A super-ecient and biosafe hemostatic cotton gauze with controlled balance of hydrophilicity/hydrophobicity and tissue adhesiveness

18 Cotton gauze is a widely used topical hemostatic material for bleeding control in 19 military and civil accidents and surgical operations, but its high blood absorption 20 capacity tends to cause extra blood loss of the wounded person, which may increase the 21 risk of shock and death. Therefore, development of rapid hemostatic cotton gauze with 22 less blood loss is of great significance. Herein , we prepared a super-efficient hemostatic cotton gauze whose surface was slightly modified with a catechol compound which 24 features a flexible long hydrophobic alkyl chain terminated with a catechol group. Its 25 hemostatic performance in rat and pig injury models was far superior to standard cotton 26 gauze and Combat Gauze TM . The latter is a well-known commercial gauze for controlling massive hemorrhaging. Additionally, after stoppage of bleeding, the wound 1 sites hardly re-bleed upon the gauze was peeled off. Histological analysis proved that 2 the novel cotton gauze well kept the biosafety of cotton gauze. Interestingly, a similar 3 impressive hemostatic performance was also achieved for chitosan nonwoven gauze 4 modified with the same procedure. Density functional theory calculation and 5 instrumental measurements demonstrate that their extraordinary hemostatic capability 6 is attributable to the highly efficient formation of big and thick primary blood clot made 7 of massive aggregated erythrocytes, due to gauze’s effective controlling of blood 8 movement through its blocking effect from tissue adhesion by catechol, platelet 9 activation by cotton fiber, blood absorption by cotton, and hydrophobic effect from long 10 alkyl chain. The methodology and hemostatic mechanisms presented in this work may 11 open a new avenue for developing highly efficient hemostatic gauzes. the pores and on the yarns of the 3 rd layer. Cross-section of the 1 st layer shows a thick erythrocyte layer. (c) Through adhesive bonds li ke π - π stacking interaction and 2 hydrogen-bond between USO’s catechol group and wound tissue’s amino acid units, 3 dam-like barriers form surrounding the wound. They retard blood seeping out of the 4 tissue surface. The repelling pressure from the hydrophobic effect among long alkyl 5 chains slows down blood wicking movement. (d) Dam-like barriers forming by USO 6 between gauze layers and between fibers, retard blood diffusion in the vertical and 7 horizontal directions, largely confine blood movement in the pores between warp and 8 weft yarns, resulting in large accumulation of erythrocytes.

less blood loss is of great significance. Herein, we prepared a super-efficient hemostatic 23 cotton gauze whose surface was slightly modified with a catechol compound which 24 features a flexible long hydrophobic alkyl chain terminated with a catechol group. Its 25 hemostatic performance in rat and pig injury models was far superior to standard cotton 26 gauze and Combat Gauze TM . The latter is a well-known commercial gauze for 27 controlling massive hemorrhaging. Additionally, after stoppage of bleeding, the wound 1 sites hardly re-bleed upon the gauze was peeled off. Histological analysis proved that 2 the novel cotton gauze well kept the biosafety of cotton gauze. Interestingly, a similar 3 impressive hemostatic performance was also achieved for chitosan nonwoven gauze 4 modified with the same procedure. Density functional theory calculation and 5 instrumental measurements demonstrate that their extraordinary hemostatic capability 6 is attributable to the highly efficient formation of big and thick primary blood clot made 7 of massive aggregated erythrocytes, due to gauze's effective controlling of blood 8 movement through its blocking effect from tissue adhesion by catechol, platelet 9 activation by cotton fiber, blood absorption by cotton, and hydrophobic effect from long 10 alkyl chain. The methodology and hemostatic mechanisms presented in this work may 11 open a new avenue for developing highly efficient hemostatic gauzes. 14 Massive bleeding from nonfatal traumatic wounds is one of the leading causes of 15 death and disability in battlefields and civilian accidents, because significant blood loss 16 causes symptoms such as hypothermia, coagulopathy, acidosis, sepsis, and organ 17 failure. 1,2 More than 50% of these mortalities is preventable if emergent and efficient 18 hemostatic measures are applied. This inspires the development of advanced hemostatic 19 products and technologies for prehospital bleeding control of the wounded people, to 20 increase survival rate and reduce medical costs. 3-6 21 Cotton gauze has a long history as an effective topical hemostatic fabric for 22 compressible and noncompressible wounds, mainly due to its safety, non-allergy, low 23 cost, adaptability, breathability, stability, blood absorbency, and easy applicability 7,8 . 24 It is still the most widely used hemostat for traumatic bleeding control, although lots of 25 efficient hemostatic agents have been manufactured and clinically applied in recent two 1 decades. [9][10][11][12] The hemostatic mechanism of cotton gauze counts on the activation of 2 platelets upon contacting with cotton fiber, and its quick wicking of blood fluid, leading 3 to resting of blood cells and platelets to form blood clots. However, it's often seen that 4 excessive large volume of blood is absorbed by cotton gauze before bleeding stops, due 5 to its highly hydrophilic nature, porous structure, and capillary action among the 6 weaved fibers. Those extra blood losses may be the last straw that causes morbidity or 7 mortality because the blood volume in the circulating system is critical. 8 Many endeavors to enhance the hemostatic efficacy (reduction in blood loss and 9 bleeding time) of cotton gauze have been made in the academic and industrial circles. 10 Z-Medica in USA has commercialized a gauze brand-named QuikClot Combat Gauze ® 11 (QCG), which is made by binding inorganic mineral kaolin particles onto 12 rayon/polyester nonwoven. Kaolin can activate clotting factor XII to accelerate blood 13 coagulation reactions, leading to fast thrombus formation. This topical hemostat has 14 been clinically adopted in military, emergency care, and hospital for compressible 15 severe hemorrhage. However, its rapid hemostasis efficacy is subsided because of 16 possible loss of kaolin, and the detached kaolin particles may cause risks of unexpected 17 distal thrombus. To increase binding stability of inorganic particles to cotton fiber, 18 mesoporous chabazite zeolite particles were chemically anchored onto fiber surface by 19 an on-site growth route. 8 Such a composite cotton gauze has a better topical hemostatic 20 efficiency than QCG. In the rabbit lethal femoral artery injury model, blood loss of 21 chabazite zeolite-cotton gauze was only about 40% less than that of QCG. For these 22 two kinds of hybrid cotton gauzes, it's obvious that they still absorb large precious 23 volume of blood during bleeding control. Tuning the wettability of a hemostatic fabric 1 was proposed to address this concern. 13 A Janus gauze consisting of a top hydrophobic 2 fabric layer and a bottom hydrophilic cotton fabric layer was developed. It was thought 3 that the hydrophilic layer absorbed blood to expedite clotting, while the hydrophobic 4 layer gives a pressure to inhibit blood diffusion through gauze in lengthways. However, 5 blood permeation in warp and weft directions of the bottom cotton fabric layer is un- 6 avoidable, resulting in losses of valuable blood yet. Furthermore, in most cases, blood 7 seepage from the seam between the hemostatic fabric/wound surface, which leads to 8 massive blood loss as well. A composite gauze with a superhydrophobic 9 Poly(vinylidene fluoride)/carbon nanofiber (PVDF/CNF) bottom coating layer was 10 reported to have fast hemostatic capability and no re-bleeding potential, because of 11 synergetic effects from CNF's acceleration of fibrin fiber formation and PVDF's 12 repellency of blood. 14

13
Therefore, controlling of movement of blood fluid at the gauze/tissue contact 14 surface and in gauze is the key in designing a highly efficient hemostatic gauze. Inspired 15 by mussel foot protein's good adhesion to wet substrates, [15][16][17] we speculate that cotton 16 gauze with catechol group on its surface may help it adhere to blood-wetted tissue, in 17 order to hinder blood seepage from the seam of gauze/tissue contact surface; meanwhile, 18 the strong interaction among cotton fibers through catechol linkage may slow down 19 blood diffusion in gauze. Thus, in view of the great importance of fiber structure, 20 wettability, and wet biological tissue adherence to hemostatic materials, herein, we 21 design and prepare a novel hemostatic cotton fabric, which is made by slightly grafting 22 a catechol compound (Fig. 1a) with a C15 alkyl side chain onto the fabric surface. This 23 hemostatic fabric integrates wet tissue adhesiveness from catechol group, 1 hydrophobicity from long alkyl chain, absorbency and breathability from cotton fiber, 2 into one hemostatic device. In rat femoral artery and liver laceration, and pig femoral 3 artery injury models, this modified cotton gauze shows very limited blood permeation 4 through the gauze, no blood oozing out from the seam of gauze/wound contact surface, 5 short bleeding time and reduced blood loss. Its biosafety is comparable to cotton gauze, 6 as suggested by the cytocompatibility and histological examinations. This idea and 7 methodology were also successfully applied to chitosan nonwoven fabric to fabricate a 8 gauze with high hemostatic capability. 9 10 Fig. 1 (a) Reaction scheme for grafting catechols onto cotton gauze. Surface grafting 11 reactions of (b) 4-allyl-1,2-benzenediol (ABO) and 1,2-benzenediol-3-(9,11,13-12 pentadecatrienyl) (USO), and (c) hexadecyltrimethoxysilane (HTMS) onto cotton 13 cellulose macromolecular chain. 14

1
Chemical composition and surface structure 2 Free radicals can be readily generated on cotton gauze by plasma treatment. 3 Through radical initiated reaction with the double bonds in USO, which can be grafted 4 onto cotton gauze surface. Thus, a long alky side chain with a catechol end group 5 was introduced onto cotton fiber (Fig. 1a, b). Its chemical structure is confirmed by its 6 solid state 13 C NMR spectrum (Fig. 2a). The peaks at region of 60-70 ppm are attributed 7 to C6 of cellulose, while the signals at region of 70-80 ppm is assigned to C2, C3, and 8 C5 of cellulose. The peaks at region of 80-95 and 100-110 ppm correspond to C4 and 9 C1, respectively. 18 Compared with cotton gauze, several new peaks appear for USO-g- 10 gauze, i.e., broad peaks at 120-150 ppm corresponding to unsaturated carbons on 11 benzene ring; peaks at region of 10-40 ppm assigning to methyl and methylene groups 12 of the alkyl chain of USO. Therefore, 13 C-NMR spectra indicates that USO is 13 successfully grafted onto cotton gauze surface. 14 15 Fig. 2 (a) Solid-state 13 C NMR spectra of gauzes. SEM images of (b-d) cotton gauze 16 and (e-g) USO-g-gauze at different magnifications. 17 FTIR and XPS also prove that ABO, HTMS and USO were successfully grafted 18 onto the surface of cotton gauze (Figs. S2 and S3). For cotton gauze, characteristic 19 absorption bands at 3200-3600 cm -1 (υ O-H ), 1610 cm -1 (υ C-C ), 1435 cm -1 (δ C-O-C ), 1125 cm -1 1 (υ C-O-C ) are present. In the spectra of ABO-g-gauze, HTMS-g-gauze, and USO-g-2 gauze, a new band at 2846 cm -1 appears for υ C-H of -CH2-groups of ABO, HTMS, 3 and USO grafted on the gauze. C1s XPS spectra shows the relatively strongest peak of 4 USO-g-gauze is C-C bond, while that of cotton gauze is C-O-/C-OH bond (Fig. S3b). 5 This is in accordance with the abundance of C-C bond from the long aliphatic chain 6 grafted on USO-g-gauze. 7 Surface morphology of gauzes 8 Cotton gauze consists of interwoven cotton yarns ( Fig. 2b-g). Pores including 9 macropores (among yarns), capillaries (among fibers), meso-and micro-pores (inner 10 fiber) are abundant in the gauze (Fig. 2b). After plasma treatment and radical initiated 11 graft reaction with USO, the interwoven fiber network structure is well maintained, 12 while the surface roughness of cotton fiber slightly increases (Fig. 2g). On the one hand, 13 the improved surface roughness is due to the deepening effect of plasma etching. 19 On 14 the other hand, the USO thin layer grafted on the fiber surface may also enhance 15 roughness. 20 In fact, the well maintenance of fiber/yarn/fabric morphology is also found 16 for ABO-g-gauze and HTMS-g-gauze (Fig. S4). 17 18 Wettability of gauzes 19 The wettability of a gauze has a large effect on blood fluid absorption, protein 20 adsorption and blood cell adhesion. For instance, it is reported that a material with a 21 water contact angle (WCA) of 40-70° is suitable for adhesion of various cells. 21 The 22 wettability of gauze by simulated body fluid (SBF) and fresh rat blood was evaluated 23 by applying 200 μL of these liquids onto a gauze. As shown in Fig. 3a, a water droplet 24 immediately spreads over and diffuses into gauze upon dripping onto gauze surface, 1 due to cotton gauze has a robust intrinsic hydrophilicity and capillary structure. ABO-2 g-gauze exhibits similar wetting behavior, but the blood spreading area on ABO-g- 3 gauze surface is smaller than that on cotton gauze within a same time period (Fig. 3b). 4 HTMS-g-gauze shows a high hydrophobicity with a contact angle of 132.6°. Water or 5 blood droplet neither spreads radially nor diffuses downwards, but stands on HTMS-g- 6 gauze surface. This is due to this gauze is covered by 16-carbon hydrophobic alkyl 7 chains, whose hydrophobicity prevents blood from diffusing into interior. The instant 8 static WCA of USO-g-gauze is ca. 68.2°. Water or blood exhibits a unique and 9 interesting wetting behavior on USO-g-gauze, namely the droplet gradually and 10 vertically diffuses into gauze in 60 s, while the wetted area on surface is almost identical 11 to the size of the droplet. This is totally different from the easy all-directional liquid 12 pervasion on cotton gauze and ABO-g-gauze.  15 and (d) water absorption ratio of gauzes. 16 Water vapor permeation rate and water absorption/movement in gauze 1 Hemostatic gauze with proper water vapor transmission and liquid absorption can 2 prevent dehydration and excessive accumulation of exudates. Therefore, they can 3 control water/blood loss and create an ideal moist environment for wound healing. 22 4 The water vapor permeation rate of cotton gauze, ABO-g-gauze, HTMS-g-gauze, and 5 USO-g-gauze was 1028, 1023, 1021, and 1015 g·m -2 ·day -1 at 37 o C, respectively ( Fig.   6 3c), suggesting that water vapor permeability of the three surface modified gauzes is as 7 good as that of standard cotton gauze, due to their breathable knitted fabric structure. 8 Cotton gauze and ABO-g-gauze are rapidly wetted by water and sink to the bottom 9 when they are put in water. The hydrophobic HTMS-g-gauze always floats on water 10 surface. However, USO-g-gauze shows a very different behavior, i.e., it initially floats 11 on water surface, but sinks after 30 min (Suppl. Video 1). USO-g-gauze has a thin USO 12 layer on its surface, weakening its hygroscopicity. This lowers the water absorption rate 13 of USO-g-gauze, which is proven by its water absorption dynamic (Fig. 3d). Therefore, 14 it takes a relatively long period for fully swollen to sink. The water absorption ratio of 15 cotton gauze, ABO-g-gauze, HTMS-g-gauze, and USO-g-gauze is approximately 16 337.7%, 313.9%, 7.6%, and 260.5% at the saturated state ( Fig. 3d), respectively. 17 Compared to cotton gauze, the water absorption ratio of USO-g-gauze significantly lose 18 by 77.2%, attributing to the great effect from the very thin hydrophobic USO layer on 19 surface. 20 The movement of water in gauze was also quantitatively measured by moisture 21 management test (MMT). The variation of wetting time of gauzes (Table 1) is in good 22 correlation to their wettability as shown in Fig. 3a and b. The wetting time (top side) is 23 as short as 0.19 s for cotton and ABO-g-cotton gauzes, and substantially increases to 1 1.59 s for USO-g-cotton, and to 15.53 s for hydrophobic HTMS-g-gauze. As indicated 2 in Table 1, the shorter wetting time is, the faster the water absorption rate and spreading 3 speed is. The water absorption rate of gauzes follows a decreasing order of Cotton 4 gauze > ABO-g-gauze > USO-g-gauze > HTMS-g-gauze, while the water spreading 5 speed is in the order of Cotton gauze ≈ ABO-g-gauze >> USO-g-gauze >> HTMS-g- 6 gauze. The changing patterns of these indices suggest that the surface chemical structure 7 of gauze effectively guides moisture movement (wetting, spreading, and diffusion) in 8 gauze, agreeing well with the results shown in Fig. 3. The designed concomitant 9 hydrophobic/hydrophilic structure of USO-g-cotton imparts it with not only proper 10 wetting time and spreading rate, but its ability of water diffusion from one side to the 11 other side (indicated by the cumulative one-way transport capacity, Table 1). Such a 12 unique property would be very helpful for controlling blood movement in gauze and at 13 the gauze/tissue contact surface when it is practically applied as a topical hemostat, as 14 will be shown in the following sections. 15  Erythrocyte and platelet adhesion on gauze 1 Aggregated erythrocytes and platelets are essential components of blood clot for 2 bleeding control. 23 SEM was used to explore the interaction of erythrocyte/platelet with 3 gauzes. As shown in Fig. 4, erythrocytes adhere on gauze fiber surface to form 4 aggregates. The unique double concave disk structure of erythrocyte is well maintained, 5 indicating that gauzes don't affect the normal physiological state of erythrocytes. 6 However, the amounts of erythrocyte adhering on gauze fiber surface depend on its 7 wettability. Due to the hydrophobicity of HTMS-g-gauze, on it significantly fewer 12  Bleeding control efficiency in rat femoral artery injury model 5 The hemostasis efficacy of gauzes on in vitro trauma was evaluated by using the 6 rat femoral artery injury model. The hemostatic performance of the five gauzes is helpful for hemostat. When applied to a bleeding trauma, the hydrophobic HTMS-g- 1 gauze inhibits blood wetting, absorption, wicking and diffusion into the upper gauze 2 layer ( Fig. 3 and Table 1), resulting in poor hemostatic performance. As for QCG gauze, 3 an acclaimed gauze for effective controlling of severe bleeding, the outflowing blood 4 diffuses into the four-layer gauze in 1 s, but the blood-stained area on the outmost layer 5 gauze is smaller than that on cotton gauze and ABO-g-gauze. The unfolded gauze, 6 which is removed from the post-hemostatic wound, has blood-stains on all four gauze On the rat femoral artery injury model, HTMS-g-gauze has the longest hemostasis 1 time of ca. 356 ± 15 s; The hemostasis time of ABO-g-gauze is ca. 174 ± 13 s, down 2 from ca. 180 ± 20 s of cotton gauze; that of USO-g-gauze astonishingly fells to ca. 34 3 s from 147 s of QCG gauze (Fig. 5b). Compared with QCG gauze, the hemostatic time 4 of USO-g-gauze drops by 77%. Accordingly, the blood loss follows a decreasing order 5 of HTMS-g-gauze (1.04 ± 0.09 g) > cotton gauze (0.52 ± 0.03 g) ≈ ABO-g-gauze (0.49 6 ± 0.08 g) > QCG gauze (0.42 ± 0.03 g) > > USO-g-gauze (0.12 ± 0.03 g) (Fig. 5c). 7 Thus, the blood loss of USO-g-gauze is impressively reduced by about 71%, as 8 compared to QCG gauze. 9 10 Fig. 5 (a) From left to right: gauze was put on the bleeding rat femoral artery injury, 11 gauze was removed from the wound after hemostatic state was reached; The stacked 12 four gauze layers were unfolded. "1" was the layer directly contacted with the wound, 1 "4" was the outmost layer. (b) Hemostatic time and (c) total blood loss in the rat femoral 2 artery model. (d) Enlarged photo of rat femoral artery injury after hemostasis. This 3 photo was taken right after the gauze was removed. 4 5 Hemostasis in rat liver injury model 6 The gauze's hemostatic performance in a non-compressible wound is evaluated by 7 using the rat liver laceration model. Their behaviors of blood diffusion, flow at the 8 gauze/tissue contact surface, and re-bleeding are similar to the hemostasis on the rat 9 femoral artery injury model ( Fig. 6a and Fig. 6-unfolded gauze, Suppl. Video 5). 10 Obviously, the hemostatic efficacy of USO-g-gauze is significantly better than the other 11 four gauzes. When USO-g-gauze is removed from the wound in post-hemostasis, no 12 fresh blood wells out from the liver wound. The hemostatic time of cotton gauze, ABO-13 g-gauze, HTMS-g-gauze, QCG gauze, and USO-g-gauze is 172 ± 20 s,153 ± 15 s, 344 14 ± 19 s, 96 ± 12 s, and 32 ± 4 s, respectively (Fig. 6b). In addition, the blood loss of 15 cotton gauze, ABO-g-gauze, HTMS-g-gauze, QCG, and USO-g-gauze is 0.39 ± 0.08 g, 16 0.37± 0.03 g, 0.98 ± 0.09 g, 0.13 ± 0.07 g, and 0.03 ± 0.01 g, respectively (Fig. 6c). 17 Compared with QCG gauze, the hemostatic time and blood loss of USO-g-gauze is 18 reduced by 67% and 77%, respectively. Therefore, the high hemostatic ability of USO- 19 g-gauze on the non-compressible liver injury is also confirmed. 20 In both rat femoral artery injury and liver laceration models, the survival rate 21 within 120 min varies from one gauze to another. It is 100% for rats treated with USO- 22 g-gauze, compared to 20% for QCG, and to no survival for cotton, ABO-g-gauze and Hemostatic performance in pig femoral artery injury model 10 In order to further evaluate the hemostatic performance of gauzes on massive 11 bleeding wounds, the pig femoral artery injury model is used. The cotton gauze is 12 wrapped around the wound for 3 min, then opened up to check the wound. As shown 13 in Fig. 7a and Suppl. Video 6, the outmost cotton gauze layer is immediately wetted by 1 blood upon contact with the trauma. The blood diffusion area steadily increases as time 2 goes by. Fresh blood continues flowing out from the wound when the cotton gauze is 3 removed after 3 min (Fig. 7c). In contrast, USO-g-gauze shows far better hemostatic 4 performance (Suppl. Video 7). Only a small blood-stain is observed on the outmost 5 gauze layer in 60 s (Fig. 7b). Three min later, re-bleeding does not happen upon the 6 gauze is removed from the wound (Fig. 7d). Because this is a severe bleeding wound, 7 a relatively big area is blood-wetted in the first layer, but the 4 th layer is stained by a 8 small blood domain only. This suggests blood diffusion in the vertical and radial 9 directions is largely restricted in USO-g-gauze. It should be pointed out that the accurate 10 hemostatic time of every gauze is not measured on this injury model, because it is hard 11 to judge when the wound stops bleeding due to the wound is wrapped by gauze and re- 12 bleeding often occurs upon uncovering wound. Therefore, at the time interval of 3 min, 13 the blood mass absorbed by gauze is measured to roughly reflect the blood loss. This is 14 0.80 ± 0.12 g for USO-g-gauze, while that of cotton gauze, ABO-g-gauze, QCG, and 15 HTMS-g-gauze is 5.12 ± 0.34 g, 4.16 ± 0.40 g, 3.93 ± 0.30 g, and 8.20 ± 0.34 g, 16 respectively (Fig. 7e). Therefore, the blood loss from the pig femoral artery wound 17 treated with USO-g-gauze is only 15.6% and 20.4% of that with cotton gauze and QCG, 18 respectively. In fact, the hemostasis time of USO-g-gauze on this wound model is less 19 than 3 min, as suggested by the fact that no occurrence of re-bleeding upon uncovering 20 the wound (Fig. 7d). 21 Hemostasis on the pig skin laceration model is also examined since its structure is 22 very similar to human skin. A regular cut with a length of 2 cm and a depth of 1 cm was 23 made with a scalpel, then a four-layer gauze was applied onto the wound. The dynamic 24 hemostatic process of the four gauzes demonstrates that blood diffusion and absorption, 25 and blood flowing underneath gauze, are very similar to the hemostasis on the rat 1 femoral artery injury and liver laceration models. Blood even rarely stains the 2 hydrophobic HTMS-g-gauze, but oozes out from the seam of gauze/skin surface (Fig.   3 S7). The blood loss of the cuts treated with cotton gauze, ABO-g-gauze, QCG, HTMS-4 g-gauze, and USO-g-gauze is 0.55 ± 0.04 g, 0.32 ± 0.03 g, 0.22 ± 0.02 g, 0.71 ± 0.02 g, 5 and 0.032 ± 0.01 g (Fig. 7f), respectively. Compared with cotton gauze and QCG, the 6 blood loss for USO-g-gauze reduces by ca. 94% and 85.5%, respectively. The results 7 of the pig femoral artery injury and skin laceration models further justify that USO-g- 8 gauze has excellent hemostatic efficacy for severe bleeding wounds. 9 10 Fig. 7 Hemostasis in the pig femoral artery injury models. Hemostasis process of (a) 11 cotton gauze, and (b) USO-g-gauze. The status of wound after (c) cotton gauze and (d) 12 USO-g-gauze were removed after it was treated for 3 min. Blood loss of gauzes in the 13 pig (e) femoral artery and (f) skin injury models. 14 1 In each injury model, the bleeding wounds treated with ABO-g-gauze is cleaner, 2 and the blood diffusion area on gauze and blood loss are smaller than those treated with 3 standard cotton gauze. ABO-g-gauze not only concentrates blood components due to 4 its quick blood absorption ability (Fig. 3d), but also catches blood cells by its tissue 5 adhesive catechol groups (Fig. 4b), so the hemostatic efficiency of ABO-g-gauze is 6 slightly improved. The hemostatic performance of HTMS-g-gauze is significantly 7 inferior to that of cotton gauze, because of strong repellence of blood fluid by the highly 8 hydrophobic HTMS alkyl chain (Figs. 3 and 4). However, the USO-g-gauze containing 9 a hydrophobic long alkyl chain with a tissue adhesive catechol end group exhibits 10 impressively excellent hemostatic efficacy. 11 Why has USO-g-gauze the most exceptional hemostatic capability among those 12 gauzes? To better understand the nature behind this feature of this new hemostatic cotton 13 gauze, we initially carried out the detailed density functional theory (DFT) calculations 14 to investigate the adsorption interaction of sixteen different kinds of amino acid 15 molecules (they are essential components of tissue keratin protein 25 ) with USO-g-gauze, 16 as illustrated in Fig. S8. More details have been provided in Section 8 of SI. The 17 adsorption energies (ΔEads) of these amino acids to USO-g-gauze are calculated by 18 considering the main non-covalent interaction modes including π-π stacking and 19 hydrogen bond interactions. 20 Our computed results reveal that the amino acids containing π-conjugated benzene 21 ring, such as phenylalanine (F) and tyrosine (Y), can be effectively adsorbed on the 22 catechol of USO-g-gauze through synergistic actions of π-π stacking and hydrogen  (Table S2), and the calculated ΔEads values 3 are as big as 0.570~ 0.639 eV (Fig. S8), indicating strong interaction force between them. 4 Furthermore, we also examine the effect of relative position between two OH 5 groups on the benzene ring on ΔEads of a model amino acid glycine (G) on USO-g-gauze 6 ( Fig. S9). The computed results reveal that when the relative position between the two 7 OH groups is changed from the original ortho-to meta-to para-arrangements, ΔEads 8 values for G is reduced from 0.617 to 0.502/0.419 and then to 0.397/0.384 eV. Further, 9 when the two OH groups are even separated by three H atoms, small ΔEads (ca. 10 0.383/0.355eV) is attained. Clearly, with increasing the spacing distance between the 11 two OH groups, ΔEads for the model amino acid decreases significantly, in view of the 12 fact that double hydrogen bonds cannot be effectively formed or only a single hydrogen 13 bond can be formed (Fig. S9). Therefore, the relative position between two OH groups 14 on the benzene ring has an important influence on ΔEads, where the ortho position can 15 bring the largest ΔEads due to formation of two hydrogen bonds. Further, when removing 16 either of the two OH groups in catechol of USO-g-gauze (Fig. S10), the computed ΔEads 17 values (0.248 and 0.355 eV) is about half of that of the corresponding structure with two 18 hydrogen bonds, which means that both hydrogen bonds can be effectively formed 19 simultaneously between the relevant amino acids and catechol in USO-g-gauze. 20 Obviously, all of these can reflect the superior structural match to form double H-bonds 21 between the amino acids and the catechol of USO. 22 Overall, USO-g-gauze with a long alkyl chain terminated with a catechol group 23 can effectively interact with all these amino acids via double hydrogen bonds or the 24 synergistic action of π-π stacking and hydrogen bonding. These non-covalent 25 interactions contribute to USO-g-gauze's strong tissue adhesiveness, where the catechol 1 at the end of alkyl chain can play a crucial role.
2 Apart from the above molecular level analyses of adhesion interaction, the 3 adhesion force (or peeling force) of those gauzes on fresh wet rat femoral tissue is 4 measured and shows obvious variation from one to another. As expected, the 5 hydrophobic HTMS-g-gauze has the lowest peeling force of 24 mN, while USO-g- 6 gauze shows the largest force of 90 mN, which is ca. two times as much as that of cotton 7 gauze (Fig. S11). From Suppl. Video 3, the adhesion of ABO-g-gauze and USO-g- 8 chitosan to wound tissue is obviously perceivable when they were peeled off from the 9 wounds, while it is less noticeable in the cases of cotton gauze etc (Suppl. Video 2). 10 This vividly confirms the existence of adhesive interaction between catechol and tissue, 11 but such non-covalent adhesion can be broken with mild peeling forces. In the case of 12 the double -OH groups of catechol were modified such as with chelation with Fe 3+ or 13 oxidation into quinone, the tissue adhesion force sharply decreases to values close to 14 that of HTMS-g-gauze (Fig. S11). This further confirms that catechol group plays an 15 essential role in the wet tissue adhesion. 26-28 16 In order to demonstrate the importance of tissue adhesion of catechol to high 17 hemostatic efficiency, the catechol group on USO-g-gauze is transformed to lower its 18 tissue adhesiveness. Hence, USOFe-g-gauze and USOQu-g-gauze were fabricated. On 19 USOFe-g-gauze, catechol groups readily react with Fe 3+ to form a complex, while on 20 USOQu-g-gauze the catechol groups are oxidized into quinone (Fig. S12). Both 21 catechol-Fe 3+ and quinone groups have no or weak adhesion to wet tissue. 29

22
When USOFe-g-gauze is applied onto the rat femoral artery injury, blood spills (1s 23 in Fig. 15a, pointed by yellow arrow) and seeps out of the seam of gauze/wound surface 1 (15s in Fig. 15a, pointed by yellow arrow), it also diffuses throughout the whole four 2 gauze layers in 15 s, but thereafter the blood-stained area on the top layer does not 3 expand much, instead blood continuously oozes out from the gauze/tissue contact 4 surface (Fig. S13a). The gauze is removed 5 min later when the wound fully stops 5 bleeding, re-bleeding is observed and much red fresh blood is around the wound (Fig.   6 S13a). Compared to USO-g-gauze, the hemostasis time and blood loss on this wound 7 treated by USOFe-g-gauze significantly increases to 289 ± 5 s, and 1.32 ± 0.11 g, 8 respectively (Fig. S13b, c). The instant static WCA of USOFe-g-gauze is 119 o (higher 9 than 68 o of USO-g-gauze), but slowly reduces to 0 o within 60 s with diffusion of water 10 droplet (Fig. S12b), similar to that occurred for USO-g-gauze. Therefore, the weak 11 adhesiveness of catechol-Fe 3+ groups to wet skin tissue, would be responsible for the 12 substantially longer hemostatic time and more blood loss than USO-g-gauze and cotton 13 gauze. Similar phenomena (blood diffusion, seepage, and re-bleeding) are found for the 14 USOQu-g-gauze on the rat femoral artery injury (Fig. S13a). This is also due to the less 15 tissue adhesiveness of quinone groups existing on the USOQu-g-gauze surface. The 16 hemostasis time and blood loss by USOQu-g-gauze are 193 ± 5 s and 0.87 ± 0.21 g, 17 respectively (Fig. S13b, c). 18 The hemostatic performance of USOFe-g-gauze and USOQu-g-gauze on the rat 19 liver laceration model is similar to that on the rat femoral artery injury model (Fig.   20 S14a). Blood flows down the liver and wets the gauze (underneath the rat liver), which 21 does not occur when USO-g-gauze is applied. On this injury, the hemostatic time and 22 blood loss by USOFe-g-gauze are 101 ± 4 s and 0.48 ± 0.13 g, respectively; and by 1 USOQu-g-gauze are 65 ± 4 s and 0.28 ± 0.10 g, respectively (Fig. S14b, c). Thus, the 2 catechol group does play a crucial role in controlling traumatic bleeding. The far 3 superior hemostatic potential of USO-g-gauze to ABO-g-gauze (ABO has a three 4 carbons alkyl chain) strongly suggests that a long hydrophobic alkyl chain is also of 5 great importance. Therefore, USO, a catechol compound with a side alkyl chain 6 having 15 carbons is a good candidate compound for surface modification of fabric 7 gauze (such as cotton gauze and chitosan nonwoven) to prepare novel highly efficient 8 hemostatic gauzes. 9 Certainly, other properties of USO-g-gauze such as its moisture management 10 ability (water absorption, wettability, diffusion, and one-way transportation) are also 11 essential to the high hemostatic efficiency. The blood fluid movement in gauze and 12 around the gauze/tissue surface governed by the unique wetting property and tissue/cell 13 adhesiveness facilitates aggregation of astonishingly massive erythrocytes, as shown in 14 Fig. 8B. There are so many congested erythrocytes that they even fill the quadrilateral 15 macro-pores among warp and weft yarns of the first two gauze layers (the layer directly 16 contacts with tissue is the first layer), but none in the 3 rd layer, which are consistent 17 with the observation shown in Figs. 5 and 6-unfloded gauze. The thickness of the 18 erythrocyte layer accumulated on the 1 st layer reaches as high as 220 μm. As well- 19 known, erythrocytes are the key component of the primary blood clot. Thus, more 20 erythrocytes are aggregated, bigger clot is formed, shorter bleeding time and less blood 21 loss are attained. Therefore, the thick erythrocyte layers in the first two USO-g-gauze 22 layers serve as clots for effectively controlling bleeding. However, the erythrocyte 1 accumulation ability of cotton gauze is poor as suggested by its sparse distribution on 2 cotton gauze yarns with none in the quadrilateral macro-pores of the whole four gauze 3 layers (Fig. 8A). This is because erythrocytes move along with the fast blood wicking 4 to everywhere in the cotton gauze patch, rather than group together to form a big 5 erythrocyte plug. But such a movement is retarded in USO-g-gauze by the anchoring 6 hydrophobic chain barriers at the interface of USO-g-gauze/tissue and among fibers and 7 yarns. 8 Therefore, the hemostatic mechanism of USO-g-gauze is proposed in Fig. 8C and   9 d. When USO-g-gauze is applied onto a bleeding wound, catechol groups of USO 10 quickly anchor to skin tissue through non-covalent bonds such as hydrogen bond and 11 π-π stacking, to form dam-like barriers around wound (Fig. 8C). This can hinder and 12 eventually prohibit blood from seepage at the gauze/tissue contact surface. The 13 repelling pressure from the massive hydrophobic interaction among the long alky 14 chains retards blood diffusion into the upper gauze layers, so only the first and second 15 layers are blood-wetted as shown in the rat and liver injury models (Figs. 5 and 6-16 unfolded gauze). Even in the pig injury models, blood has a difficulty in diffusing 17 radially, and only little volume of blood reaches the outmost layer to result in a small 18 blood stain (Fig. 7b). This should contribute to the massive dam-like barriers formed 19 among yarns and fibers due to the interaction between USO catechol and cotton 20 cellulose, hence big blood stream moves in the pores between warp and weft yarns, 21 with small blood streams in other pathways (Fig. 8D). Finally, the blood wicking 22 capability of the moderately hydrophilic USO-g-gauze fibers facilitate platelets 1 adhesion and activation, and erythrocyte accumulation (Fig. 8D), therefore promoting 2 the formation of blood clot. Thus, the synergistic effects of tissue adhesion, 3 hydrophobic interaction, and hydrophilic fiber structure make USO-g-gauze an 4 excellent hemostatic gauze. The other impressive feature of USO-g-gauze is that no re-5 bleeding occurs upon removing it from wound after hemostasis, while re-bleeding is 6 often experienced when cotton gauze is used. In the case of cotton gauze, with sparse 7 aggregation of erythrocytes on yarns, it is an essential part of the blood plug (Fig. 8A). 8 The plug would be easily broken since erythrocytes are removed along with peeling- 9 off cotton gauze, leading to secondary bleeding. From Fig. 8B, it's seen that the 10 erythrocytes accumulated at the injury site is so enormous and thick that removal of 11 USO-g-gauze would take away part of erythrocytes, but some erythrocytes (Fig. 5d) 12 remain on site to avoid re-bleeding. The mechanism for no re-bleeding of USO-g-gauze 13 is different from that of Bandage ® , which is a well-known no re-bleeding hemostatic 14 fabric strip for small bleeding wounds. Bandage's anti-adhesion to tissue relies on the 15 hydrophobic membrane covering on the water-absorbent fabric layer (Fig. S15). The 16 poor adhesion of hydrophobic membrane to fibrin prevents bandage from being a part 17 of blood clot. The no re-bleeding is actually very important when wounded person are 18 relocated or injured tissue/organs are moved accidently, so new blood loss can be 19 substantially avoided. Since the hemostasis of USO-g-gauze is a physical blocking 20 effect rather than change of the body's normal physiologic clotting mechanisms, it 21 would also show hemostatic efficacy for patients with coagulopathy. Biocompatibility of USO-g-gauze 10 The growth and proliferation of L929 fibroblasts on USO-g-gauze are shown in 11 Fig. 9. On the gauze yarn, many green-dyed living fibroblasts with several red-dyed 12 dead cells are observed. Obviously, the fibroblasts proliferate very well with incubation 13 time increasing from 1 to 3 d. The good cell biocompatibility of USO-g-gauze is 14 comparable to that of cotton gauze and ABO-g-gauze ( Fig. S16a and b). The fibroblasts 15 can also dwell and grow on HTMS-g-gauze, but the proliferation rate on this 16 hydrophobic gauze is significantly smaller than that on the other three hydrophilic 17 gauzes (Fig. S16c). Therefore, USO-g-gauze can favor cell adhesion and proliferation, 18 due to good balance of hydrophilicity/hydrophobicity and presence of catechol groups 19 on this gauze.   Fig. 10 shows histological changes of the subcutaneous muscle tissue treated with 6 cotton gauze and USO-g-gauze at specific time points (3, 7, 14, 21 days) using H&E staining 7 and toluidine blue staining. In the cotton gauze-treated tissue, plenty of neutrophils arose 8 around the gauze on the third day after implantation (Fig.10a) and the corresponding number 9 were counted to be about 50 ± 5 (Fig. 10c). However, very few neutrophils were observed 10 7 days later, which reduced to about 2 ± 1 after 21 days. Meanwhile, USO-g-gauze had a 11 similar inflammatory response. The density of neutrophils increased in the USO-g-gauze 12 treated tissues since implantation, but rapidly decreased to 4 ± 1 after 7 days. In addition to 13 neutrophils, mast cell was another critical effector of inflammation. As shown in the 14 toluidine blue staining ( Fig. 10b and d), several mast cells were observed in the tissue 15 section contacting with the cotton gauze and the USO-g-gauze on the seventh days after 16 implantation, but sharply reduced 14 days later within the cotton and the USO-g-gauze 17 treated tissue. The subcutaneous implantation examinations reveal that the surface 18 modification with USO did not compromise the biocompatibility of cotton gauze, and 1 caused no significant inflammatory responses. blood absorption capacity of conventional cotton gauze, making it to significantly 5 reduce additional blood loss from bleeding traumas, and therefore increasing survival 6 rate and decreasing medical such as transfusional costs. Excellent hemostatic efficacy 7 (including short hemostatic time, low blood loss, and no re-bleeding) of this gauze was 8 observed on the rat femoral artery and liver laceration models, pig skin laceration and 9 femoral artery massive bleeding wound. Its hemostatic performance is much superior 10 to standard cotton gauze and QCG. The USO-g-gauze displays similar cell and tissue 11 compatibility to cotton gauze. Such an idea and methodology were also successfully 12 applied to make USO-g-chitosan gauze, whose hemostatic efficacy is much better than 13 the chitosan control. We speculate this magic gauze may find very promising 14 emergency-care and clinical applications for controlling traumatic massive bleeding for 15 wounded soldiers in battlefields, civilians in accidents, patients in operation rooms, and 16 patients with coagulopathy. 17 18

1.
Cotton gauze was sequentially washed with distilled water and ethanol, two times in 15 each solvent and 30 min each. Then it was dried under nitrogen stream. To introduce 16 free radicals onto gauze, it was treated by a low temperature plasma in N2 at 400 Pa, 80 17 W for 3 min (PT-5S，Sanhe Poda Co., Ltd, China). The preparation conditions for 18 USO-g-gauze were optimized as described in (Table S1, Section 1 of SI). In an ideal 19 condition, USO-g-gauze was fabricated by placing the plasma-treated gauze in a 20 mixture of 2.0 wt% USO/ethanol and refluxing at 70 o C for 2 h. Then, after washed 3 21 times with ethanol, it was dried in a vacuum oven at 80 o C for 2 h to obtain USO-g- 22 gauze. About 0.1 wt% USO was grafted, as determined by the gravimetrical method. 23 4-allyl-1,2-benzenediol (ABO) containing a catechol group was prepared 24 according to the scheme shown in Scheme S1. 32 Typically, eugenol (8.6 mL, 56 mmol) 25 and TES (19 mL, 112  analysis (Section 1 of SI). The ABO grafted gauze (ABO-g-gauze) was made by the 10 same way as that for USO-g-gauze, except that ABO was used to replace USO in the 11 reaction mixture (Fig. 1b). HTMS with a long alkyl chain in its chemical structure was 12 grafted onto the cotton gauze to make HTMS grafted gauze (HTMS-g-gauze) in a 13 similar way for preparing USO-g-gauze (Fig. 1c). 14 15 Characterization. 16 The morphology of gauzes was observed by scanning electron microscopy 17 (JEOL-7500LV, Japan). The solid-state 13 C NMR spectra were characterized by a 18 superconducting fourier transform nuclear magnetic resonance spectrometer (Bruker 19 Avance III 400 WB, AVANCE III, Switzerland). X-ray photoelectron spectra was 20 measured by scanning XPS microprobe instrument (Thermo K-Alpha+, UK). Fourier 21 transform infrared (FTIR) spectra in KBr form were obtained on an FTIR spectrometer 22 (PerkinElmer, 1600) in the range of 4000 -400 cm -1 . The wettability was determined 23 by DSA 100 (Krüss, German y). 24 25 Moisture management test 1 The tests (wetting time, wetted radius, water absorption rate, water spreading 2 speed, and cumulative one-way transport capacity) were performed on a moisture 3 management tester (M290, SDL ATLAS, USA) according to AATCC 195 by 4 measuring the electrical resistance of the top and bottom sides of the gauze. 33 The gauze 5 size was 5 cm x 5 cm. 0.2 mL of standard test solution was dropped onto the gauze in 6 the test. 5 replicates were run for each gauze. 7 8 Water vapor transmission rate and water absorption capacity of gauzes. 9 To measure water vapor permeability, a beaker containing 50 mL of distilled 10 water was covered with a gauze. The circumferential border was tightly sealed to 11 prevent any water vapor loss through the boundary. The water vapor transmission rate 12 (W evap ) was determined by measuring the mass loss of water in the beaker after 24 h at 13 37 o C. The W evap (g·m -2 ·day -1 ) was calculated according to Equation (1): 15 where A, m b , and m a are the area of the beaker mouth (m 2 ), the weight of the beaker 16 before and after water evaporation, respectively. 17 For measurement of water absorption ratio, a square gauze with size of 2 cm × 2 18 cm was immersed in simulated body fluid (SBF), then at certain time interval it was 19 taken out and placed on a filter paper to absorb free water, followed by weighing its 20 mass. The gauze mass before (m b ) and after water absorption (m a ) was measured, and 21 water absorption ratio (W abs ) was calculated by Equation (2) Hemostasis evaluation 1 All animal experiments were carried out in accordance with the guidelines for the 2 protection and use of experimental animals in Fujian Normal University. The experiments 3 were approved by the Animal Ethics Committee of Fujian Normal University. The gauze 4 was tailored into rectangle swatch with a size of 12 cm × 2.5 cm, and four swatches were 5 stacked together before used on a bleeding wound. The hemostatic study included four 6 animal injury models: rat femoral artery and liver injury, pig femoral artery and skin 7 laceration. Five rats or pigs were randomly selected as a group in each animal model and 8 assigned to each sample. QCG is the standard military hemostatic agent recommended by 9 the Tactical Combat Injury Care Committee for use as a control. Anesthesia was injected 10 intraperitoneally with a 10 wt% chloral hydrate solution (0.4 mL/100 g). After complete 11 hemostasis was reached, all animals were observed for 2 h or until death. The survived 12 animals were euthanized with 10% chloral hydrate at the end of experiment. 13 14 Rat femoral artery injury model 15 The proceeding of femoral artery injury model was conducted as follows. 34 Rats 16 were randomly selected and anesthetized. Then the fur on the rat thigh was shaved off to 17 expose the femoral artery. Pre-weighed cotton gauze was placed beneath the thigh. Then 18 the artery was transected. After bleeding for 2 s, pre-weighed gauze was gently applied 19 or compressed onto the trauma (compressing for 150 s). Blood diffusion in the gauze 20 was recorded. Bleeding time and blood loss were measured (n=6). We defined that 21 hemostatic stage is reached when the blood-stained area on gauze doesn't expand and no 22 blood seeps out of the seam at gauze/wound contact surface. 23 For observing the micro-morphology of the blood-stained gauze after hemostasis, a 24 patch stacked with four gauze layers was dressed onto the rat femoral artery injury to reach 25 hemostasis. The gauze was placed in centrifuge tubes and fixed with 2.5% (v/v%) GA/PBS 1 solution at room temperature for 4 h. Then it was rinsed twice with distilled water, followed 2 by sequential dehydration in 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% and 3 100% (v/v%) ethanol/PBS solution. consequently, it was dried at 37 o C. The gauze was 4 sputter-coated with gold before SEM observation. 5 6 Rat liver laceration model 7 Liver laceration model was conducted according to a standard procedure. 35 Rats 8 were randomly selected and anesthetized. Subsequently, the epithelial tissue of the 9 abdomen was cut to expose the liver, and pre-weighted cotton gauze were placed under 10 the liver. A scalpel was used to make a linear incision trauma of about 1 cm in length on 11 the left lobe of the liver. After bleeding for 2 s, pre-weighed gauze was applied onto the 12 trauma. Blood permeation in the gauze, bleeding time, and blood loss were recorded (n 13 = 6). 14 15 Pig skin laceration and femoral artery injury models 16 Two-month-old Bama miniature pigs (weight 1.6 -1.8 kg) were used to simulate 17 massive bleeding trauma. 36 Anesthesia was injected intraperitoneally with a 10 wt% 18 chloral hydrate solution (0.4 mL/100 g). After the hairs on the back and leg were shaved 19 off, the following two different traumas were created: (1) A linear incision of 2 cm 20 (length) x 1 cm (depth) was made on the back of the pig with a scalpel; (2) A distally 21 extending 4 cm longitudinal incision was made in the right femur region to expose the 22 femoral artery. 37

23
Peeling force of gauze on wet rat femoral tissue 1 Rats were pre-treated as in the femoral artery injury model, then the peeling test 2 was conducted before the artery was transected. Gauze (5 x 2.5 cm 2 ) was immediately 3 put on the exposed wet femoral tissue and kept there for 10 min. The gauze was then 4 peeled off with a Digital Push Pull Gauge (Locosc Ningbo Precision Technology Co., 5 Ltd., China), and the peak force was recorded (n = 5). 6 7 Effect of catechol functional groups on hemostasis 8 In order to elucidate the effect of catechol structure on hemostatic efficiency of 9 USO-g-gauze, two models were designed: (1) The catechol groups were protected by 10 chelating with Fe 3+ ions. 38 Typically, USO-g-gauze was soaked in 100 mL of 0.1 M FeCl 3 11 aq. solution for 10 minutes at 37 o C to allow the occurrence of coordination reaction 12 between catechol groups and Fe 3+ . After washed with distilled water, it was dried in a 13 vacuum oven at 80 o C for 0.5 h. As-obtained gauze was coded as USOFe-g-gauze.
(2) 14 The catechol groups were oxidized to quinone. The USO-g-gauze was treated in 100 mL 15 of 20 mM Tris-HCl buffer solution (pH 9.8) for 10 minutes at 37 o C to convert the 16 phenolic structure to quinoine structure. 39 After washed with distilled water, it was 17 dried in a vacuum oven at 80 o C for 0.5 h. As-obtained gauzed was coded as USOQu-g- 18 gauze. 19 20 Accumulation of erythrocytes and platelets on gauzes 21 The platelet-rich plasma (PRP) was obtained by centrifuging rat whole blood at 22 3000 rpm for 20 min at 4 o C. 40 In a volume ratio of 1:10, the whole blood or PRP was 23 added into a phosphate buffer solution (PBS, pH 7.4) containing a piece of gauze (1 cm 1 × 1 cm), followed by incubating for 90 min at 37 o C. Subsequently, the gauze was rinsed 2 three times with PBS to remove physically adhered blood cells and platelets, and fixed 3 with 2.5% (v/v%) GA/PBS for 2 h. Then it was rinsed twice with distilled water, 4 followed by sequentially dehydrating with 25%, 50%, 75%, 85%, 90%, and 100% (v/v%) 5 ethanol/PBS solution. Finally, it was air-dried at 37 o C. The adhesion of erythrocyte and 6 platelet on gauze surface was observed by SEM. 7 8 Biocompatibility of gauze 9 The proliferation of L929 cells on gauze was measured by using a live/dead assay 10 kit (Beyotime Biotechnology, Shanghai, China). 41 Briefly, after treated with 75% 11 medical alcohol and 0.1 mg/mL physiological saline, the gauze was transferred into a 12 24-well plastic culture plate. 1 mL of fibroblast with a density of 2.5×10 5 cells/mL was 13 gently added on the surface of the gauze surface and cultured at 37 o C in a humidified 14 atmosphere of 5% CO 2 for 1, 2, and 3 d. Then, 10 μL of the combined Live/Dead cell- 15 staining solution (2 µM Calcein-AM and 4 µM PI) was added into 400 μL of culture 16 medium and was incubated at 37 o C for 4 h. Finally, images of the live (green 17 fluorescence) and dead (red fluorescence) cells were obtained using an inverted 18 fluorescence microscope. 19 20 In vivo inflammatory assay 21 Sprague Dawley rats maintained under a 12 h light/12 h dark schedule with a 22 continuous supply of food and water. After a week of adaptation, the rats were randomly 23 divided into 2 groups: cotton gauze group and USO-g-gauze. Each group contained 20 1 rats. The gauze was cut into a square (1 cm × 1 cm) and were sterilized by ultraviolet 2 irradiation for 2 h. The rats were anesthetized by inhalant anesthetics-ether and fixed on 3 the surgical plate. After the shave and the iodine disinfection, a longitudinal incision 4 about 2 cm were made symmetrically on both sides of the spine. Then, the cotton gauze 5 or USO-g-gauze were respectively implanted into the subcutaneous sac of the back. The 6 incision was sutured by the simple intermittent suture method. Finally, the cut was 7 sprayed with penicillin powder and then covered with aseptic dressing paste to prevent 8 infection. The animals were returned to cages alone after surgery. At specific times (3, 9 7, 14, 21 days), the rats were sacrificed and the wounds along with surrounding tissues 10 were collected. The collected tissues were fixed by 4% paraformaldehyde. After 11 formaldehyde fixation for 1 week, the calcium was decalcified with 10% EDTA reagent 12 for 2-4 weeks. Subsequently, H&E staining and toluidine blue staining were performed, 13 and staining was performed using a fluorescence inverted microscope for microscopic 14 examination and image acquisition analysis. 15 16 Statistical analysis 17 All data are shown as the mean ± standard deviation (SD) with five replicates for 18 each test. One-way analysis of variance (ANOVA) was applied for statistical analysis. 19 *p < 0.05 and **p < 0.01 were considered significant and greatly significant difference, 20 respectively. 21