Carbon Nitride Quantum Dots Inhibit Ice Growth - Wiley Online Library

Communication Quantum Dots

www.advmat.de

Oxidized Quasi-Carbon Nitride Quantum Dots Inhibit Ice Growth Guoying Bai, Zhiping Song, Hongya Geng, Dong Gao, Kai Liu, Shuwang Wu, Wei Rao, Liangqia Guo,* and Jianjun Wang* methods to control ice formation are under intensive investigation for both fundamental research and practical applications. For example, the current state-ofthe-art method of cryopreservation is the addition of high concentrations of cellpermeating cryoprotectants such as glycerol or dimethyl sulfoxide to minimize ice formation.[2] In nonvitreous cryopreservation, ice recrystallization during thawing is a major cause of cellular damage; in the vitrification approach, ice nucleation is a key problem because it may cause devitrification during the thawing process and consequently damage cells.[3] Note that these organic solvents at high concentrations are toxic for the cryopreserved cells/tissue or other cells/tissue of the transplant or transfusion recipients. Therefore, complicated and time-consuming post-thaw desolvent procedures are necessary.[2b,4] In nature, some organisms living in polar or desert regions have evolved special capabilities to regulate ice formation by producing antifreeze proteins or glycoproteins (AF(G)Ps). These AF(G)Ps protect them from freezing damage by effectively inhibiting ice growth and recrystallization.[5] Extracting natural AF(G)Ps from living organisms is typically an intricate and low-yield process. Although some AF(G)Ps have been synthesized or expressed,[6] they are often expensive and of low thermal stability.[7] Therefore, synthetic mimics of AF(G)Ps are more desirable because of their possibility of mass production. Over the past few decades, some synthetic AF(G)P-inspired compounds have been reported[8] and many novel cryoprotectants have been used for the cryopreservation of red blood cells (RBCs), including poly(vinyl alcohol), poly(ampholyte), simple carbohydrates, and oligosaccharides, and even some small-molecule phenolic-glycosides and aryl-glycosides.[2a,b,9] Recently, a supermolecule assembled from safranine O was demonstrated to exhibit planespecific thermal hysteresis (TH) (the temperature difference between the equilibrium melting point and freezing point of ice) activity.[10] Although many studies have been conducted in this field, fundamental understanding of AFPs for controlling ice formation and protecting living organisms from freezing damage remains elusive. Consequently, the synthesis of AFP mimics often relies on a trial-and-error method; general criteria for the design of AFP mimics are missing.

Antifreeze proteins (AFPs), a type of high-efficiency but expensive and often unstable biological antifreeze, have stimulated substantial interest in the search for synthetic mimics. However, only a few reported AFP mimics display thermal hysteresis, and general criteria for the design of AFP mimics remain unknown. Herein, oxidized quasi-carbon nitride quantum dots (OQCNs) are synthesized through an up-scalable bottom-up approach. They exhibit thermal-hysteresis activity, an ice-crystal shaping effect, and activity on ice-recrystallization inhibition. In the cryopreservation of sheep red blood cells, OQCNs improve cell recovery to more than twice that obtained by using a commercial cryoprotectant (hydroxyethyl starch) without the addition of any organic solvents. It is shown experimentally that OQCNs preferably bind onto the ice-crystal surface, which leads to the inhibition of ice-crystal growth due to the Kelvin effect. Further analysis reveals that the match of the distance between two neighboring tertiary N atoms on OQCNs with the repeated spacing of O atoms along the c-axis on the primary prism plane of ice lattice is critical for OQCNs to bind preferentially on ice crystals. Here, the application of graphitic carbon nitride derivatives for cryopreservation is reported for the first time.

The formation of ice crystals is ubiquitous in nature but sometimes highly undesirable because of its detrimental or even fatal effect on most biological systems, as well as its serious threat to the property and safety of humans.[1] Therefore,

G. Bai, H. Geng, D. Gao, K. Liu, S. Wu, Prof. J. Wang Key Laboratory of Green Printing Institute of Chemistry Chinese Academy of Sciences Beijing 100190, P. R. China E-mail: [email protected] G. Bai, H. Geng, D. Gao, K. Liu, S. Wu, Prof. J. Wang School of Chemistry and Chemical Engineering University of Chinese Academy of Sciences Beijing 100049, P. R. China Z. Song, Prof. L. Guo College of Chemistry Fuzhou University Fuzhou 350116, P. R. China E-mail: [email protected] Prof. W. Rao Technical Institute of Physics Chemistry Chinese Academy of Sciences Beijing 100190, P. R. China

DOI: 10.1002/adma.201606843

Adv. Mater. 2017, 29, 1606843

1606843 (1 of 8)

© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com

www.advmat.de

Most recently, we discovered that graphene oxide (GO) can adsorb onto ice-crystal surfaces and inhibit ice-crystal growth.[11] This inhibition is likely due to the arrangement of hydroxyl groups on the honeycomb-structured graphene surface matching the lattice of the ice crystal, which is similar to the arrangement of hydroxyl groups on the ice-crystal-binding face of the AFP from the beetle Tenebrio molitor.[12] In this work, we study the effect of graphitic-carbon nitride (g-CN) derivatives on ice growth. Notably, g-CN possesses a layered structure analogous to the structure of graphite and each plane is

composed of tri-s-triazine units containing alternating sp2hybridized C and N atoms and tertiary nitrogen N(C)3 groups as the junctions (Figure 1a).[13] Interestingly, the tertiary N possessing a lone electron pair can function as the acceptor of a hydrogen bond. In addition, the size of one tri-s-triazine unit is ≈7.13 Å, which is similar to the repeated spacing (7.35 Å) between O atoms in the lattice of hexagonal ice crystals along the c-axis and thus introduces the possibility of adsorbing onto the ice.[12,13] Moreover, the distance between two neighboring tertiary N atoms can be varied by substituting sp2-hybridized

Figure 1. a) Schematic of a perfect g-CN sheet. The lone pairs of the sp2-hybridized N atoms in hexatomic rings are orientated toward the pores and cannot form hydrogen bonds with H2O because of the spatially confined effect. However, the tertiary N, whose lone pair is not entirely conjugated with tri-s-triazine, can act as an acceptor of a hydrogen bond. b) AFM and c) TEM images of OQCNs-180-3 (the inset in (b) is the corresponding height profile of OQCNs-180-3 along the line). The possible in-plane structures of d) OCNs and e) OQCNs. The obvious difference between their in-plane structures is that in the OQCNs, the triazine rings are mostly substituted by benzene rings, leading to a greater distance between two adjacent tertiary Ns (N(C)3) (≈7.42 Å, which is estimated according to the bond lengths and angles: the average CN bond length in N(C)3 groups is 1.44 Å; the average CC bond length in benzene rings is 1.42 Å; the bond angles of CNC in N(C)3 groups and CC in benzene rings are all ≈120°). f) Primary prism face of hexagonal ice. ( ) O atoms; H atoms have been omitted for clarity; ( ) O atoms in another layer 0.92 Å behind those in the paper plane.

Adv. Mater. 2017, 29, 1606843

1606843 (2 of 8)



www.advmat.de

N by graphitic C as the difference of the average bond length between CNC in triazine rings and CCC in benzene rings. Therefore, the effect of the distance between adjacent hydrogen-bond binding sites on controlling ice formation can be investigated. Here, we synthesized different types of g-CN derivatives, oxidized g-CN quantum dots (OCNs) and oxidized quasicarbon nitride quantum dots (OQCNs), with different distances between the two adjacent tertiary N atoms to study their effects on controlling ice formation. OCNs were prepared via the topdown method described in our previous report,[14] whereas OQCNs were synthesized via a simple bottom-up approach that consists of a single step. The bottom-up method adopted here is much simpler than the common top-down method of preparing OCNs, which typically consists of complicated steps, including the synthesis of bulk g-CN, chemical oxidation, and ultrasonic exfoliation.[15] In addition, the reaction for the synthesis of OQCNs in an autoclave is solvent-free and up-scalable. Collectively, these features enable the large-scale production of OQCNs, which is meaningful for using this type of g-CN derivative as a material. Our experimental results show that these two types of g-CN derivatives exhibit distinct behaviors in controlling ice growth: ice crystals in the aqueous dispersion of OCNs behave similarly to those in pure water; interestingly, OQCNs shape ice crystals into hexagons and substantially inhibit their growth. This behavior is probably ascribed to the different distances between the two neighboring tertiary N atoms resulting from their different in-plane structures. Moreover, the application of OQCNs for the cryopreservation of sheep RBCs is demonstrated for the first time. Without the addition of any organic solvents, OQCNs significantly improve cell recovery to more than twice that obtained by using commercial cryoprotectant (hydroxyethyl starch (HES)) under the stringent test condition of slow thawing at 4 °C. The g-CN derivatives, a promising type of 2D material, have shown enormous potential for applications

extending from energy-related topics to other new emerging fields, such as bioimaging, disinfection, electrochemiluminescence detection, and ferromagnetism.[16] In this work, they are demonstrated for the first time to regulate ice growth and can be utilized for cryopreservation. Five types of OQCNs were synthesized by the bottom-up approach under different temperatures using citric acid and urea mixed in various ratios as precursors (see the Supporting Information for the experimental details). For simplicity, the as-prepared samples are named as OQCNs-T-r, where T and r represent the reaction temperature and molar ratio of the precursors (nurea/ncitric-acid), respectively. Typical atomic force microscopy (AFM) and transmission electron microscopy (TEM) images of OQCNs-180-3 (Figure 1b,c; see Figure S1, Supporting Information for TEM images of the other types of OQCNs) show that the OQCNs are ≈10 nm in diameter and 1.7 nm thick, which are close to the diameter and thickness of the OCNs prepared in our previous report.[14] This similarity suggests that the morphology of the OQCNs does not obviously differ from that of the OCNs. The X-ray diffraction (XRD) patterns of both OQCNs and OCNs (Figure S2 and S3, Supporting Information, respectively) show the characteristic g-CN interplanar stacking peak at 27.5°.[13,17] However, the characteristic diffraction peak at ≈13.1° arising from the in-plane repeated tri-s-triazine units is not observed in any of the XRD patterns of OCNs and OQCNs,[13,17] suggesting that the in-plane long-range-ordered tri-s-triazine structure does not exist for either OCNs or OQCNs because of the introduction of oxygen-containing functional groups. Notably, the OCNs and OQCNs have different specific in-plane structures and compositions according to the comparisons of their characterization results in Table 1 (see the Supporting Information for the characterizations). In the Fourier transform infrared (FTIR) spectrum of the OCNs, the absorption at 800 cm−1 indicates the presence of the tri-s-triazine rings;[18] by

Table 1. Comparisons of the preparation methods and characterization results for as-prepared OCNs and OQCNs. Samples

OCNs

Preparation method Precursors

OQCNs-180-3 OQCNs-180-6 OQCNs-200-1 OQCNs-200-3 OQCNs-200-6

Top-down

Bottom-up

g-CN

nurea:ncitric acid 3

Lateral dimension [nm] Thickness [nm] Major XRD peaks (2θ) −1]

Main FTIR peaks [cm

6

1

3

3

10

1.7

1.7

27.5°

27.5°

3387 νOH;

3441 νOH; 3177 νNH; 1710 νCO; 1662 νCN; 1594 νCC

6

3233 νNH; 1733 νCO; 1652, 1540, 1448, and 1383 heptazine-derived repeating units; 800 tri-s-triazine rings 0.64

nN/nC

0.17

0.26

0.17

1.63

2.16

1.08

3.49

4.39

7.79

10.84

8.45

10.84

11.00

29.33

37.80

34.12

36.09

a)

N1 [%]

20.25

N2b) [%]

7.79

C1c) [%]

16.42

41.99

a)(sp2-hybridized

0.26

0.27

N (CNC)); b)(tertiary N (N(C)3)); c)(graphitic C).

Adv. Mater. 2017, 29, 1606843

1606843 (3 of 8)



www.advmat.de

contrast, in the FTIR spectra of OQCNs, the peak at ≈800 cm−1 is not observed. The X-ray photoelectron spectroscopy (XPS) results clearly show that the atomic percent ratios of N/C for all OQCN samples are much lower than that for the OCNs, which can be rationalized by considering that the content of sp2hybridized N (CNC) is substantially reduced, whereas the content of tertiary N (N(C)3) changes little, and the content of graphitic C increases greatly according to the N 1s and C 1s core-level spectra, respectively. All of these factors show that the OQCNs differ from OCNs because the in-plane tri-s-triazine units do not exist on the OQCNs; most of the sp2-hybridized N atoms, ranging from 78% to 95% for five types of OQCNs, are substituted by graphitic C, resulting in the triazine rings being mostly substituted by benzene rings. Therefore, the OQCNs are not oxidized g-CN or GO. They are a 2D nanomaterial whose in-plane structure is intermediate between GO and OCNs. Therefore, we term this type of 2D nanomaterial as oxidized quasi-carbon nitride quantum dots. Based on the characteristic peaks of the functional groups in the FTIR spectra and the results of the XPS analysis, we propose the possible in-plane structures of OCNs and OQCNs as those shown in Figure 1d,e. Both OCNs and OQCNs are modified with oxygen-containing functional groups. However, the in-plane structure of OQCNs exhibits a greater distance of ≈7.42 Å between the two adjacent tertiary N atoms, which is closer to the hexagonal ice lattice along the c-axis on the primary prism plane (7.35 Å) (Figure 1f). To investigate the effects of the as-prepared OCNs and OQCNs on ice growth, we studied the growth of single ice crystals in their aqueous dispersions using an Otago nanoliter osmometer (see the Supporting Information for the experimental details). As illustrated in Figure 2, under supercooling temperature (ΔT, which is obtained by subtracting the desired temperature during experiments from the melting temperature) of 0.08 °C, the ice crystals in pure water present a flat disk shape; the ice grows so rapidly that the observation window is full of ice within 15 s (Movie S1, Supporting Information). Interestingly, the ice-crystal growth in the OCN aqueous dispersion displays almost no difference with that of pure water (see Movie S2, Supporting Information); by contrast, OQCNs shape the ice crystals into hexagons (we present here in Figure 2 only OQCNs-180-3 for demonstration, the images of the hexagon ice crystals shaped by other types of OQCNs are shown in Figure S10 in the Supporting Information), which is similar to the shape of ice crystals faceted by certain AFPs, such as AFPs from the ryegrass Lolium perenne, beetle T. molitor, and Microdera punctipennis dzungarica.[19] The hexagonal shaping of ice crystals in the OQCN aqueous dispersion may be ascribed to the binding of OQCNs to the prism plane of the ice crystal.[7e,20] In addition, the ice crystals in the OQCNs-180-3 aqueous dispersion grow much slower than those in pure water, suggesting that OQCNs-180-3 can inhibit ice growth (see Movie S3, Supporting Information). When the ΔT increases to 0.10 °C, explosive growth parallel to the a-axes of the ice crystal is observed for the ice crystal grown in the OQCNs-180-3 aqueous dispersion (Movie S4, Supporting Information); this growth is analogous to the sudden burst behaviors of ice crystals grown in AFP solutions when the ΔT is greater than their TH gaps.[21] The major differences between the OCNs and OQCNs are the in-plane structures and surface charge densities (Table S3,

Adv. Mater. 2017, 29, 1606843

Figure 2. Growth processes of single ice crystals in H2O, OCNs, and OQCNs-180-3 aqueous dispersions. The concentrations of both the OCNs and OQCNs-180-3 are 4.0 mg mL−1. The start times in the leftmost panels are set as the times when the temperatures started to decrease to the desired ΔT at the same rate of ≈3 °C min−1 (to make sure the ice crystals at the beginning are the same size, we did not select the times when the temperatures had been decreased to the desired ΔT, as the start times for the growth rates of ice crystals in different dispersions varied widely). Scale bar: 50 µm.

Supporting Information). After systematically investigating the effects of a series of nanoparticles with various charge densities and sizes on shaping ice crystals (Table S4, Supporting Information), one can safely conclude that the distinct effects of these g-CN derivatives on ice shaping are due to the different in-plane structures, i.e., the different distances between the two adjacent tertiary N atoms that serve as acceptors of hydrogen bonds. In the case of OCNs, the distance between the neighboring tertiary N atoms is ≈7.13 Å, resulting in a mismatch[7e,22] of ≈2.99% with the distance between the O atoms along the c-axis on the primary prism plane of hexagonal ice crystals. By contrast, the distance between two adjacent tertiary N atoms on OQCNs is ≈7.42 Å, giving rise to a smaller mismatch (≈0.95%). Therefore, we concluded that the smaller mismatch of OQCNs with the ice lattice contributes to their adsorption onto the icecrystal surface, which were verified by the modified ice affinity experiments in the following. To assess the ice growth inhibition capability of g-CN derivatives quantitatively, we further measured the growth rates of ice crystals in pure water, OCN, and various OQCN aqueous dispersions under different ΔTs. Meanwhile, the TH gaps of these materials were also measured. In this study, the growth rate of ice crystals before the sudden burst is only a few percentage

1606843 (4 of 8)



www.advmat.de

of the one after the sudden burst, therefore we define the TH as the temperature difference between the melting point and the temperature at which the sudden burst of ice crystal occurs so that the capability of different OQCNs in inhibiting ice growth can be compared quantitatively, though the growth rate of ice crystals before the sudden burst in this study is higher than the largest threshold growth rate (0.2 µm s−1) used for defining the TH in previous reports.[23] As shown in Figure 3a, the growth rate of ice crystals in the OCN aqueous dispersion is similar to that in pure water, which further verifies that the OCNs exert little effect on the growth of ice crystals. In contrast, only when ΔT is above a critical value, the growth rates of ice crystals in all types of OQCN aqueous dispersions increase substantially, suggesting that all of the types of OQCNs possess TH activity. The concentration dependence of TH activity is shown in Figure S11 in the Supporting Information. As the ice crystals grew rapidly and irregularly, which led to large errors in measuring the growth rates once the crystals burst suddenly, we measured the growth rates at the ΔTs ranging from 0.01 °C to the temperatures at which the sudden burst was observed. To validate the adsorption of OQCNs onto the ice-crystal surface and to reveal the origin of the difference in the TH activity of various types of OQCNs, we performed a modified ice affinity experiment (see the Supporting Information for the experimental details) that was inspired by the AFP purification method first reported by Davies and co-workers[24] and used in our previous work.[11] For comparison, exactly the same experiment was conducted on OCNs as a negative control because OCNs do not observably affect ice shaping and growth. Briefly, ice was formed by placing liquid nitrogen above the

OQCN or OCN aqueous dispersions with an initial concentration of 1.0 mg mL−1. The formed ice was then removed, and the concentration of OQCNs or OCNs in the ice was measured by UV–vis spectrophotometry or fluorescence emission spectrophotometry. This procedure represents one cycle and was repeated 21 times. With increasing number of experimental cycles, ice-growth rate reduces gradually as the distance between the liquid nitrogen and the dispersion increases gradually. Modified ice affinity experiments on each OQCN sample and on the OCN control were repeated three times, and the mean value was calculated. Figure 3b shows the change of the concentrations of OQCNs and OCNs in the ice melt with increasing cycle number. The content of OQCNs in the ice melt clearly decreased during the first several cycles and finally became constant. The initial decrease of the OQCN content was likely due to the decreasing number of OQCNs entrapped between the dendritic ice crystals formed because of the high growth rate, i.e., the constitutional supercooling.[25] With the decrease of the ice-growth rate, the constitutional supercooling was finally avoided, and the constant content of OQCNs in the ice melt was reached because of the adsorption equilibrium, verifying that OQCNs do adsorb onto the ice-crystal surface. In stark contrast, the content of OCNs decreased monotonously, indicating that OCNs do not adsorb preferentially on the icecrystal surface. Therefore, we propose that the adsorption of OQCNs onto the ice-crystal surface leads to curvatures on the ice-crystal surface between adsorbed OQCNs and thus suppress the further growth of ice because of the Kelvin effect, as illustrated in Figure 3c.[26] Note that the equilibrium adsorption concentrations of different OQCNs have different values,

Figure 3. a) The growth rates of ice crystals in various samples under different ΔTs. All of the sample concentrations are 4.0 mg mL−1. b) Modified ice affinity experiment results: the trends of the concentrations of OQCNs and OCNs in the ice phase with the experimental cycle number. c) Proposed mechanism of OQCNs on the ice growth inhibition. d) The relationship between the TH gaps of OQCNs and the value of the content of potential adsorption sites multiplied by the ratio CX–H–Y%/IXRD,2θ =27.5°.

Adv. Mater. 2017, 29, 1606843

1606843 (5 of 8)



www.advmat.de

indicating their different adsorption capacities. The variation of the equilibrium adsorption concentrations is in accordance with the variation trend of the TH activities (Figure S12, Supporting Information). Thus, the TH activities of different types of OQCNs depend on their adsorption capabilities on the ice surface. This dependence may be because higher equilibrium adsorption amounts results in denser OQCNs on the ice-crystal surface, larger curvatures on the ice-crystal surface, and greater suppression of the freezing point based on the Kelvin effect. We further analyzed on the molecular level the reason why different types of OQCNs exhibit different TH activities. First, we classified the hydrogen-bond binding sites on OQCNs (Figure S6, Supporting Information) into two categories: the tertiary N sites and the remaining hydrogen-bond binding sites, including OH, COOH, and NH2. Because the distance between the neighboring tertiary N atoms on OQCNs matches the repeated spacing of oxygen atoms along the c-axis on the primary prism plane of the ice lattice, the tertiary N atoms serve as “potential adsorption sites.” Because the distance between the second category does not match the ice lattice, these hydrogen-bond binding sites cannot preferentially adsorb onto the ice-crystal surface with a great excess of liquid water, as verified by the results of modified ice affinity experiments with the OCNs. However, these hydrogen-bond binding sites can form hydrogen bonds with H2O, hindering the interlayer π–π stacking; COOH or NH2 can facilitate the solvation and electrostatic repulsion of OQCNs, contributing to the dispersion of OQCNs in water (Figure S13, Supporting Information).

Therefore, their presence increases the number of potential adsorption sites exposed to water for a fixed mass concentration of OQCNs. According to the experimental data analysis, the TH activities do not directly depend on the content of the tertiary N sites (Table 1 and Table S2, Supporting Information), i.e., potential adsorption sites. This independence of the TH activities is attributable to some regions on the OQCNs not being exposed to water because of the interlayer π–π stacking, which results in some potential adsorption sites being covered and inactivated for preferred adsorption. To directly quantify the degrees of interlayer π–π stacking of different types of OQCNs in water, we calculated the ratios between the content of OQCNs’ hydrogen-bond binding sites (CX–H–Y) and the intensity of the XRD peaks at 2θ = 27.5° (IXRD,2θ=27.5°, which is measured under exactly the same conditions for different samples and indicates the degree of OQCNs’ interlayer π–π stacking) (see Table S5, Supporting Information). As the hydrogenbond binding sites of the OQCNs contribute to preventing their interlayer π–π stacking in water, the ratio correlates the exposure degrees of OQCNs to water. Indeed, as shown in Figure 3d, the TH activity increases with the increasing value of the content of potential adsorption sites multiplied by the ratio (CX–H–Y%/IXRD,2θ=27.5°). That is, more potential adsorption sites exposed in water result in greater TH activity. If we assume that the process of the OQCN adsorption is simply irreversible, then according to the Langmuir-type kinetic model[27] (Section 3 of the Supporting Information), the OQCNs of a fixed mass with a higher content of exposed adsorption sites should provide

Figure 4. IRI of OQCNs and their application for cryopreservation of sheep RBCs. a,b) Optical microscopy images of ice crystals grown in OQCNs180-3 PBS dispersion (10.0 mg mL−1) (a) and PBS after annealing at −6 °C for 30 min (b). c) Quantitative assessment of the grain size of ice crystals grown in different OQCN PBS dispersions with various concentrations. d) The recovery of sheep RBCs cryopreserved in OQCNs-180-3 PBS dispersions with different concentrations and under a stringent test condition of slow thawing at 4 °C. The PBS and HES PBS dispersions are used as negative and positive controls, respectively.

Adv. Mater. 2017, 29, 1606843

1606843 (6 of 8)



www.advmat.de

a larger surface coverage fraction on the ice-crystal surface, resulting in greater curvatures on the ice-crystal surface and, thus, greater TH activity. Therefore, a mechanism by which the OQCNs inhibit ice growth is proposed. The prerequisite to the OQCNs’ TH activity is their preferred adsorption onto the ice surface, which necessitates a match of the distance of the OQCNs’ hydrogen-bond binding sites with that of the ice lattice. After the above condition has been met, the amount of these matched sites on the OQCNs exposed to water, which determines the adsorption amount on the ice crystal, becomes the key factor affecting the TH activity. The preferred adsorption of AFPs onto the ice-crystal surfaces is widely accepted as endowing them with ice-recrystallization inhibition (IRI) capability as well as the unique property of protecting living organisms from freezing damage,[28] although Ben and co-workers have demonstrated that some materials with no TH activity can also exhibit IRI activity because of the importance of hydration, which disturbs the three-dimensional hydrogen-bonding network of the ice–water interface.[29] Therefore, quantitative evaluations of the IRI activities of OQCNs were conducted via the splat cooling method (see the Supporting Information). Figure 4a,b displays typical optical microscopy images of recrystallized ice crystals from the PBS aqueous solutions with and without the addition of OQCNs. OQCNs can clearly suppress the recrystallization of ice crystals. Further IRI experimental results for all types of OQCNs in PBS dispersions with various concentrations are shown in Figure 4c. All of the OQCNs exhibit the capability to reduce the grain sizes of recrystallized ice crystals at various concentrations. On the basis of the obvious IRI activity, the OQCNs were investigated for the cryopreservation of sheep RBCs. The results of cytocompatibility assays (Figure S14, Supporting Information) indicate that OQCNs show no obvious cytotoxicity to sheep RBCs at concentrations up to 30 mg mL−1. OQCNs-180-3 with a series of concentrations less than 30.0 mg mL−1 were chosen to serve as cryoprotectants in sheep RBC cryopreservation. The experimental procedures follow the reported method[30]: the samples were rapidly frozen by immersion into liquid nitrogen and stored in liquid nitrogen (−196 °C); the thawing was performed by placing the samples in a refrigerator at 4 °C, which was chosen to maximize cell stress, thus providing a stringent test of the cryopreservative performance (see the Supporting Information for the experimental details). As shown in Figure 4d, the cell recovery first increased with increasing concentration of OQCNs-180-3 and then decreased when the concentration was greater than 10 mg mL−1. This decrease of the cell recovery at higher concentrations has also been reported by other groups; it is likely due to the ice shaping.[2a,31] Note that OQCNs-180-3 with a concentration of 10.0 mg mL−1 significantly improved the cell recovery to ≈55% without the addition of any organic solvents; this cell recovery is superior to that achieved by commercial HES (≈25%) and to previously reported recoveries achieved by poly(vinyl alcohol).[2a,30] In summary, to mimic AFPs in controlling ice formation and protecting living organisms from freezing damage, we synthesized OQCNs through a simple and up-scalable approach. Our investigations show that OQCNs possess the ability to shape ice crystals and inhibit ice growth/recrystallization. We also

Adv. Mater. 2017, 29, 1606843

demonstrated through the cryopreservation of sheep RBCs that OQCNs are promising as a cryoprotectant. The unique capability of OQCNs to control ice formation possibly benefits from the specific arrangement of tertiary N atoms, which facilitates the preferred adsorption of OQCNs onto the ice-crystal surface.

Supporting Information Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgements The authors are grateful for the financial support from the 973 Program (2013CB933004), the Chinese National Nature Science Foundation (51436004 and 21421061), and the Strategic Priority Research Program of the Chinese Academy of Sciences (CAS) (Grant No. XDA09020000). Sheep RBCs were purchased on demand from Beijing Abcam Technology Ltd.

Conflict of Interest The authors declare no conflict of interest.

Keywords antifreeze protein mimics, ice-growth inhibition, oxidized carbon nitride quantum dots, oxidized quasi-carbon nitride quantum dots Received: December 19, 2016 Revised: March 24, 2017 Published online: May 17, 2017

[1] a) P. Eberle, M. K. Tiwari, T. Maitra, D. Poulikakos, Nanoscale 2014, 6, 4874; b) J. Lv, Y. Song, L. Jiang, J. Wang, ACS Nano 2014, 8, 3152. [2] a) R. C. Deller, M. Vatish, D. A. Mitchell, M. I. Gibson, Nat. Commun. 2014, 5, 3244; b) J. G. Briard, J. S. Poisson, T. R. Turner, C. J. Capicciotti, J. P. Acker, R. N. Ben, Sci. Rep. 2016, 6, 23619; c) J. G. Day, G. N. Stacey, Cryopreservation and Freeze-Drying Protocols, Humana, Totowa, NJ, USA 2007. [3] B. Wowk, E. Leitl, C. M. Rasch, N. M. Karimi, S. B. Harris, G. M. Fahy, Cryobiology 2000, 40, 228. [4] a) J. F. Carpenter, T. N. Hansen, Proc. Natl. Acad. Sci. USA 1992, 89, 8953; b) G. M. Fahy, J. Saur, R. J. Williams, Cryobiology 1990, 27, 492; c) C. Polge, A. U. Smith, A. S. Parkes, Nature 1949, 164, 666. [5] a) L. L. C. Olijve, K. Meister, A. L. DeVries, J. G. Duman, S. Guo, H. J. Bakker, I. K. Voets, Proc. Natl. Acad. Sci. USA 2016, 113, 3740; b) Y. Yeh, R. E. Feeney, Chem. Rev. 1996, 96, 601; c) P. L. Davies, J. Baardsnes, M. J. Kuiper, V. K. Walker, Philos. Trans. R. Soc., B 2002, 357, 927. [6] a) Y. Tachibana, G. L. Fletcher, N. Fujitani, S. Tsuda, K. Monde, S. I. Nishimura, Angew. Chem., Int. Ed. 2004, 116, 874; b) L. Qiu, J. Ma, J. Wang, F. Zhang, Y. Wang, Cryobiology 2010, 60, 192. [7] a) G. Deng, R. A. Laursen, Biochim. Biophys. Acta 1998, 1388, 305; b) H. Chao, R. S. Hodges, C. M. Kay, S. Y. Gauthier, P. L. Davies, Protein Sci. 1996, 5, 1150; c) S. Y. Gauthier, C. B. Marshall, G. L. Fletcher, P. L. Davies, FEBS J. 2005, 272, 4439; d) C. B. Marshall, A. Chakrabartty, P. L. Davies, J. Biol. Chem. 2005, 280, 17920; e) C. Budke, T. Koop, Chem. Phys. Chem. 2006, 7, 2601.

1606843 (7 of 8)



www.advmat.de

[8] a) Y. Mastai, J. Rudloff, H. Cölfen, M. Antonietti, Chem. Phys. Chem. 2002, 3, 119; b) E. Baruch, Y. Mastai, Macromol. Rapid Commun. 2007, 28, 2256; c) T. Inada, S. S. Lu, Cryst. Growth Des. 2003, 3, 747; d) T. Inada, P. R. Modak, Chem. Eng. Sci. 2006, 61, 3149; e) C. J. Capicciotti, M. Leclère, F. A. Perras, D. L. Bryce, H. Paulin, J. Harden, Y. Liu, R. N. Ben, Chem. Sci. 2012, 3, 1408. [9] a) D. E. Mitchell, N. R. Cameron, M. I. Gibson, Chem. Commun. 2015, 51, 12977; b) C. Stoll, J. L. Holovati, J. P. Acker, W. F. Wolkers, Biotechnol. Prog. 2012, 28, 364; c) G. B. Quan, Y. Han, M. X. Liu, L. Fang, W. Du, S. P. Ren, J. X. Wang, Y. Wang, Cryobiology 2011, 62, 135; d) C. J. Capicciotti, J. D. R. Kurach, T. R. Turner, R. S. Mancini, J. P. Acker, R. N. Ben, Sci. Rep. 2015, 5, 9692. [10] R. Drori, C. Li, C. Hu, P. Raiteri, A. L. Rohl, M. D. Ward, B. Kahr, J. Am. Chem. Soc. 2016, 138, 13396. [11] H. Geng, X. Liu, G. Shi, G. Bai, J. Ma, J. Chen, Z. Wu, Y. Song, H. Fang, J. Wang, Angew. Chem., Int. Ed. 2017, 56, 997. [12] Y. C. Liou, A. Tocilj, P. L. Davies, Z. Jia, Nature 2000, 406, 322. [13] X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen, M. Antonietti, Nat. Mater. 2009, 8, 76. [14] Z. Song, T. Lin, L. Lin, S. Lin, F. Fu, X. Wang, L. Guo, Angew. Chem., Int. Ed. 2016, 55, 2773. [15] a) X. Zhang, H. Wang, H. Wang, Q. Zhang, J. Xie, Y. Tian, J. Wang, Y. Xie, Adv. Mater. 2014, 26, 4438; b) J. Oh, R. J. Yoo, S. Y. Kim, Y. J. Lee, D. W. Kim, S. Park, Chem. Eur. J. 2015, 21, 6241; c) W. Wang, J. C. Yu, Z. Shen, D. K. L. Chan, T. Gu, Chem. Commun. 2014, 50, 10148. [16] a) J. Zhang, Y. Chen, X. Wang, Energy Environ. Sci. 2015, 8, 3092; b) W. Ma, D. Han, M. Zhou, H. Sun, L. Wang, X. Dong, L. Niu, Chem. Sci. 2014, 5, 3946. [17] a) A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J.-O. Müller, R. Schlögl, J. M. Carlsson, J. Mater. Chem. 2008, 18, 4893; b) J. Li, Y. Zhang, X. Zhang, J. Han, Y. Wang, L. Gu, Z. Zhang, X. Wang, J. Jian, P. Xu, B. Song, ACS Appl. Mater. Interfaces 2015, 7, 19626. [18] a) B. Jürgens, E. Irran, J. Senker, P. Kroll, H. Müller, W. Schnick, J. Am. Chem. Soc. 2003, 125, 10288; b) B. V. Lotsch, M. Döblinger,

Adv. Mater. 2017, 29, 1606843

J. Sehnert, L. Seyfarth, J. Senker, O. Oeckler, W. Schnick, Chem. Eur. J. 2007, 13, 4969; c) B. V. Lotsch, W. Schnick, Chem. Mater. 2006, 18, 1891. [19] a) K. D. Kumble, J. Demmer, S. Fish, C. Hall, S. Corrales, A. DeAth, C. Elton, R. Prestidge, S. Luxmanan, C. J. Marshall, D. A. Wharton, Cryobiology 2008, 57, 263; b) M. Bar, Y. Celik, D. Fass, I. Braslavsky, Cryst. Growth Des. 2008, 8, 2954; c) K. Liu, C. Wang, J. Ma, G. Shi, X. Yao, H. Fang, Y. Song, J. Wang, Proc. Natl. Acad. Sci. USA 2016, 113, 14739. [20] J. A. Raymond, A. L. DeVries, Proc. Natl. Acad. Sci. USA 1977, 74, 2589. [21] R. Drori, P. L. Davies, I. Braslavsky, Langmuir 2015, 31, 5805. [22] P. V. Hobbs, Ice Physics, Oxford University Press, Oxford, UK 1974. [23] a) C. I. DeLuca, H. Chao, F. D. Sönnichsen, B. D. Sykes, P. L. Davies, Biophys. J. 1996, 71, 2346; b) O. Mizrahy, M. B. Dolev, S. Guy, I. Braslavsky, PLoS One 2013, 8, e59540. [24] M. J. Kuiper, C. Lankin, S. Y. Gauthier, V. K. Walker, P. L. Davies, Biochem. Biophys. Res. Commun. 2003, 300, 645. [25] S. Deville, E. Maire, G. B. Granger, A. Lasalle, A. Bogner, C. Gauthier, J. Leloup, C. Guizard, Nat. Mater. 2009, 8, 966. [26] C. A. Knight, Nature 2000, 406, 249. [27] K. Bartlomiej, K. J. M. Bishop, Lagzi, D. Wang, Y. H. Wei, S. B. Han, B. A. Grzybowski, Nat. Mater. 2012, 11, 227. [28] M. B. Dolev, I. Braslavsky, P. L. Davies, Annu. Rev. Biochem. 2016, 85, 515. [29] a) P. Czechura, R. Y. Tam, E. Dimitrijevic, A. V. Murphy, R. N. Ben, J. Am. Chem. Soc. 2008, 130, 2928; b) R. Y. Tam, S. S. Ferreira, P. Czechura, J. L. Chaytor, R. N. Ben, J. Am. Chem. Soc. 2008, 130, 17494; c) A. K. Balcerzak, C. J. Capicciotti, J. G. Briard, R. N. Ben, RSC Adv. 2014, 4, 42682. [30] D. E. Mitchell, J. R. Lovett, S. P. Armes, M. I. Gibson, Angew. Chem., Int. Ed. 2016, 128, 2851. [31] J. F. Carpenter, T. N. Hansen, Proc. Natl. Acad. Sci. USA 1992, 89, 8953.

1606843 (8 of 8)
