Optimisation of fragrance finishing on cotton by grafting of β-cyclodextrin based microcapsules: Application of the experimental design methodology

Cotton was impregnated into an aqueous solution containing fragrant microcapsules mixed with CA and DHP for 5 min at 195°C. Fabric samples were then squeezed and dried at 90°C for 15 min. Functionalised textiles were finally rinsed and dried again to remove non-grafted microcapsules.

The functionalisation rate of the fabrics was assessed by the mass gain of the samples upon treatment with β-cyclodextrin based microcapsules (expressed as wt.%). The fabric mass gain G was reported as the ratio of the mass difference of dry fabric after and before the treatment with the fragrant microcapsules to the initial mass of dry fabric. It was calculated as a percentage using Eq. (1) as follows: (1) G(%)=m′−mm′×100 G\left( \% \right) = {{m^\prime - m} \over {m^\prime}} \times 100 where (m) and (m′) are the masses of the dry fabric before and after functionalisation, respectively.

RSM was used to optimise the impregnation process as reported by Box et al. [29], D’Agostino and Stephens [30] and Khuri and Cornell [31]. The studied experimental factors were the concentrations of the microcapsules (X₁), the CA crosslinker (X₂) and the Na₂HPO₄ catalyst (X₃). The investigated experimental domain and the corresponding coded levels (−1, 0 and +1) are defined for each factor as shown in Table 1 according to Kumar et al. [32], and Singh et al. [33, 34]. The choice of the levels of each factor is made according to results obtained in section 3.1. A Box–Behnken design was built to fit a second-order quadratic model to experimental data, allowing the prediction of the fabric mass gain G [29]. Fifteen experiments were carried out, including three repetitions of the central point (Table 2). All these experiments were done in triplicate. The Minitab software (Version 19, State College, PA, USA) was used for multiple linear regression, analysis of variance (ANOVA) and statistical analysis.

The shape and surface features of textile fabrics before and after functionalisation were observed by scanning electron microscopy (SEM) after drying using a Jeol JCM 5000 (Jeol, Japan) microscope. Specimens were examined at room temperature without surface metallization at 5 kV and 10 kV acceleration under moderate vacuum.

Tensile test measurements of untreated and treated fabrics were carried out based on the ISO 13934-1 standard by using a Lloyd dynamometer LS series (Ametek STC, France) under standard conditions (20 ± 2°C and 65 ± 2% of relative humidity) after preconditioning all samples in a standardised conditioning room for 24 h. Measurements in column and row directions were performed using a specimen with dimensions of 200 mm length and 50 mm width. The distance between clamps and the movement speed were 50 mm and 10 mm ⋅ s⁻¹, respectively. Three samples were tested for each type of fabric.

Air permeability of untreated and treated fabrics was measured based on the EN ISO 9237 standard by using a FX3300-III instrument (TEX-TEST AG) under standard conditions (20 ± 2°C and 65 ± 2% of relative humidity). Preconditioning of all samples was done in a standardised conditioning room for 24 h. Measurements were performed under an air pressure of 100 Pa per 100 cm² fabric surface. Tests were done in triplicate and the measurements were carried out by scanning the entire textile surface on both sides in order to ascertain the reproducibility of the results. Results were expressed as rates of air flow per unit area and time expressed in “mm ⋅ s⁻¹”.

Wash fastness was studied in accordance with ISO Standard 105-C10 of 2010. Functionalised fabrics were washed for 30 min at 40°C in an Autowash device. Then, specimens were rinsed with water for 5 min and dried under ambient conditions.

The fate of bound microcapsules and the residual amount of fragrance were estimated by SEM observations and gas chromatography (GC) analysis using a 7,890 A Gas Chromatograph from Agilent Technologies (USA) equipped with a flame ionisation detector (FID). Residual neroline after washing was extracted using chloroform. Analytes from the 1 µL samples were injected at a split ratio of 1:20 and separated on a 19091J HP-5 column (length, 30 m; internal diameter, 0.32 mm; film thickness, 250 mm). The oven temperature was initially set at 200°C for 10 min and the detector temperature was 280°C. Nitrogen was used as the carrier gas with a constant flow rate of 20 mL ⋅ min⁻¹ and an inlet pressure of 1.5 bar.

The microcapsules concentration varied between 90 g ⋅ L⁻¹ and 120 g ⋅ L⁻¹ for a CA crosslinking agent concentration of 80 g ⋅ L⁻¹ and a DHP CAT of 50 g ⋅ L⁻¹. The variation of G in Figure 2 shows that the mass gain increased with respect to the concentration of microcapsules according to obvious expectation. However, there was no further mass gain for concentrations >110 g ⋅ L⁻¹as the excess microcapsules clumped together in the reaction medium.

Fig. 2

For a microcapsules concentration of 110 g ⋅ L⁻¹ and DHP CAT of 50 g ⋅ L⁻¹, the effect of the CA crosslinking agent concentration on the mass gain (G) was studied at different concentrations, from 80 g ⋅ L⁻¹ to 140 g ⋅ L⁻¹.

As shown in Figure 3, the mass gain increased with respect to the concentration of CA at levels above [CA] = 120 g ⋅ L⁻¹. However, this ‘optimum’ concentration depends on both the quantity of microcapsules introduced and the textile surface area to be treated. As a polycarboxylic acid, CA acts as a crosslinking agent by reacting with the hydroxyl groups of the β-cyclodextrin molecules in the microcapsules shell on the one hand and the hydroxyl groups of the cellulose on the other.

Fig. 3

For a microcapsules concentration of 110 g ⋅ L⁻¹ and a CA crosslinking agent concentration of 120 g ⋅ L⁻¹, the mass gain percentage (G) was assessed for DHP CATs varying between 50 g ⋅ L⁻¹ and 70 g ⋅ L⁻¹ (Figure 4).

Fig. 4

It appears that the optimum CAT is about 60 g ⋅ L⁻¹. Indeed, the mass gain decreases for higher concentrations. This may be due to the fact that, at higher concentrations, the reaction takes place mainly between the microcapsules instead of between the microcapsules and the cellulose.

Since the preliminary experiments showed that the variations of mass gain with respect to variables were not linear, and that there was a maximum, a second-order quadratic model has therefore been developed to explain the mass gain. To reach this goal, the three experimental factors were simultaneously varied according to a Box–Behnken design. Results were investigated using the software Minitab 19 which led to Eq. (2), where X_i are the coded factors: (2) G(%)= −101.4+0.983X1+0.483X2+ 1.099X3−0.00415X12−0.001611X22− 0.00702X32−0.000087X1X2− 0.00160X1X3−0.002250X2X3 \matrix{ {G\left( \% \right)} \hfill & { = \; - 101.4 + 0.983{X_1} + 0.483{X_2}} \hfill \cr {} \hfill & { + \;1.099{X_3} - 0.00415{X_1}^2 - 0.001611{X_2}^2} \hfill \cr {} \hfill & { - \;0.00702{X_3}^2 - 0.000087{X_1}{X_2}} \hfill \cr {} \hfill & { - \;0.00160{X_1}{X_3} - 0.002250{X_2}{X_3}} \hfill \cr } with Adjusted R² = 0.9204 and Predicted R² = 0.5458. At the 5% level in the Student t-test, all the regression coefficients were significant. The usual determination coefficient R² of the regression equation (Eq. (2)) was 0.9716. All these R²’s proved the satisfactory adequacy of the model. The mass gain ranges from 1.47% up to 5.22% depending on the applied operating conditions (Table 2). Each experiment was carried out in triplicate and the 95% confidence interval of G has been estimated at about ±0.1% (standard deviation ~0.04%) indicating high reproducibility of the G measurements.

ANOVA was first used to evaluate the fitting quality of the model [36]. The assessment is based on an F-test. Results of ANOVA in Table 3 shows the high significance of the regression model for the mass gain prediction (Eq. (2)) (p-value < 0.01).

To finally evaluate the validity of the model, the residuals were also determined by calculating the difference between the experimental G and the G predicted by the model (Eq. (2)). These residuals <0.8% indicated that the mass gain G was satisfactorily explained by the regression model.

Once validated, the model was used to visualise the response surfaces as bi-dimensional plots of two factors (X₁ and X₃), while keeping the third one (X₂) constant. Figure 5 shows the contour plots of the response surface for G at CA crosslinking agent concentrations of 80 g ⋅ L⁻¹, 100 g ⋅ L⁻¹ and 120 g ⋅ L⁻¹.

Fig. 5

The contour plots in Figure 5 confirm the results drawn by varying parameters one by one: the concentrations of microcapsules and CA should be maximum and the concentration of catalyst should be minimum within the studied window. There are synergistic effects, however. The most relevant ones are those between the concentration of catalyst on the one hand and either the concentration of microcapsules or CA on the other hand. The optimum concentration of catalyst drawn from the one parameter at a time study was 60 g ⋅ L⁻¹ for a microcapsules concentration of 110 g ⋅ L⁻¹ and a CA crosslinking agent concentration of 120 g ⋅ L⁻¹. The quasi-symmetrical feature of the contour plots with respect to the diagonal in Figure 5 shows that the optimum of CAT shifts from 60 g ⋅ L⁻¹ to 50 g ⋅ L⁻¹ as the concentration of microcapsules increases. The same trend is observed for an increase of CA concentration.

The optimal conditions of the cotton grafting process using fragrant microcapsules were predicted by the response optimiser tool of Minitab 19 software for maximum response. Response optimisation described in Figure 6 shows that optimal experimental conditions are a microcapsules concentration of 107.6 g ⋅ L⁻¹, a CA crosslinking agent concentration of 111.9 g ⋅ L⁻¹ and a DHP CAT of 50 g ⋅ L⁻¹, which yields a mass gain of 5.32%.

Fig. 6

To verify the validity of the model, the corresponding experiment was performed to compare the experimental results with the response predicted by the model. The theoretical optimal value of mass gain under these optimised conditions is 5.32% while the experimental one is 5.58%. The difference between these two values is well within the 95% confidence interval of the experimental error, thus proving the good predictive performance of the developed model.

Grafted cotton fabric was analysed by SEM after being treated in a bath containing fragrant β-cyclodextrin based microcapsules, CA crosslinking agent and DHP catalyst, under optimised conditions. Figure 7 shows SEM micrographs of the treated textile at ×500 and ×2,000 magnifications (Figure 7B) via the previously optimised grafting protocol compared to the non-treated one at ×500 magnification (Figure 7A). These observations confirm the efficacy of the optimised grafting process designed from the ANOVA results. Nevertheless, the present SEM data does not provide a visual comparison allowing an assessment of the optimised efficacy of the process since pictures were collected on fibers grafted using the optimised process only. Indeed, it is difficult to infer binding density of microcapsules from SEM images because it requires a large number of images of the same sample for obtaining statistical significance. We did collect several pictures for ensuring that the pictures shown in Figure 7 were representative; the number of pictures we collected was not sufficient for performing a statistical analysis. Owing to these constraints, the determination of grafting density by SEM is not to be done.

Fig. 7

Though SEM does not provide a quantitative analysis, SEM pictures show that a sufficiently high density of grafting has been reached. Pictures reveal a large collection of grafted microparticles distributed evenly onto the cotton textile fabric: between the textile yarns, the fibers and in the cotton cavity. The density of grafted microcapsules was similar to that of previous work obtained by the impregnation process using PU or epoxy resin [37] shell microparticles based on β-cyclodextrin [14, 23] or isosorbide [38, 39].

Fragrant microcapsules are not damaged following the heat treatment undergone during the process (grafting and drying temperatures of 195°C and 90°C respectively). The microparticles are also tightly fixed by means of covalent bonds after the reaction and a simple rinse.

The effect of the optimised grafting treatment on the tensile properties was evaluated by a comparison between the tensile curves of the treated and the non-treated samples in both knitting directions (row and column directions), and in particular, the maximum force, the elongation at maximum force as well as the stiffness (Figure 8).

Fig. 8

The general tensile behaviours of the different knitted fabrics were similar in the two stress directions and whether they were measured before or after functionalisation. Tensile strength at low loads mainly originates from the friction resistance of yarns in the wales [40]. Meshes progressively line up leading to fabric elongation. The yarns get deformed at higher loads and eventually neck and break.

A comparison between the maximum force, the elongation at maximum force as well as the stiffness of the untreated and the treated fabrics is shown in Figure 9.

Fig. 9

The treatment of grafting the microcapsules to the cotton knitwear increased the maximum force in both directions. The establishment of new covalent bonds contributes, in some way, to the improvement of the tensile strength. The elongation at maximum force is slightly higher for the grafted textile than that of the untreated textile. The cause of such changes could be the treatment at high temperatures and without applied tension during grafting that increases the amorphous areas of already mercerised cotton. Hence, the rate of crystallinity decreases and the strain at maximum force increases slightly in both directions of stress.

Examination of Figure 9 shows that the grafting treatment increases the stiffness of the textile stressed in both directions. This increase is more accentuated in the column direction rather than in the row direction, which is probably due to the establishment of new bonds, inter-fibers and inter-yarns, of covalent types by the grafting process. According to these results, it appears that the optimised grafting treatment slightly modifies the mechanical properties of the textile without altering them.

The assessment of air permeability was investigated by measuring the rate of air flow through the fabric under differential pressure.

The air permeability properties of knitted cotton fabrics before and after treatment are displayed in Figure 10. The treatment of grafting microcapsules decreased the air permeability of the functionalised textile by 50% for a mass gain of 5.58%. The textile support is not yet saturated and occluded with bound microcapsules, since porosity between the fibres was clearly seen in SEM pictures. There is no excess of the grafted microcapsules forming an occlusive layer right side up and on the other side of the textile substrate. Microcapsules are well distributed inside the textile sample in the sites between the thread and in statements between the fibers.

Fig. 10

Misra et al. [41] in 2020 reported that fabrics become less permeable to air upon increasing the mass per unit area of the textile. Although the air permeability decreased from 992 mm ⋅ s⁻¹ to 492 mm ⋅ s⁻¹, the wearing comfort is not compromised. Given the low mass per unit area of the jersey structure, the very high level of openness and air permeability of the textile support before treatment allows for a large mass gain without loss of sensory properties.

The uniformity of grafted microcapsules was confirmed by air permeability measurements performed at different points of the fabric. All tests presented similar air permeability behaviour. The present treatment is efficient in, not only promoting a high level of material attachment but also ensuring the coating uniformity [42].

The effect of washing cycles on grafted textile was investigated according to the ISO 105-C10 standard of 2010 to evaluate the efficiency and the durability of the microcapsules application to the knitted fabric. SEM observations and GC analysis were used to assess the lifetime of scent textiles both as qualitative and quantitative evaluations.

SEM micrographs of the functionalised knitted fabric after 40 washing cycles (Figure 11) show the presence of microcapsules remaining bound to the textile. The amounts of microparticles on the fabric before and after washing are similar; no noticeable fall could be detected. Thus, unlike the impregnation treatment, the grafting treatment provides a good wash fastness. Indeed, grafted microcapsules still remain attached to the textile support by means of covalent bonds. Compared to their original shape in the dispersed state, β-cyclodextrin-based microcapsules are slightly deformed following the release of the active ingredient by diffusion through the capsule membrane. For such a finishing process, in addition to the release of the encapsulated product, there is the possibility for the capture of sweat or cigarette smoke by the β-cyclodextrin, which considerably reduces the intensity of odours on the clothes. These encapsulated substances are removed in the washing machine, and the β-cyclodextrin is again free to trap odours. Textiles grafted with β-cyclodextrin-based microcapsules can also be selectively reloaded with a large diversity of guest substances by dipping or spraying methods. 4Particular interest is the treatment with perfume, antimicrobial or pharmaceutical products which can be released continuously over a long period of time upon exposure to moisture. The higher the humidity level, the more the encapsulated substance that is released. Such modified textiles may ‘respond’ to a variety of environmental conditions [35]. These two methods of reloading the β-cyclodextrin prolong the effect of the functionality of the studied grafting process [43].

Fig. 11

The presence and the quantification of residual encapsulated neroline after washing cycles were studied by GC analysis. The area of the fragrance peak decreased as repeated washing cycles have been applied. The amount of residual neroline was determined based on a calibration curve.

Figure 12 shows that the residual neroline concentration decreased slowly, confirming that the microcapsules remained loaded with perfume along subsequent washing cycles.

Fig. 12

Residual neroline concentration on the textile treated with β-cyclodextrin-based microcapsules is lower than that obtained by the impregnation process for the same mass gain. The high grafting temperature causes a partial loss of about 37% of the encapsulated perfume during the polycondensation reaction (temperature of 195°C). However, the effectiveness of the textile grafted with β-cyclodextrin-based microcapsules appeared higher than that of in our previous works [39] and those of Rodrigues et al. [44, 45] since the present fragrance loss was about 2% after five washes and 77% after 40 washes. This can be attributed to the fact that in the case of the grafted textile, there is no significant loss of the microcapsules following the washing cycles.

Thereby, this decrease in perfume concentration is essentially due to a diffusion phenomenon of the active material through the pores of the PU microcapsules and the cavity of the β-cyclodextrin. These results confirmed not only the efficient encapsulation of neroline and the features of a controlled release of the active material but also the efficiency of the grafting process. Microencapsulation is an effective method that allowed the control of neroline diffusion through the pores of the microparticles shell protecting the active ingredient from the environment on one hand. On the other hand, it appears that the grafting process further contributes to improving the lifetime of scent textiles.