Engineering of super bactericidal cotton using pyridinium/di-N-chloramine siloxane with intensified synergism

Tuning the ratio of complementary biocidal groups in a composite unit is proved to be a tactic to better minimize their weaknesses to realize higher synergism. A silane with precursors of one pyridinium and two N-chloramine sites, 6-(pyridin-4-yl)-3-(3-(trimethoxysilyl)propyl)-1,3,5-triazinane-2,4-dione, was synthesized, hydrolyzed and dehydrocondensed into a silicone modifier on cotton cellulose. Specially, isonicotinaldehyde was ammonolyzed with biuret to produce 6-(pyridin-4-yl)-1,3,5-triazinane-2,4-dione that subsequently reacted with (γ-chloropropyl)trimethoxysilane to synthesize the silane through nucleophilic substitution. The modifier on cotton was quaternized and chlorinated to transform the one pyridine and two amide N–H structures in each unit of the silicone to pyridinium and N-chloramine counterparts. The cationic pyridinium increases the hydrophilicity of the unit and draws anionic bacteria to its two adjacent highly fatal N-chloramine sites, achieving a faster contact-killing rate than not only monofunctionality but also basic synergistic integration of one cationic center and one N-chloramine. This phenomenon is therefore referred to as “intensified synergism” and provides crucial information for the design of more powerful biocides. The pyridinium/di-N-chloramine silicone coating exhibited extraordinary durability towards UV irradiation, washing cycles and long-term storage due to the good UV resistance and chemical inertness of pyridinium and silicone backbone.

the one pyridine and two amide N-H structures in each unit of the silicone to pyridinium and N-chloramine counterparts. The cationic pyridinium increases the hydrophilicity of the unit and draws anionic bacteria to its two adjacent highly fatal N-chloramine sites, achieving a faster contact-killing rate than not only monofunctionality but also basic synergistic integration of one cationic center and one N-chloramine. This phenomenon is therefore referred to as ''intensified synergism'' and provides crucial information for the design of more powerful biocides. The pyridinium/di-N-chloramine silicone coating exhibited extraordinary durability towards UV irradiation, washing cycles and long-term storage due to the good UV resistance and chemical inertness of pyridinium and silicone backbone.

Introduction
The fight against bacterial infection has run through the history of mankind. Despite the great improvement of the sanitary conditions, bacterial contamination still causes loss of billions of dollars worldwide every year (Ding et al. 2018). Antibacterial modification of material surfaces is one of the most commonly used strategies for contamination control. The modification of cotton has attracted particular attention owning to its wide usage in daily life and hygiene field while the hydrophilicity resulting from surface hydroxyl groups facilitates the growth and reproduction of pathogenic microorganisms Ren et al. 2009). Therefore, the development of biocidal coating on cotton for better performance than reported ones is still highly desirable.
Attachment of biocides to cotton needs a friendly and efficient modification method that can be cataloged as a physical or chemical one. Physical methods such as layer-by-layer assembly Gomes et al. 2015) and interpenetration in supercritical carbon dioxide (Chen et al. 2013) have been applied to the formation of biocidal coatings. The application of physical methods is not the mainstream solution because layer-by-layer assembly works well on highly charged substrates while cotton is only slightly charged and impenetration in supercritical carbon dioxide needs high pressure vessels that are usually expensive and hard to operate. In contrast, cotton is usually modified by chemical methods since it has surface reactive hydroxyl groups. However, the areal density of hydroxyl group of cotton is relatively low so that potent biocidability cannot be achieved when each hydroxyl group only bonds with one biocidal functionality. The problem is often addressed using two strategies. One is the generation of free radicals on cotton for initialization of polymerization of biocidal monomers to compensate the sparse hydroxyl sites (Luo et al. 2017;Luo and Sun 2006;Ma et al. 2015). The second and most used one is the employment of silanes as carriers of biocidal groups since each of their hydrolyzates has three silanol groups that can condense with both counterparts in other hydrolyzates and hydroxyl groups on surface of cotton to form crosslinked polymeric silicone coatings (Chen et al. 2020;Kou et al. 2009;Ren et al. 2008).
Cotton was initially decorated with single functionality including N-halamines (N-chloramines and N-bromamines) (Cheng et al. 2015;Luo et al. 2017;Zhang et al. 2019), antibiotics Qu et al. 2019), metals and metal oxides (El-Rafie et al. 2014;Ibrahim et al. 2019;Xu et al. 2018), and cationic salts (quaternary ammonia salts and pyridinium ions, etc.) (Przybylak et al. 2018;Zhang et al. 2018b). Nhalamines and cationic salts have stood out due to their low cost, broad-spectrum efficacy, and abound types. However, the use of single biocidal group has an inherent disadvantage since bacteria have a vast diversity of structures and hence it is hard for one functionality to possess all desired properties (Ates and I Cerkez 2017). For example, cationic salts kill bacteria through penetration into the anionic cytoplasmic membranes and hence are not efficient in eliminating Gram-negative bacteria that have thick cellular walls (Liang et al. 2006). N-halamines are relatively hydrophobic and cannot sufficiently contact with bacterial suspension so that a higher concentration sometimes does not lead to a faster killing ). Moreover, N-halamines are consumed during the killing process and the substrate gradually loses biocidability before regeneration.
To address the disadvantage of single functionality, combination of different types of functionalities is used as a tactic to diminish the drawbacks of each of them. For instance, silver was integrated with cationic polymers to modify cotton (Chen et al. 2015a) and PET (Zhang et al. 2018a) for better performance. Since cationic salts and N-halamines have complementary properties, their combination is hypothesized to have good synergism. In such a composite unit, the cationic salt can increase the hydrophilicity to address the hydrophobicity of N-halamine, electronically attract anionic bacteria to N-halamine, and provide certain biocidability when N-halamine sites are consumed entirely (Kang et al. 2013;Li et al. 2012;Liu et al. 2013). That is, although cationic salt is a mild biocidal group yet it can facilitate the contact of the highly lethal N-halamine with bacteria to achieve a faster killing compared with any functionality alone. Furthermore, since the biocidal ability of the cationic salt is auxiliary and the main contributor of biocidability is N-halamine, we further hypothesized that the combination of one cationic center and multiple Nhalamines can realize even higher antibacterial efficacy (referred to as intensified synergism) than the counterpart of one cationic center and one N-halamine (basic synergistic format).
Besides the combination, several other important principles should be beared in mind when designing a composite unit with not only promising biocidability but also other desirable properties. Firstly, it is desirable that the composite unit is polymerizable since polymers do not penetrate into human skins, have longer life yet lower toxicity (Krishnan et al. 2006). The second rule is that the cyclic N-halamines are preferred since they are more stable than the acyclic counterparts (Dong et al. 2010;Liang et al. 2007). Moreover, it is beneficial to conjugate Nhalamine with hydrophilic neighbors to address its hydrophilicity (Kou et al. 2009;Ren et al. 2009).
A composite unit of one cationic center and several cyclic N-halamines with hydrophilic neighbors is then assumed to have better overall properties than current ones. To test the hypothesis, a silane with a composite unit that bears biocidal precursors of a cationic pyridinium and two amide N-chloramine sites in a hydrophilic 6-membered ring was design and synthesized herein as shown in Scheme 1. Isonicotinaldehyde was firstly reacted with biuret to produce 6-(pyridin-4-yl)-1,3,5-triazinane-2,4-dione that was reacted with (c-chloropropyl)trimethoxysilane (CPTMO) to synthesize the silane named 6-(pyridin-4-yl)-3-(3-(trimethoxysilyl)propyl)-1,3,5-triazinane-2,4-dione via base-catalyzed nucleophilic substitution. Such a silane meets all of the previously discussed requirements. This is because its pyridine and two amide hydrogens can be converted to complementary pyridinium and amide N-chloramines to satisfy the requirements of intensified synergism due to the combination of one cationic center and two Nchloramines, hydrophilicity due to the water soluble pyridinium and the 6-membered ring, stability due to the cyclic N-chloramine structure. Finally, the saline is polymerizable after hydrolysis via dehydration condensation to fulfill the requirement of the formation a polymeric biocidal layer on cotton. The studies herein proved the assumed virtues of this combination.

Materials
Biuret was purchased from Nine Ding Chemistry (Shanghai) Co., Ltd. Isonicotinaldehyde was obtained from Shanghai Macklin Biochemical Co., Ltd. HCl (36%) was supplied by Yantai Far Eastern Fine Chemical Co., Ltd. KOH and anhydrous ethanol were provided by Chengdu Kelong Chemical Co., Ltd. Aqueous NaClO (10%) was purchased from Tianjin Guangfu Fine Chemical Co., Ltd. Na 2 CO 3 was supplied by Tianjin Bodi Chemical Industry Co., Ltd. (c-Chloropropyl)trimethoxysilane was purchased from Shandong West Asia Chemical Co., Ltd. 1-Chlorohexane was purchased from Sinopharm Chemical Reagent Co., Ltd. KI was purchased from Tianjin Jinbei Fine Chemical Co., Ltd. Cotton swatches were purchased from Dongguan Yunfan Textile Co., Ltd. Escherichia coli and Staphylococcus aureus were purchased from Guangdong Institute of Microbiology. All other chemicals were obtained from Shanghai Macklin Biochemical Co., Ltd.

Characterization
Infrared spectrum of each produce was illustrated by a Thermo Nicolet Magna IR-560 spectrometer using transmission technique (KBr pellet). The spectra were collected at 0.5 cm -1 resolution and 8 scans in the 400 * 4000 cm -1 wavenumber range.
Cotton fibers were vacuum-coated with platinum using a 108auto sputtering coater (Cressington scientific instruments Ltd.) and then characterized with a FEI Nano SEM450 field emission scanning electron microscope (SEM) at an accelerating voltage of 15 kV under a chamber pressure of 1 9 10 -4 Pa to analyze the morphology of the biocidal coating.
Proton nuclear magnetic resonance ( 1 H-NMR) study was performed with a Bruker Avance III HD 500 MHz spectrometer using DMSO-d6 as solvent.
X-ray photoelectron spectroscopy (XPS) spectra of samples were detected with a Thermo Scientific Escalab 250Xi spectrometer installed with an Al Ka monochromatic X-ray source. Spectral acquisitions were performed under a chamber pressure of 1 9 10 -6 Pa at a test angle of 45°. Wide scans (1-1000 eV) were acquired at an analyzer pass energy of 100 eV and a resolution of 1 eV while high resolution scans were recorded at an analyzer pass energy of 23.5 eV and a resolution of 0.05 eV.
Binding energy (BE) of aliphatic carbon (C 1s ) was set to 284.6 eV for calibration of charging effects.

Biocidal modification of cotton fibers (Scheme 2)
The pH value of the above solution was adjusted to * 5 with CH 3 COOH. The methoxysilyl groups originating from (c-chloropropyl)trimethoxysilane were hydrolyzed for 20 min to silanol groups. Cotton swatches were immersed into the hydrolysis solution for 15 min. The immersed swatches were taken out and cured at 100°C in a vacuum oven for 1 h to complete the dehydrocondensation. The swatches were then washed in an ultrasonic oscillator to remove physically adsorbed impurities and dried in the air.
The as-prepared swatches were refluxed in solution of 2 mL 1-chlorohexane and 20 mL ethanol for 6 h to quaternize the pyridine rings to pyridinium ions. The quaternized swatches were dried in the air and then chlorinated with 10% NaClO solution for 6 h at the room temperature to convert the N-H bonds to N-Cl formats. The chlorinated fibers were then rinsed with Scheme 1 Synthesis of 6-(pyridin-4-yl)-3-(3-(trimethoxysilyl)propyl)-1,3,5-triazinane-2,4-dione distilled water to remove free chlorine and dried at ambient temperature. Each unit of the silicone coating has one pyridinium and two N-chloramine sites in the best situation and the then final sample was referred to as pyridinium/di-N-chloramine cotton in the study.

Evaluation of the amount of oxidative chlorine
The determination of the total amount of oxidative chlorines (Cl ? ) is necessary for biocidal analysis and comparison with other systems. The total amount is evaluated using an iodometric/thiosulfate titration method from the formula below (Cerkez et al. 2016;Zhang et al. 2019): where N is the concentration (mol/L) of the Na 2 S 2 O 3 titration solution, V Cl ? and V 0 represent volumes (L) of Na 2 S 2 O 3 solution consumed in the titrations of pyridinium/di-N-chloramine cotton swatches and controls, respectively, and W denotes the weight in grams of the titration swatches.

Evaluation of biocidal efficacy
Biocidability of pyridinium/di-N-chloramine cotton samples was evaluated using gram-positive S. aureus and gram-negative E. coli as representative bacteria in accordance with the ''sandwich test'' method (Sun and Sun 2004;Zhao and Liu 2011). Both bacteria were allowed to grow at 37°C under 250 rpm overnight in broth medium. Afterwards, the cells were centrifugally harvested, washed and diluted with phosphate buffered saline (PBS) solution to prepare suspensions of known concentrations (CFU). 50 lL of each bacterial suspension was sandwiched in the center of two 1 in 2 pyridinium/di-N-chloramine cotton swatches that sufficiently contacted with the suspension by placing a weight on the top. After contact time of 3, 5, and 10 min, the remaining N-chloramine sites of swatches were quenched with 10 mL of sterile 0.02 N Na 2 S 2 O 3 in a centrifuge tube. After vortex, serial dilutions of the quenched suspension with PBS were placed on Luria-Bertani agar plates at 37°C overnight for record of the number of colonies to calculate antibacterial efficacy. Some pyridinium/di-N-chloramine cotton swatches were first quenched with excessive 0.02 N Na 2 S 2 O 3 and subsequently subjected to evaluation following the same procedure for the analysis of biocidal efficacy of pyridinium functionality only (denoted as pyridinium functionalized cotton). Pristine cotton swatches without any modification were used as the controls. Each sample was assayed in triple and the average value was reported.
Scheme 2 Preparation of pyridinium/di-N-chloramine cotton Investigation of biocidal durability and rechargeability The durability and rechargeability of the N-chloramine sites in the modifier provide crucial information of the application value of the pyridinium/di-Nchloramine cotton. The effects of washings, UV irradiation and storage on durability and rechargeability of the N-chloramine sites were then assayed. The durability and the rechargeability of N-Cl bonds under repeated washings were evaluated using AATCC Test Method 61-1996 . Pyridinium/di-N-chloramine cotton swatches with a size of 1 9 2 inch were subjected to repeated washing cycles in 150 mL of 0.15 wt% aqueous AATCC detergent solution in a canister containing 50 stainless steel balls. The canister ran at 49°C and 42 rpm for 45 min to accomplish one washing cycle that is equivalent of five machine washings. After 5, 10 and 15 washing cycles, each sample was washed for three times with distilled water, dried at ambient temperature afterwards, and then titrated for the determination of oxidative chlorines. Some washed samples were recharged with NaClO solution following previously described process and then titrated to assay the durability of the biocidal silicone layer and the rechargeability of hydrolyzed N-chloramine sites. Each reported value was averaged over three measurements.
The durability and rechargeability of the N-chloramine sites in the modifier of pyridinium/di-Nchloramine cotton swatch under UV irradiation (340 nm) were tested in an accelerated weathering tester (Q8 model, Hongzhan Group) at 25% RH and 20°C over a 7-day period. For each period, one set of irradiated swatches was directly titrated for the determination of the amount of the remaining oxidative chlorine, and a second set was rechlorinated using NaClO solution and then titrated for the estimation of the proton-initiated decomposition of the silicone modifier and N-chloramine sites.
The storage stability of the biocidal function was evaluated by keeping pyridinium/di-N-chloramine cotton swatches at 25°C and 65% RH under laboratory light over a 30-day period. Some swatches were titrated directly and some were rechlorinated and then titrated to investigate the storage stability of Nchloramine structures.

Biocidal modification of cotton fibers
The synthesis and hydrolysis of the silane monomers and subsequent condensation polymerization of the hydrolyzates on cotton are accomplished using a convenient one-pot procedure as previously illustrated in Scheme 2. The success of the formation of the silicone modifier cotton can be verified by comparison of the FTIR spectra of cotton before (Fig. 2a) and after the polymerization (Fig. 2b). After the polymerization, the spectrum of the sample (referred to as silane modified cotton) shows characteristic bands originating from the silicone unit including the amidic N-H stretching vibration at 3248 cm -1 , the C¼O stretching vibration at 1690 cm -1 , the C¼N stretching vibration at 1602 cm -1 , the Si-O-C stretching vibration and 1259 cm -1 and the Si-O-Si stretching vibration at 800 cm -1 . Next modification is to quaternize the pyridine rings in the modifier to pyridinium ions with 1-chlorohexane to produce quaternized cotton. Correspondingly, the stretching vibration of C¼N at 1602 cm -1 was shifted to higher frequencies and merged with the binding vibration of O-H at 1639 cm -1 as shown in Fig. 2c. Eventually, the two amidic N-H in each unit were chlorinated to N-chloramines with NaClO to produce pyridinium/di-Nchloramine cotton, which resulted in the disappearance of N-H stretching vibration at 3259 cm -1 and a blueshift of C=O stretching mode from 1689 cm -1 to a higher frequency of 1724 cm -1 . The bluefshift is the evidence of the breakage of N-HÁÁÁO=C hydrogen bonding and increase of atomic weight resulted from the transformation of N-H to N-Cl Sun and Sun 2001).
Other characterization is required for unequivocally confirmation since the coating is relatively thin, which in turn leads to subtle of changes of some FTIR signals. In this sense, XPS spectra are good supplements since this technique is very surface sensitive and only acquires chemical compositions of the top * 5 nm when placing a sample at a test angle of 45°with respect to the X-ray beam. In contrast with the one of original cotton (Fig. 3a) that only displays signals of carbon (C 1s ) at 285 eV and oxygen (O 1s ) at 533 eV, the XPS survey scan of pyridinium/di-N-chloramine cotton (Fig. 3b) exhibits additional photoelectron peaks at 102, 153, 202, 272, 401 eV that agree well with binding energies of Si 2p , Si 2s , Cl 2p , Cl 2s , and N 1s . These new characteristic peaks witness the presence of the biocidal coating layer on cotton fibers.
The chemical states of nitrogen and chlorine are further examined with high-resolution spectra to ensure the correct formats of pyridinium and Nchloramine on pyridinium/di-N-chloramine cotton. As shown in Scheme 2, there are four types of nitrogen atoms according to the chemical environments. The N 1s peak was correspondingly fitted into components at 402.2 eV for the cationic nitrogen in pyridinium ions (Chen et al. 2015b), at 400.6 eV for covalent nitrogen in imide bond (Dong et al. 2014), at 399.8 eV for covalent nitrogen in amide N-chloramine (Tamura et al. 2012), and at 399.1 eV (Chen et al. 2015b) for residual pyridine due to the incompletion of quaternization with an areal ratio of * 0.7:1:2:0.3 as shown in Fig. 4a. The Cl 1s peak of the coating was similarly curve-fitted into two components at 200.5 eV for the covalent chlorine in N-chloramine and at 197.2 eV for the anionic chlorine in pyridinium ion with an areal ratio of * 2:0.69 as shown in Fig. 4b (Sodhi et al. 1992). The two fittings agree well since both suggest a * 70% quaternization and a * 100% chlorination.
After confirmation of its existence, the coating layer on pyridinium/di-N-chloramine cotton fibers was observed using SEM for morphology analysis. Compared with the original ones (Fig. 5a) that have smooth surfaces, the pyridinium/di-N-chloramine cotton fibers (Fig. 5b) were tightly surrounded with continuous coating layers without cracks or agglomerates. The full coverage ensures good contact with bacterial suspension and hence is positive for the biocidal application.
The thickness of a modification layer on a substrate is commonly estimated from the image of crosssection. However, the fibers herein are coated with a biocidal silicone. Silicones are highly elastic and have very low glass-transition temperatures, which in turn hinders the acquisition of an accurate cross-section of silicone-coated fibers even with freeze-cutting technique (Przybylak et al. 2018). The thickness of the modifier was then estimated to be 136.8 nm by using a method based on the weight of the coating from the following equation: where t denotes the thickness of the biocidal layer; d is the diameter of the original cotton fiber; W 0 and W 1 are weights of the fibers before and after modification, respectively; q C and q P mean densities of the cotton cellulose and the biocidal silicone modifier, respectively. In addition, the concentration of oxidative chlorine is calculated to be 0.24 wt% by the previously described iodometric/thiosulfate titration from Eq. 1.

Assessment of biocidal performance
The biocidability of pyridinium/di-N-chloramine cotton swatches was tested using S. aureus and E. coli as model species. As what the biocidal data of kinetics showed in Table 1, pristine cotton swatches (controls) were not biocidal since the small losses of S. aureus (0.20 log) and E. coli (0.17 log) after 10 min contact time were a result from the adhesion of microorganisms to fibers and natural mortality (Ren et al. 2008).
In contrast, the pyridinium/di-N-chloramine cotton swatches were highly biocidal and imparted complete inactivation of both strains within 3 min. Since our hypothesis is that the combination of one cationic center and more than one N-chloramines leads to intensified synergism, a higher biocidal efficacy than not only monofunctionality but also basic synergistic integration of one cationic center and one Nchloramine, the comparisons of our combination with those cases are needed for verification. The biocidal performance our combination of one cationic pyridinium and two amide N-chloramines was firstly compared with non-synergistic single functionality (pyridinium or N-chloramines). Pyridinium functionalized cotton swatches that were formed by quenching the N-chloramine sites of pyridinium/di-N-chloramine cotton swatches with Na 2 S 2 O 3 solution showed much lower biocidal efficacies (only 1.50 log reduction for S. aureus and 0.81 log reduction for E. coli after 10 min) as shown in Table 1 since pyridinium salts are well-known mild biocidal groups. Furthermore, the pyridinium functionalized cotton swatches showed higher inactivation rate against S. aureus than E. coli due to the thicker membrane of E. coli that in turn leads to more resistance to the penetration of pyridinium structures (Hu et al. 2014;Liang et al. 2006). This low efficacy is still desirable, especially after the N-chloramines of the combined units are consumed and not rechlorinated yet. N-chloramine silane functionalized cotton with similar loadings of Cl ? was also less biocidal than pyridinium/di-N-chloramine cotton since an amide N-chloramine silane coated cotton with Cl ? loading of 0.23 wt% only fully eliminated S. aureus and E. coli within a contact time of 10 and 30 min, respectively (Cheng et al. 2015). Therefore, our presented combination is more efficient than single functionality of both pyridinium and N-chloramine. This is believed that the hydrophilic cationic salt not only overcomes the hydrophobicity of N-chloramine but also electronically draws bacteria to N-chloramine, assisting the contact and killing process.  S. aureus and E. coli at inoculum populations of 3.57 9 10 7 and 4.17 9 10 7 CFU, respectively. The accuracy for the log reductions in this study was ± 0.06 Our combination is then compared with basic synergism of one pyridinium and one N-chloramine. It took a longer contact time of 10 min for cotton swatches of basic synergism (one pyridinium and one N-chloramine) with higher loading of Cl ? (0.32 wt%) to completely kill S. aureus and E. coli , proving the intensified synergism of our design. This is because antibacterial ability is a surface property and hence only the ones on the top surface instead of the overall biocidal groups can participate in contact-killing process. The increase of the ratio of more powerful N-chloramine in a composite unit naturally results in a higher surface concentration of N-chloramine compared with basic synergistic counterparts, achieving the observed intensified synergism. The intensified synergism can then be used as a strategy for design of more powerful biocidal functionalities.

Stability of biocidal performance
Stable and regenerable biocidal performance is very desirable for practical applications. The chemistry of the biocidal coating layer is a silicone since its main chain is composed of -Si-O-units. Studies have verified that silicones are ideal carriers of functional pendants since they are nontoxic polymers with a sturdy and hydrophobic backbone that ensures safe and long-lasting biocidal performance even under intense usages Zhao et al. 2015). Additionally, silicones can be photodecomposed into inorganic SiO x that shields beneath compositions from further photolysis (Ouyang et al. 2000;Phely-Bobin et al. 2000), increasing the stability of the biocidal coating layer. The stability and rechargeability of oxidative chlorines on pyridinium/di-N-chloramine cotton swatches under repeated washings are summarized in Table 2. The content of oxidative chlorines on pyridinium/di-N-chloramine cotton decreased gradually as the increase of the number of washing cycles, losing a 83% of initial loading (from 0.24 to 0.04 wt%) after 15 washing cycles (equivalence of 75 machine washing cycles). The samples after 15 washing cycles were still biocidal, providing total kills of * 10 7 CFU of E. coli. and S. aureus in 60 and 45 min, respectively. The rechargeability of lost chlorines is promising since the content reached 0.10 wt%, corresponding to a 42% rechargeability, after 15 washing cycles by rechlorination with NaClO. It is believed that the recoverable parts of lost chlorines are caused by hydrolysis of N-Cl to N-H and the nonrecoverable ones are caused by peeling of biocidal silicone coating during washings.
Similarly, UV photolysis can also induce permanent loss (decomposition of the biocidal layer) and temporary loss (cleavage of N-Cl) of N-chloramine sites. Data in Table 3 shows that the loading of oxidative chlorines radically decreased within 1 h, losing 49% of oxidative chlorine. Then, the loss gradually increased with the increase of irradiation time, losing 93% and 100% of initial value at 24 h and 7 d. However, 96%, 61% and 52% of oxidative chlorines were recharged after the swatches were shined for 1 h, 24 h and 7 d and rechlorinated. The durability and rechargeability of oxidative chlorines under UV irradiation are both better than some nonsilicone coatings Jiang et al. 2014) due to the previously mentioned shielding effect of SiO x . In those literatures, Cerkez and coworkers reported 54%, 68%, and 100% losses of chlorines after irradiation for 1 h, 2 h, and 24 h, respectively, while Jiang and coworkers observed a 100% loss of chlorines after irradiation for only 3 h. In addition, UV absorbing effect of pyridinium could also contribute to the photolytic stability of the coating (Li et al. 2016). The biocidal modifier adsorbs moisture from air under storage, which in turn leads to the hydrolysis of N-chloramine structures to N-H moieties. The remained and recovered chlorines as a function of long-term storage are shown in Table 4. The chlorine loading gradually decreased from 0.24 wt% to 0.11 wt% (a loss of 54%) within 30 days under laboratory light at 25°C and 65% RH. The minimum contact times of the samples stored for 30 days for total kills of *10 7 CFU of E. coli. and S. aureus were 25 min. *83% of chlorines was recovered by rechlorination for the samples stored for 30 d. The good durability under long-term storage is attributed to the hydrophobicity of silicone main chains to prevent moisturetriggered hydrolysis of N-chloramine structures. In contrast, * 85% Cl ? loss was measured in a hydrophilic chitosan-based coating that absorbed a large amount of moisture from the surroundings that caused rapid hydrolysis of N-chloramine sites under the same conditions for 30 days (Cao and Sun 2008). Therefore, the biocidal performance of the modified cotton is anticipated to be well preserved after a relatively extended storage period.
The above results verified that the increase of the ratio of N-chloramine in a composite unit of cationic center/ N-chloramine is a strategy for deign of biocides with intensified synergism for antibacterial modification of substrates.

Conclusions
A silane, 6-(pyridin-4-yl)-3-(3-(trimethoxysilyl)propyl)-1,3,5-triazinane-2,4-dione, was synthesized for condensation polymerization on cotton to form a silicone coating. After quaternization and chlorination, each silicone unit is designed to have one cationic pyridinium and two amide N-chloramines. The results proved the hypothesis that the increase of N-chloramine in a composite unit leads to intensified synergism, an even higher biocidability than basic synergism of one cationic center and one N-chloramine. In addition, the design employs cyclic and hydrophilic N-chloramine, UV adsorbing pyridinium, and sturdy and UV-resistant silicone backbone. These virtues ensure promising durability and recoverability of antibacterial functionality under repeated washings, UV irradiation and long-term storage. The pyridinium/ di-N-chloramine cotton therefore has superior biocidability and stability for practical usage. The accuracy for the chlorine evaluations was ± 2% Errors are estimated to be ± 0.01