (P value<0.05). Temperature (D), followed by molar ratio of activator: H2O2 (B) and concentration of activator (A), respectively, present the highest statistical relevance since a higher F ratio indicates greater relevance of the corresponding factor. Although all the interactions between the factors were analyzed, the interactions between the concentration of activator (A) and the molar ratio of activator: H2O2 (B), between the concentration of activator (A) and the temperature (D), between the molar ratio of activator: H2O2 (B) and the temperature (D) were statistically significant. In terms of F value, the two-way interaction between the concentration of activator (A) and the molar ratio of activator: H2O2 (B) is greater than the other two-way interactions (F= 17. 75).
2D contour plots, representing a 3D surface on a two-dimensional plane, illustrate the response function of two factors while keeping all other factors constant for the optimization of process conditions. Figure 2(a)-(j) show the two-way interactions of process parameters on WI while all other parameters are at fixed level. It was observed from Figure 2(a) that a whiteness index higher than 70 was obtained at 9.47 g / L (at 0-level) and higher activator concentrations and at a molar ratio of 1: 6-1: 10 activator: H2O2.
It is seen from Figure 2(b) that the maximum WI can be obtained at 11.7 g/L of activator concentration (at 1-level) and the increase in the activator concentration to a certain value (approximately to 12.8 g/L) either at a low or high level of alkali (NaOH) range used in the study causes a decrease in the WI. Figure 2(c) indicates that WI higher than 80 can be achieved between 11.7 g/L and 13.93 g/L of activator concentrations at a temperature of 80 oC. Similar with the trend of alkali concentration in Figure 2(b), (e) and (h), process time has no significant effect on WI as shown in Figure 2(d) (g) and (j). A whiteness index higher than 75 was obtained at a ratio of 1: 6-1: 10 activator: H2O2 at a temperature of about 68 to 80 ° C (Figure 2(f)), indicating that peracid became more active resulting in improved bleaching efficiency.
3.2 Computational Results
Mapped Surface for the LUMO on the TBUCB activator given in Figure3a is localized on the butanoyl and butyrolactam groups, especially on the carbonyl groups. One can expect the attack of perhydroxyl anion should be directed at these groups because they are the most electron deficient sites in the activator structure. HOMO orbital which has low possibility for the nucleophilic attack is mainly on the chloride anion and partly on the ethyl ammonium groups (Figure 3b). ESP surface showed that carbonyl groups on the butanoyl and butyrolactam are the only groups that have both electron rich part, depicted by red color on oxygen which can form a hydrogen bond with perhydroxyl anion, in addition to the electron deficient carbonyl carbon depicted by blue color which can undergo nucleophilic attack by peroxide anion, perhydroxyl (Figure 3b). Although there are two carbonyl groups, only the carbonyl carbon of butanoyl is the one that can lead to the formation of peracid (Figure 3c). Atomic charges on the carbon atoms were compared by two different methods to determine which one will be preferred for the anion attack. Both ESP charges based on the electrostatic potential fitting method and charges based on the partitioning of the molecular electron density showed higher positive charge on the carbon atom of the butanoyl carbonyl (Figure 3d-e), which is the main driving force for the initial diffusion and attack by perhydroxyl anion. At last, Fukui reactivity indices for the nucleophilic attack were compared for the two carbonyl carbons, which showed higher potential of the cationic bleach activation to undergo nucleophilic attack by peroxide anion at butanoyl carbonyl carbon rather than butyrolactam carbonyl carbon (Figure 3f). Fukui reactivity indices based on different atomic charge calculation methods gave similar results.
Reactions were found to follow two step mechanisms, which are perhydroxyl anion attack at the carbonyl carbon and the peracid formation. Transition state characterized by the single imaginary frequency was calculated for the perhydroxy anion attack at the TBUCB and free energy profile is given in Figure 4a. Reaction barrier for the perhydroxyl anion attack at the carbonyl carbon calculated by using sum of electronic and thermal free energies at 70˚C was determined as 12.55 kcal/mol under water solvation effect. This activation barrier was compared with the TAED activation barrier, and it was determined that perhydroxyl anion attack has lower reaction barrier for TBUCB determined as 13.72 kcal/mol in addition to the higher solubility than TAED (Figure 4b). There is an interaction between carbonyl carbon and perhydroxyl anion as well as a hydrogen bond formation between the oxygen of the carbonyl and perhydroxyl hydrogen in the transition state. This confirms the importance of initial hydrogen-bonding through these interactions that leads to the nucleophilic attack. Although carbonyl carbon that is part of the butyrolactam has a partial positive charge as a potential site for nucleophic attack, no transition states for perhydroxyl anion attack at the carbonyl carbon of butyrolactam ring was found, which confirms that there is no reaction happening inside ring, which led to the conclusion that the carbonyl group of butyrolactam is not reactive towards perhydroxyl anion attack to form any peracid. The bond distance between carbonyl carbon of butanoyl and the nitrogen atom of butyrolactam significantly increased after perhydroxyl bonding, followed by bond dissociation and peracid formation in the second step of the reaction. The perhydroxyl anion attack in the first step is the rate determining step, and dissociation of peracid for TBUCB and TAED in the second step occurs with a lower energy barrier at 7.88 and 4.58 kcal/mol, respectively. Imaginary frequencies and cartesian coordinates of atoms for the transition states are given in the Supporting Information.
The main limitation of TBUCB is the conformational freedom due to the alkyl group in butanoyl group. Results showed that positively charged triethyl ammonium group is attracted toward the more negative oxygen atoms of either butanoyl (Figure 5a and 5c) or butyrolactam (Figure 5b and 5d) groups, resulting in a bent conformation with decreased reactivity, thereby preventing more peracid formation. In addition, steric interactions cause a barrier for perhydroxyl ion to reach to the activator in bleaching bath for peracid formation. Although all trans coplanar conformation is the lowest energy structure, gauche1 (Figure 5a and 5c) and gauche2 (Figure 5b and 5d) conformers of these alkyl groups were easily obtained with less than 2 kcal/mol energy difference. The closest distance between ethylammonium hydrogens and butanoyl oxygen which was 4.51 Å in the lowest energy structure, decreased to 2.51 Å in gauche1 structure. The closest distance between ethylammonium hydrogens and butyrolactam oxygen which was 6.02 Å in the lowest energy structure, decreased to 4.51 Å in gauche2 structure (Figure c-d).
When phenyl is used in the molecule structure (Figure 5e) similar with the structure reported as N-[4-(triethylammoniomethyl)benzoyl]caprolactam chloride (TBCC) in the literature [Hou, Zhang & Zhou, 2010; Xu, Hinks, & Shamey, 2010a; Luo et al., 2015], the interaction between triethyl ammonium and the oxygen atoms of the carbonyl group decreased since the butanoyl segment can adopt different conformations while the phenyl group is highly rigid. However, the reactivity of the butanoyl carbonyl towards nucleophilic attacks significantly decreased by phenyl substitution. Alkene modification can be a better alternative to achieve more rigid structure to prevent folding of this aliphatic chain in the future studies (Figure 5f). Unlike butanoyl group, rigid conformations of alkene group hinder coplanar structure from intramolecular interaction and keep the plane of molecule parallel to cellulose surface without compromising nucleophilic reactivity. In general, future work of the developments of novel activator should focus on more rigid, linear and coplanar bleach activator/peracid to achieve better substantivity for cellulose, which would mimic the substantivity of direct dyes to cellulose.