Investigations on Aging Behavior and Mechanism of Polyurea Coating in Marine Atmosphere

In this investigation, the aging behaviors of polyurea coating exposed to marine atmosphere for 150 days were studied and the mechanism was analyzed. The influences on surface and mechanical properties, surface morphology, thermal stability behavior, as well as chemical changes evolution of the coating were investigated. By attenuated total reflectance fourier transform infrared spectroscopy (ATR–FTIR) and X-ray photoelectron spectroscopy (XPS), changes in the chemical properties of polyurea coatings before (PCB) and after 150 d (PCA) of aging were analyzed, and emphasis was given to the effect of aging on functional group change, the hydrogen bonding behavior, and phase separated morphology. The results displayed prominent chain scission during aging, such as N–H, C=O, and C–O–C and the hydrogen bonded urea carbonyl content showed a decrease trend. The relative content of soft and hard segments showed a significant change, which increased the degree of phase separation.


Introduction
Polyurea as a thermoset elastomer has excellent performance for its special phase structure and physical crosslinks, which is formed by reacting a diisocyanate with an amine terminated compound by a step growth polymerization process [1][2][3][4]. Due to its exceptional mechanical and physical properties, chemical and moisture resistance, polyurea has been extensively applied in a wide number of applications. [1,3] For example, polyurea has been found to change the failure and fragmentation behavior upon impact by hypervelocity projectiles and is applied in military fields such as bulletproof vests [5]. In recent years, polyurea has been increasingly used in marine engineering, such as dock steel piles, sea-crossing bridges, and drilling platforms. Polyurea has also been studied as a binder for marine joint structures exposed to impact loads of varying amplitudes and strain rates. Thus, polyurea's moisture resistance properties and its impact mitigation capacity over a wide range of frequencies make it an ideal candidate for applications that involve exposure to or submergence in sea water [6].
As is known, there are aggressive factors such as ultraviolet radiation, floating and pushing fouling, aggressive chemicals, mechanical stresses, moisture and so on in the ocean atmosphere, which can not only cause failure of coating, but also pose a threat to property and life safety [7,8]. Based on this background, research work on the constitutive properties of polyurea are being given increasingly more interest [9][10][11][12][13]. Youssef [14] investigated the effect of UV radiation on the dynamic mechanical properties of polyurea samples and found that the dynamic creep modulus increased with an increase in the UV exposure duration. Whitten [15] studied the color change from white to deep tan after extended

Properties of Polyurea Coating
The polyurea used in this investigation is pure sheet-like coating without spraying on any substrate, and is mainly applied to marine steel structure protection in engineering. It contains zero Volatile Organic Content, which is extremely friendly to the environment. Due to the fast curing speed and insensitivity to moisture and temperature, it can be continuously sprayed on a sloping or vertical surface without sagging or running.

Experimental Method
The experimental site is located at the Qingdao University of Technology, which is situated 4 kilometers away from the actual coastline. Qingdao is situated in the southeastern part of the Shandong Peninsula, borders the Yellow Sea, and is of a temperate monsoon climate, where the annual average relative humidity, temperature, as well as precipitation are 70%, 12.3 °C, and 680 mm, respectively. The polyurea coating used in the experiment should be hung on the exposed frame in the artificial sea pool in accordance with the requirements of GB/T9276-1996 standard. The exposure frame is oriented 45° to the south, ensuring that it is fully exposed to natural sunlight, as shown in Figure 2. The exposure time is 150 days, from mid-March to mid-August.

Properties of Polyurea Coating
The polyurea used in this investigation is pure sheet-like coating without spraying on any substrate, and is mainly applied to marine steel structure protection in engineering. It contains zero Volatile Organic Content, which is extremely friendly to the environment. Due to the fast curing speed and insensitivity to moisture and temperature, it can be continuously sprayed on a sloping or vertical surface without sagging or running.

Experimental Method
The experimental site is located at the Qingdao University of Technology, which is situated 4 kilometers away from the actual coastline. Qingdao is situated in the southeastern part of the Shandong Peninsula, borders the Yellow Sea, and is of a temperate monsoon climate, where the annual average relative humidity, temperature, as well as precipitation are 70%, 12.3 • C, and 680 mm, respectively. The polyurea coating used in the experiment should be hung on the exposed frame in the artificial sea pool in accordance with the requirements of GB/T9276-1996 standard. The exposure frame is oriented 45 • to the south, ensuring that it is fully exposed to natural sunlight, as shown in Figure 2. The exposure time is 150 days, from mid-March to mid-August.

Contact Angle Measurements
Static contact angle measurements were carried out at room temperature on an SDC-200 contact angle goniometer (Dongguan Shengding Precision instrument Co., Ltd., Dongguan, China), using a syringe with a needle to drop a droplet of distilled water on the surface of the sample to be tested, which was then stopped for a few seconds, waiting for the drop to stabilize, and then digital images of the droplet silhouette were captured. The contact angle was measured within 30 s. The polar and dispersive components of the surface energy were calculated using the Fowkes method. Before this measurement, the sample surface before and after aging was ultrasonically cleaned with alcohol to remove impurities and oil.

Mechanical Properties
Tensile properties were conducted using MZ-4000D1 universal testing machine (Jiangsu Mingzhu Test Machinery Co., Ltd., Yangzhou, China) at a constant speed of 500 mm / min and at room temperature (25 °C ). The dimensions of tensile test samples were taken according to the ASTM D 412 (Standard Test Methods for Vulcanized Rubber and Thermoplastic Elastomers-Tension).

Scanning Electron Microscope
The microstructure of the sample before and after aging was observed by a JSM-7500F scanning electron microscope produced by JEOL (Beijing, China). The resolutions were 1.0 nm (15 kV) and 1.4 nm (1 KV), respectively. The accelerating voltage range was from 0.1KV to 30 KV, and the electron gun used a tungsten filament lamp with magnifications ranging from 25 times to 1,000,000 times. Before the test, the sample was gently wiped with anhydrous ethanol to remove surface dust and was then dried. Then, the sample was sputtered-coated with gold to form a conductive film to improve the image quality and resolution.

Atomic Force Microscopy
Tapping mode atomic force microscopy (AFM) was performed with an Ntegra Prima Scanning Probe system from NT-MDT Prima (NT-MDT Spectrum Instruments Corp., Moscow, Russia). Topographic (height) and phase images were recorded simultaneously under ambient conditions. A silicon cantilever probe with a force constant of 5 N·m −1 and a resonance frequency of 70 kHz was used to work on the tapping mode. Before the test, a 15 mm × 15 mm square specimen of PCB and PCA were prepared, and the cleaned samples were obtained by absolute ethanol to remove surface dust.

Contact Angle Measurements
Static contact angle measurements were carried out at room temperature on an SDC-200 contact angle goniometer (Dongguan Shengding Precision instrument Co., Ltd., Dongguan, China), using a syringe with a needle to drop a droplet of distilled water on the surface of the sample to be tested, which was then stopped for a few seconds, waiting for the drop to stabilize, and then digital images of the droplet silhouette were captured. The contact angle was measured within 30 s. The polar and dispersive components of the surface energy were calculated using the Fowkes method. Before this measurement, the sample surface before and after aging was ultrasonically cleaned with alcohol to remove impurities and oil.

Mechanical Properties
Tensile properties were conducted using MZ-4000D 1 universal testing machine (Jiangsu Mingzhu Test Machinery Co., Ltd., Yangzhou, China) at a constant speed of 500 mm/min and at room temperature (25 • C). The dimensions of tensile test samples were taken according to the ASTM D 412 (Standard Test Methods for Vulcanized Rubber and Thermoplastic Elastomers-Tension).

Scanning Electron Microscope
The microstructure of the sample before and after aging was observed by a JSM-7500F scanning electron microscope produced by JEOL (Beijing, China). The resolutions were 1.0 nm (15 kV) and 1.4 nm (1 KV), respectively. The accelerating voltage range was from 0.1KV to 30 KV, and the electron gun used a tungsten filament lamp with magnifications ranging from 25 times to 1,000,000 times. Before the test, the sample was gently wiped with anhydrous ethanol to remove surface dust and was then dried. Then, the sample was sputtered-coated with gold to form a conductive film to improve the image quality and resolution.

Atomic Force Microscopy
Tapping mode atomic force microscopy (AFM) was performed with an Ntegra Prima Scanning Probe system from NT-MDT Prima (NT-MDT Spectrum Instruments Corp., Moscow, Russia). Topographic (height) and phase images were recorded simultaneously under ambient conditions. A silicon cantilever probe with a force constant of 5 N·m −1 and a resonance frequency of 70 kHz was used to work on the tapping mode. Before the test, a 15 mm × 15 mm square specimen of PCB and PCA were prepared, and the cleaned samples were obtained by absolute ethanol to remove surface dust.

Thermogravimetric Analysis
Thermogravimetric analysis (TGA) was performed on a STA449C (Netzsch Gerateball, Selb, Germany) under the N 2 atmosphere at 20 mL/min. In the experiment, a sample weighing approximately 8 mg was heated at 10 • C/min from room temperature to 700 • C.
2.3.6. Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy ATR-FTIR (Attenuated Total Reflectance) spectra of PCB and PCA were investigated with VERTEX 70 FTIR spectrometer (Bruker Optics, Inc., Ettlingen, Germany) with 32 scans and a resolution of 4 cm −1 .The spectral region was from 4000 to 500 cm −1 and the surface of samples for measurements was cleaned by anhydrous ethanol. The degree of degradation could be detected by detecting changes in the dipole moment associated with the stretching vibration of the functional group.

X-ray Photoelectron Spectroscopic
The X-ray photoelectron spectroscopic (XPS) experiments were performed on Thermo ESCALAB 250XI system (AXIS SUPRA DLD , Shimadzu Kratos Inc., Kyoto, Japan) with Al K α (hv = 1486.6 eV) radiation. The binding energy (BE) scale was regulated by setting the C1s transition at 284.8 eV. The accuracy of the BE values was ± 0.2 eV. The surface of the measurement samples was cleaned by anhydrous ethanol and then naturally dried.

Surface and Mechanical Properties
Gloss refers to the ability of a surface to reflect light projected thereon and can be used to indicate the degree of aging of the coating. Figure 3 showed the change of gloss of polyurea coating for 150 days. With extended exposure time, the glossiness value showed a decreasing tendency. As the aging time reached 150 days, the glossiness value decreased 91.95%. Literature [38] stated that the greater the roughness of the coating surface, the lower the reflectivity and gloss value obtained, so the decrease of the coating gloss corresponded well to the cracks of the coating surface.   Determining the water contact angle can provide information about the change in the surface topography through the wetting properties [39,40]. The contact angle results and calculated surface energies of the polyurea coatings were presented in Figure 4. As observed, a steady decrease in the value of the contact angle was evident. The contact angle value of the coating before aging was about 78 • ± 1.2 • and dropped to 37.7 • ± 2.1 • , as the aging time reached 150 days, which decreased by about 51.28%. The result indicated a significant change in its wettability on the sample. As the value decreased, Materials 2019, 12, 3636 6 of 15 the levels of hydrophilicity increased, which could increase the adhesion of contaminants [41]. This was one of reasons for the decrease in gloss.

Surface Morphology and Topography
The morphology changes on the surface of samples before and after ocean atmosphere exposure for 30 d, 60 d, 90 d, 120 d, and 150 d were obtained by SEM in Figure 6. As shown in Figure 6, the In this section, the aging impact on the tensile properties of the polyurea coating was investigated. Figure 5 showed the result of a polyurea coating during natural exposure to the ocean atmosphere for 150 days. A slight increase in the tensile strength values was observed in the sample during the early stage of aging. This related to the enhancement of molecular cross-linking in the early aging period, and in this period, the effect of molecular cross-linking was greater than molecular bond rupture [26]. However, the tensile strength presented the descent trend on the whole, in the long term, and the tensile strength reduced by 7.44% after exposure for 150 days.

Surface Morphology and Topography
The morphology changes on the surface of samples before and after ocean atmosphere exposure for 30 d, 60 d, 90 d, 120 d, and 150 d were obtained by SEM in Figure 6. As shown in Figure 6, the

Surface Morphology and Topography
The morphology changes on the surface of samples before and after ocean atmosphere exposure for 30 d, 60 d, 90 d, 120 d, and 150 d were obtained by SEM in Figure 6. As shown in Figure 6, the surface of the unaged coating was relatively smooth and homogeneous, and no obvious holes and cracks appeared. With an increase in the aging time, the coating surface generated more cracks, which resulted in the increase of the surface roughness. The defect area calculated by the image J software is shown in Figure 7. The increase in area defects correlated to the exposure duration, which was consistent with research conducted by Youssef [14]. The contact angle decreased with an increase in the defect area, and the two were basically negatively correlated. The appearance of cracks on the coating was related to the breakage of molecular bonds. As found in the literature [38], under the combined action of violet radiation and other influencing factors, the molecules in the coating could degrade to form micropores, and then microcracks had developed. The degradation products might also leave the surface, creating a rough surface.
surface of the unaged coating was relatively smooth and homogeneous, and no obvious holes and cracks appeared. With an increase in the aging time, the coating surface generated more cracks, which resulted in the increase of the surface roughness. The defect area calculated by the image J software is shown in Figure 7. The increase in area defects correlated to the exposure duration, which was consistent with research conducted by Youssef [14]. The contact angle decreased with an increase in the defect area, and the two were basically negatively correlated. The appearance of cracks on the coating was related to the breakage of molecular bonds. As found in the literature [38], under the combined action of violet radiation and other influencing factors, the molecules in the coating could degrade to form micropores, and then microcracks had developed. The degradation products might also leave the surface, creating a rough surface.  The surface morphology of PCB and PCA were probed using tapping mode AFM, and the phase contrast, and topographic images were shown in Figures 8 and 9. Figure 8 provided direct visual evidence of two distinct phases-the darker areas correspond to the soft segment domain and the lighter areas correspond to the hard segment domain [42], the hard segment dispersed randomly within the soft segment, forming a microphase separation structure. Compared to the PCB, the PBA had a higher ordinate, indicating a greater degree of microphase separation. Figure 9 showed AFM resulted in the increase of the surface roughness. The defect area calculated by the image J software is shown in Figure 7. The increase in area defects correlated to the exposure duration, which was consistent with research conducted by Youssef [14]. The contact angle decreased with an increase in the defect area, and the two were basically negatively correlated. The appearance of cracks on the coating was related to the breakage of molecular bonds. As found in the literature [38], under the combined action of violet radiation and other influencing factors, the molecules in the coating could degrade to form micropores, and then microcracks had developed. The degradation products might also leave the surface, creating a rough surface.  The surface morphology of PCB and PCA were probed using tapping mode AFM, and the phase contrast, and topographic images were shown in Figures 8 and 9. Figure 8 provided direct visual evidence of two distinct phases-the darker areas correspond to the soft segment domain and the lighter areas correspond to the hard segment domain [42], the hard segment dispersed randomly within the soft segment, forming a microphase separation structure. Compared to the PCB, the PBA had a higher ordinate, indicating a greater degree of microphase separation. Figure 9 showed AFM The surface morphology of PCB and PCA were probed using tapping mode AFM, and the phase contrast, and topographic images were shown in Figures 8 and 9. Figure 8 provided direct visual evidence of two distinct phases-the darker areas correspond to the soft segment domain and the lighter areas correspond to the hard segment domain [42], the hard segment dispersed randomly within the soft segment, forming a microphase separation structure. Compared to the PCB, the PBA had a higher ordinate, indicating a greater degree of microphase separation. Figure 9 showed AFM 2D (Figure 9a,b) and 3D (Figure 9c,d) topographic image representations of PCB and PCA. We can clearly observe in Figure 9c,d, using a qualitative approach, that an increase in surface roughness was achieved after exposure in the marine atmosphere. According to the NOVA software analysis, the root mean square roughness (R rms ) value for the original sample was 6.03 nm and the value of average roughness (R a ) was 3.24 nm. The surface topography was relatively flat. As the exposure time reach 150 days, an obvious increase in the roughness degree values could be observed, the R rms value reached 49.50 nm while the R a was close to 36.3 nm. Surface roughness was the main reason for reducing the contact angle [41] which explains the reason for the above-mentioned contact angle reduction. The surface height of PCB and PCA was analyzed by the NOVA software, and the results are shown in Figure 9e. As observed, there are two different roughness profile zones-some very rough zones with a surface height from 50 to 300 nm and other zones had very low roughness values in the range of less than 50 nm.
reducing the contact angle [41] which explains the reason for the above-mentioned contact angle reduction. The surface height of PCB and PCA was analyzed by the NOVA software, and the results are shown in Figure 9e. As observed, there are two different roughness profile zones-some very rough zones with a surface height from 50 to 300 nm and other zones had very low roughness values in the range of less than 50 nm. reducing the contact angle [41] which explains the reason for the above-mentioned contact angle reduction. The surface height of PCB and PCA was analyzed by the NOVA software, and the results are shown in Figure 9e. As observed, there are two different roughness profile zones-some very rough zones with a surface height from 50 to 300 nm and other zones had very low roughness values in the range of less than 50 nm.

Thermal Stabilities
Thermal gravity (TG) weight loss curves and the differential thermal gravity (DTG) for PCB an PCA are shown in Figure 10. The onset degradation temperatures are defined by the temperatures 5% weight loss in TGA curves, whereas temperatures of the maximum degradation rate we evaluated by the peaks in the DTG curves. It can be seen that PCB and PCA both exhibited only o degradation step in nitrogen atmosphere during 250-500 °C . Compared to PCB, PCA had a bett

Thermal Stabilities
Thermal gravity (TG) weight loss curves and the differential thermal gravity (DTG) for PCB and PCA are shown in Figure 10. The onset degradation temperatures are defined by the temperatures of 5% weight loss in TGA curves, whereas temperatures of the maximum degradation rate were evaluated by the peaks in the DTG curves. It can be seen that PCB and PCA both exhibited only one degradation step in nitrogen atmosphere during 250-500 • C. Compared to PCB, PCA had a better thermal stability, for it started to degrade at as high as 297 • C and the maximum weight loss rate occurred at 413 • C, while the coating before aging started to degrade and the maximum weight loss rate occurred at the temperature of 272 • C and 403 • C, respectively. The results of this experiment had the following implications. First, it is possible for PCA to degrade into small molecules, which has a better heat resistance than macromolecules. Second, the soft segment content is reduced, and the coating with a high hard segment content exhibits better heat resistance [42]. Practical reasons need further experimentation.

Thermal Stabilities
Thermal gravity (TG) weight loss curves and the differential thermal gravity (DTG) for PCB and PCA are shown in Figure 10. The onset degradation temperatures are defined by the temperatures of 5% weight loss in TGA curves, whereas temperatures of the maximum degradation rate were evaluated by the peaks in the DTG curves. It can be seen that PCB and PCA both exhibited only one degradation step in nitrogen atmosphere during 250-500 °C . Compared to PCB, PCA had a better thermal stability, for it started to degrade at as high as 297 °C and the maximum weight loss rate occurred at 413 °C , while the coating before aging started to degrade and the maximum weight loss rate occurred at the temperature of 272 °C and 403 °C , respectively. The results of this experiment had the following implications. First, it is possible for PCA to degrade into small molecules, which has a better heat resistance than macromolecules. Second, the soft segment content is reduced, and the coating with a high hard segment content exhibits better heat resistance [42]. Practical reasons need further experimentation.

Chemical Changes
The FTIR spectra of PCB and PCA are shown in Figure 11. The spectral bands corresponding to the hard segment domains were observed at 3,284-3,287 cm −1 for N-H stretching vibrations and at 1,600-1,700 cm −1 attributed to C=O stretching vibrations. The absorption bands pertaining to the C-N stretching vibrations were at around 1,540 cm −1 , all above were associated with the existence of urea bond in the interior of the composites. The Figure 11 shows that the major difference between the spectra corresponding to PCB and PCA was their intensity of bands. A decrease in the intensity signal Plane/(μm) Figure 10. TG and DTG curves for polyurea before and after aging.

Chemical Changes
The FTIR spectra of PCB and PCA are shown in Figure 11. The spectral bands corresponding to the hard segment domains were observed at 3284-3287 cm −1 for N-H stretching vibrations and at 1600-1700 cm −1 attributed to C=O stretching vibrations. The absorption bands pertaining to the C-N stretching vibrations were at around 1540 cm −1 , all above were associated with the existence of urea bond in the interior of the composites. The Figure 11 shows that the major difference between the spectra corresponding to PCB and PCA was their intensity of bands. A decrease in the intensity signal of the aged sample indicated the beginning of chain scission or the disappearance of a certain group. For example, the absorbance peaks at 1020-1100 cm -1 decreased, confirming the cleavage of the C-O-C bond at exposure aging, and at the same time, the peak of C=O absorption decreased because of the scission of C=O band. The secondary amino N-H stretching vibration peak near 3287 cm −1 was weakened and broadened, and shifted towards a lower frequency, which indicated the existence of N-H bond cleavage and the reduction in the degree of hydrogen bonding. The C-H stretching vibration in the 2860-2970 cm −1 range was obviously weakened, probably the C-H bond in the D2000 side methyl and methylene had fractured [43]. The results coincided perfectly with the evolution of the surface energy and the tensile properties.
scission of C=O band. The secondary amino N-H stretching vibration peak near 3,287 cm −1 was weakened and broadened, and shifted towards a lower frequency, which indicated the existence of N-H bond cleavage and the reduction in the degree of hydrogen bonding. The C-H stretching vibration in the 2,860-2,970 cm −1 range was obviously weakened, probably the C-H bond in the D2000 side methyl and methylene had fractured [43]. The results coincided perfectly with the evolution of the surface energy and the tensile properties. Figure 11. FTIR spectra at the region of 4,000-500 cm −1 for unaged and aged polyurea.
Hydrogen bonding characteristic was one of the most notable features of the polyurea materials, especially polyurea materials with a high hard segment content [44]. As major objectives of hydrogen bonding behavior research, amino, and carbonyl largely determine the hard segment structure and affect the structure and properties of the material [30]. In order to investigate the effect of aging on hydrogen bonding degree, emphasis was given to the amino region and the urea carbonyl region of the spectrum.
The peak at about 3,440 cm −1 was assigned to the amino that had not formed a hydrogen bond. In the N-H region, the stretching vibration band of N-H of PCB and PCA showed in the frequency of 3,287.28 cm −1 and 3,284.46 cm −1 , respectively. The magnitude of the shift of the infrared absorption peak number was the indicator of the strength of the hydrogen bond, thus, the length of hydrogen bond can be calculated through the following Equation (1) [45].
where ∆V represented the shift distance from the amino band without hydrogen bonding to the existing amino band, while R denote the length of hydrogen bond. From the above formula, the lengths of the amino hydrogen bonds before and after aging were both calculated to be about 2.93 Å , thus, aging has little effect on the hydrogen bond length in the N-H region. A detailed analysis of hydrogen bonding characteristics was carried out by curve fitting the broad band in the carbonyl region of PCB and PCA. The band deconvolution analysis was performed on the carbonyl region at 1,620-1,690cm −1 using the OMNIC 8.2 software (Thermo nicolet, Madison, WI, USA). Hydrogen bonded carbonyl ureas were characterized by the presence of disordered and ordered urea carbonyls, consistent with previous reports [46,47]. The peaks due to the urea domain were characterized by bands at 1,626-1,645 cm −1 and 1,651-1,669 cm −1 , attributed to the ordered and disordered hydrogen-bonded urea carbonyls, respectively [43,47,48]. The absorption bands arising around 1,673-1,685 cm −1 corresponded to free hydrogen-bonded urea carbonyl. The wavenumber assignment areas as well as the hydrogen bonding degree of the deconvoluted bands are shown in Table 1. The degree of hydrogen bonding corresponding to the urethane carbonyl was calculated using Equation (2) [47]: Figure 11. FTIR spectra at the region of 4000-500 cm −1 for unaged and aged polyurea.
Hydrogen bonding characteristic was one of the most notable features of the polyurea materials, especially polyurea materials with a high hard segment content [44]. As major objectives of hydrogen bonding behavior research, amino, and carbonyl largely determine the hard segment structure and affect the structure and properties of the material [30]. In order to investigate the effect of aging on hydrogen bonding degree, emphasis was given to the amino region and the urea carbonyl region of the spectrum.
The peak at about 3440 cm −1 was assigned to the amino that had not formed a hydrogen bond. In the N-H region, the stretching vibration band of N-H of PCB and PCA showed in the frequency of 3287.28 cm −1 and 3284.46 cm −1 , respectively. The magnitude of the shift of the infrared absorption peak number was the indicator of the strength of the hydrogen bond, thus, the length of hydrogen bond can be calculated through the following Equation (1) [45].
where ∆V represented the shift distance from the amino band without hydrogen bonding to the existing amino band, while R denote the length of hydrogen bond. From the above formula, the lengths of the amino hydrogen bonds before and after aging were both calculated to be about 2.93 Å, thus, aging has little effect on the hydrogen bond length in the N-H region. A detailed analysis of hydrogen bonding characteristics was carried out by curve fitting the broad band in the carbonyl region of PCB and PCA. The band deconvolution analysis was performed on the carbonyl region at 1620-1690cm −1 using the OMNIC 8.2 software (Thermo nicolet, Madison, WI, USA). Hydrogen bonded carbonyl ureas were characterized by the presence of disordered and ordered urea carbonyls, consistent with previous reports [46,47]. The peaks due to the urea domain were characterized by bands at 1626-1645 cm −1 and 1651-1669 cm −1 , attributed to the ordered and disordered hydrogen-bonded urea carbonyls, respectively [43,47,48]. The absorption bands arising around 1673-1685 cm −1 corresponded to free hydrogen-bonded urea carbonyl. The wavenumber assignment areas as well as the hydrogen bonding degree of the deconvoluted bands are shown in Table 1. The degree of hydrogen bonding corresponding to the urethane carbonyl was calculated using Equation (2) [47]: where A o , A diso , and A free denote the area of ordered, disordered, and free hydrogen bonded urea carbonyl, respectively. Whereas, X o , X diso , and X b showed the percentage of ordered, disordered, and free hydrogen bonding. A comparative analysis of the relative trends in the hydrogen bonding degree of PCB and PCA needed to be considered separately. According to the data in Table 1, major differences between the two sets of samples towards the hydrogen bonding behavior were the degree of ordered, disordered, and total urea hydrogen bonding. The X o value of PCB and PCA were 57.66% and 46.13%, respectively, whereas the X b values were 83.41% and 74.56%. In contrast to PCB, a significant decrease in the X o and X b values of PCA was observed. The X o value showed a slight increase tendency compared with PCB. Based on the observations above, the ordered hydrogenated urea carbonyl content in PCA was less than PCB, which was related to the intermolecular force, for a reduction in the degree of hydrogen bonding can result in the weakening of the aggregation between hard segments and a decrease in the order of the hard segment structure [49]. Thus, results clearly demonstrated that the intermolecular force of the polyurea after aging was reduced. Literature pointed out [46], the hydrogen bonding in the region of carbonyl can limit the phase separation, as compared to the cases with less hydrogen bonding. The results showed the reduction of hydrogen bonding urea carbonyl content, which improved the degree of microphase separation in the soft and hard sections, in good agreement with the results of AFM.
In the XPS spectrum shown as Figure 12, three obvious signals were detected at 285.3 eV (C 1s), 399.8 eV (N 1s), and 532.7eV (O 1s). It can be seen that relative content of N, C, and O elements was slightly changed and no new elements were produced. The contents of N, C, and O elements and the ratio of elements of PCB and PCA were shown in Table 2. After aging, the relative content of the N element increased more obviously, while the O element and the C element decreased slightly. The oxygen-to-carbon (O/C) and the nitrogen-to-carbon (N/C) ratio of PCB was 0.2598 and 0.0565, respectively, whereas that of PCA was 0.2605 and 0.0765, respectively. The increase of O/C and N/C ratio indicated that the coating degraded during exposure in the marine atmosphere, and the degradation caused the inefficiency of coating.    In order to elucidate the changes in chemical groups, the C1 S spectrum was chosen and deconvolved with the XPSPEAK software as shown in Figure 11. The C1 S spectrum of PCB can be resolved into four contributions (Figure 13a). The peak with lowest BE, 284.8 eV, was assigned to carbon atoms which were linked to hydrogen (C-H) or carbon (C-C). The peak observed at 285.7 was assigned to the carbon atom bonded to the nitrogen atom (C-N). The peak located at 286.3 eV corresponded to carbon atoms single-bonded to oxygen atoms (C-O-C) [50]. According to the data in Table 3, the relative content of C-O-C before aging was estimated to be 39.860%, while C=O was about 1.859%. After 150 days of aging, the percentage of C-O-C and C=O content changed greatly, the C-O-C content fell to 24.626%, while the C=O content increased to 15.963%. This phenomenon can be explained as the bond breakage degree of the C-O-C in the soft segment after aging, which was much higher than that of the hard segment urea bond -NHCONH-, and the soft segment phase was more damaged than the hard segment. This indicated the different sensitiveness of the hard and soft segment to the attack of the marine atmosphere exposure, and that the soft segments were more susceptible to erosion, which explained the reason for a greater degree of microphase separation of PCA.  In order to elucidate the changes in chemical groups, the C1S spectrum was chosen and deconvolved with the XPSPEAK software as shown in Figure.11. The C1S spectrum of PCB can be resolved into four contributions (Figure 13a). The peak with lowest BE, 284.8 eV, was assigned to carbon atoms which were linked to hydrogen (C-H) or carbon (C-C). The peak observed at 285.7 was assigned to the carbon atom bonded to the nitrogen atom (C-N). The peak located at 286.3 eV corresponded to carbon atoms single-bonded to oxygen atoms (C-O-C) [50]. According to the data in Table 3, the relative content of C-O-C before aging was estimated to be 39.860%, while C=O was about 1.859%. After 150 days of aging, the percentage of C-O-C and C=O content changed greatly, the C-O-C content fell to 24.626%, while the C=O content increased to 15.963%. This phenomenon can be explained as the bond breakage degree of the C-O-C in the soft segment after aging, which was much higher than that of the hard segment urea bond -NHCONH-, and the soft segment phase was more damaged than the hard segment. This indicated the different sensitiveness of the hard and soft segment to the attack of the marine atmosphere exposure, and that the soft segments were more susceptible to erosion, which explained the reason for a greater degree of microphase separation of PCA.
(a) (b) Figure 13. Component-fitted XPS spectra of the C1s peaks of polyurea coating (a) before and (b) after aging. Table 3. Component-Fitted C1s Spectra of polyurea before and after aging. Through the above analysis results, we could draw up the partial aging mechanism shown in Scheme 1. As observed, aging had a great influence on the hydrogen bonding of the hard segment, especially the carbonyl region. After aging, the degree of hydrogen bonding in the carbonyl region was significantly reduced. The decrease of hydrogen bonding degree could weaken the aggregation between hard segments and reduce the order of the hard segment structure, which resulted in the decrease in the mechanical properties. Further studies found that the sensitivity of polyurea between soft and hard segment were significantly different, and the degree of C-O-C bond breakage in the soft segment was higher than that in the hard segment, which caused the soft segment to be etched, and then resulted in the increase of roughness and the microphase separation degree. Macroscopically, the gloss of the material was decreased, and the micro-cracks were generated on the surface of the coating on the microscopic surface, which was one of the important reasons for the decline in the mechanical properties.

Element C-C C-H C-N C-O-C C=O
Through the above analysis results, we could draw up the partial aging mechanism shown in Scheme 1. As observed, aging had a great influence on the hydrogen bonding of the hard segment, especially the carbonyl region. After aging, the degree of hydrogen bonding in the carbonyl region was significantly reduced. The decrease of hydrogen bonding degree could weaken the aggregation between hard segments and reduce the order of the hard segment structure, which resulted in the decrease in the mechanical properties. Further studies found that the sensitivity of polyurea between soft and hard segment were significantly different, and the degree of C-O-C bond breakage in the soft segment was higher than that in the hard segment, which caused the soft segment to be etched, and then resulted in the increase of roughness and the microphase separation degree. Macroscopically, the gloss of the material was decreased, and the micro-cracks were generated on the surface of the coating on the microscopic surface, which was one of the important reasons for the decline in the mechanical properties.