Probing Physicochemical Performances of 3D Printed Carbon Fiber Composites During 8‐Month Exposure to Space Environment

Carbon fiber reinforced polymers (CFRPs) offer exceptional properties that make them highly relevant in the aerospace industry, such as high thermal conductivity and an outstanding strength‐to‐weight ratio. Advances in additive manufacturing have expanded the aerospace applications of CFRPs, even allowing for in‐space fabrication of complex structures. Understanding the stability of CFRPs in the harsh conditions of low Earth orbit (LEO) is crucial. LEO exposes materials to extreme environmental factors, such as vacuum, radiation, atomic oxygen, and temperature fluctuations, which can accelerate degradation. To investigate the space‐environment effect on material, changes in properties of 3D‐printed CFRPs are compared with CFRPs made through forging and conventional compression molding. Surface analyses examine morphological, chemical, and matrix composition changes, along with an evaluation of mechanical integrity. Remarkably, the naked 3D printed CFRPs withstood 8 months of LEO exposure similar to the compression molded CFRP samples, with changes in chemical properties limited to the sample's outer surface. Further, despite no protective coatings are used, limited surface erosion and no variation in mechanical strength are observed. These results provide relevant information for the development and deployment of novel 3D printed CFRPs materials for a wide spectrum of terrestrial and space applications.


Introduction
Carbon fiber reinforced polymers (CFRP) are routinely adopted in the aerospace industry due to numerous advantages they such, as 3D printed CFRPs gain more prominence in aerospace applications, understanding their chemical and structural stability, particularly in the demanding conditions of low Earth orbit (LEO), becomes crucial for ensuring the safety and success of future space missions.
LEO, which spans from 200 to 1600 km from the Earth's surface, [2] subjects materials to harsh environmental factors that accelerate material degradation. [3]3c,4] These extreme conditions can lead to premature failure of critical components, emphasizing the need for a better insight in new CFRP materials for satellite and spacecraft design.While ground-based facilities attempt to simulate space environmental conditions, the reliability of the results obtained is often questionable.The failure-review board of the Hubble Space Telescope (HST) highlighted the underestimation of damage in space compared to durability tests conducted on Earth. [5]his discrepancy was observed in fluorinated ethylene propylene (FEP) surfaces, where degradation in space was more severe than predicted based on accelerated tests of a factor as high as ten.
Notably, while there is limited available literature on compression molded CFRPs exposed to LEO, there is complete absence of scientific data for 3D printed CFRPs.In this study, we performed a comprehensive evaluation of the performances of five CFRPs after direct protracted exposure to extreme LEO conditions.This comparative analysis involved exposing two 3D printed CFRPs, two forged and one conventional compression molded CFRPs samples outside the International Space Station (ISS) for over 8 months on the Nanoracks External Platform (NREP) located on the Japanese Experimental Module (JEM).
While coatings are often investigated as shielding components for underlying CFRP structures [6] no protective coatings were adopted in this study as their use would confound the characterization of the naked material.To comprehensively assess the changes between the flight samples and ground controls, we employed state-of-the-art surface analysis techniques.We examined morphological alterations using optical microscopy, scanning electron microscopy (SEM), and 3D surface profilometry.Chemical changes in terms of elemental composition and chemical bond prevalence were analyzed using energy-dispersive Xray spectroscopy (EDX), Fourier-transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), and time-offlight secondary ion mass spectrometry (ToF-SIMS).Furthermore, the overall composition changes in the matrix were evaluated using differential scanning calorimetry (DSC).Finally, the structural integrity of the materials was tested by assessing their flexural strength through a three-point bending test and a compression indentation test.
This CFRP material investigation provides valuable new insights into 3D printed CFRPs, as well as forged short-fiber and compression molded long-fiber CFRP materials for a broad spectrum of terrestrial and aerospace applications stretching past low Earth orbit (LEO) into deep space exploration.

LEO Exposure Experiment
Five CFRP samples were launched aboard the Cygnus NG-12 mission and successfully returned to Earth on the Dragon SpX-21 spacecraft.Table 1 shows a list of these five composite samples with their manufacturing process, constituent materials, purpose and the effect on composite property.These samples were strategically positioned outside the International Space Station (ISS) (Figure 1a) on the Japanese Experimental Module (JEM) (Figure 1b).To secure them during the mission, the samples were affixed to the space development acceleration capability (SDAC) flight test unit and integrated into the Nanoracks External Platform (NREP), depicted in Figure 1c.The deployment outside the ISS endured precisely 8.63 months (259 days), subjecting the samples to an estimated four thousand temperature cycles.These cycles ranged from temperatures of up to 121 °C when exposed to direct sunlight, to lows of −157 °C when in complete shade.In addition to these extreme temperature fluctuations, materials in low Earth orbit (LEO), located between 200-1000 km above the Earth's surface, face a barrage of harsh environmental conditions.These include high vacuum, solar UV radiation, galactic cosmic ray charged particle (ionizing) radiation, plasma, surface charging and arcing, impacts from micrometeoroids and orbital debris (MMOD), and environment-induced contamination.Of particular concern is the presence of atomic oxygen (AO), which forms in the LEO environment through the photodissociation of diatomic oxygen (O 2 ) and poses a significant threat to the structural integrity of exposed materials.When AO collides with a spacecraft surface due to the orbital velocity and thermal velocity of the atoms, polymers can experience hydrogen abstraction, oxygen addition, or oxygen insertion, which can compromise their integrity. [7]The exposure facility housing the samples was positioned to face the wake direction, resulting in reduced AO exposure compared to a ram-oriented facility. [8]However, the facility still experienced moderate solar radiation similar to the ram orientation, as well as typical omnidirectional charged particle and cosmic radiation. [9]

Optical and SEM Imaging
Figure 1 shows a comprehensive analysis of the samples after their exposure to Low Earth Orbit (LEO).The first row of images presents optical views of the entire sample, while the second row offers a close-up perspective of one corner, utilizing a stereo-microscope.Notably, all samples display evident discoloration.The samples were held in place using a metallic holder that shielded the edges from UV radiation.Solar UV radiation, with wavelengths typically ranging from 0.1 to 0.4 μm, possesses enough energy to cause polymer damage through cross-linking (hardening) or chain scission (weakening).7a] Furthermore, under high vacuum conditions, UV radiation can generate oxygen vacancies in oxides, resulting in pronounced color changes, often manifesting as yellowing or general darkening.The third row of images in Figure 1d showcases SEM images at the interface between the exposed and non-exposed surfaces of the samples.In certain samples (D1 and D2), a distinct difference in brightness between the two areas is evident.It is important to note that an increase in brightness in SEM images is commonly associated with a higher atomic number, but in this case, it is likely attributed to changes in surface conductivity of the CFRP.This change in brightness may be indicative of carbon fiber surfacing due to degradation of the epoxy matrix.However, for the remaining samples, the interface between the two areas could not be clearly identified using SEM.
6a] However, the potential failure mechanism of CFRP in LEO lies in microcracking generated from the degradation of the fiber/matrix interface.Such microcracking can propagate due to combined environmental factors such as thermal cycling and mechanical stresses.While microcracking was not observed in our investigation, the observed color and conductivity changes on the surface serve as indicators of physical/chemical alterations, warranting further investigation.

Profilometry
To investigate potential material erosion resulting from atomic oxygen (AO) exposure, we employed a high-resolution 3D profilometer, specifically a Keyence optical and laser scanning profilometer.Figure 2a-e presents an optical/laser hybrid image of the samples, with a vertical dashed line approximating the interface between the non-exposed and exposed areas.Figures 2f-j display a color depth map of the samples, while Figure 2k-o provide a 3D depth reconstruction.
To evaluate surface erosion, we calculated the average step height between the exposed and non-exposed areas of the samples (Figure 1p).This involved averaging the z-axis values within each region of interest and subsequently comparing them.For both D1 and D2, we observed a decrease in height, specifically −133 and −835 nm, respectively.Conversely, C1 exhibited a step height increase of 270 nm in the exposed area.It is worth noting that these measured step height differences are comparable to the surface roughness of the samples and therefore may not necessarily indicate significant erosion.
Figure 2q presents the calculated average peak curvature across the two regions of interest: non-exposed versus exposed.In the case of sample C1, a significantly higher peak curvature was observed in the exposed area, indicating the presence of sharper peaks.This finding aligns with the theory that AO exposure leads to oxidation and erosion of the polymeric matrix, resulting in a non-uniform surface with more pronounced carbon fiber -patterns.This can be observed in Figure 2c,h, where the carbon fiber pattern in the exposed region appears more distinct.Surprisingly, D1 and D2 exhibited the opposite effect, with the exposed region displaying a lower mean peak curvature, suggesting a smoother surface.
Additionally, we calculated the developed interfacial area (Sdr) for the molded samples (Figure 2r).Similarly, both D1 and D2 exhibited lower Sdr in the exposed region compared to the non-exposed region, while C1 showed higher Sdr in the exposed region.The behavior of C1 differed from the other two molded CFRP samples, indicating that the observed topographical surface changes are material-dependent and result in either a smoother or rougher surface.It is noteworthy that, although changes in peak curvature and developed surface were observed, a clear increase in surface roughness for all exposed surfaces, as expected, was not observed.These results somewhat contrast with previous literature findings. [10]Typically, surfaces of materials that produce volatile oxidation products, such as hydrocarbon polymers, gradually develop cones pointing in the direction of incoming atomic oxygen, leading to an increase in surface roughness over time.This roughness increase follows the square root of the amount of atomic oxygen exposure. [11]6a] It should be noted that for the 3D printed samples, due to their high intrinsic surface roughness from manufacturing process, surface parameters such as those observed in the molded samples would not provide meaningful information and are thus not reported.Overall, while profilometry offers nanometric depth resolution, the inherent roughness of the samples prevents us from definitively confirming material erosion.

Energy-Dispersive X-ray Spectroscopy
To assess possible changes in the elemental chemistry of the exposed surface, we conducted Energy-dispersive X-ray spectroscopy (EDX) analysis (Figure 3a-d).In all exposed samples, we observed the presence of silicon contamination, which has been previously documented in samples that have orbited in LEO for extended periods of time. [8]This contamination arises from silicone materials that were not adequately vacuum-baked and contain volatile short-chain molecules that can easily migrate onto adjacent surfaces.When these contaminated spacecraft surfaces are exposed to AO in LEO, the silicones undergo oxidation and form silica (or silicates).The presence of silica creates a cloud in LEO, and over time, the contamination can gradually accumulate, leading to deposits that have been observed to be several microns thick, as seen in the case of a Mir solar array after 10 years in space. [8,12]  Additionally, we observed an increase in oxygen concentration (Figure 3d) in all samples except for D1, and decrease in carbon concentration (Figure 3b).When the highly reactive AO impinges on the CF composite surface with sufficient collision energy, the C─C bond can react with AO, resulting in the production of volatile gases such as CO, CO 2 , H 2 O, and NO, or the formation of carbon-oxygen bonds, leading to surface carbon leaving as gas and additional oxygen being introduced into the composite surface.Some of these chemical changes on the surface were expected and are commonly associated with AO and UV exposure of polymers. [8]Further, the variation in nitrogen concentration on the non-exposed and exposed surfaces may be related to the amine and amide groups available on each surface.

FTIR
To gain further insight into the chemical changes occurring on the material surface, we conducted attenuated total reflection Fourier-transform infrared spectroscopy (ATR-FTIR).Each peak in the spectrum corresponds to a specific vibrational mode of a chemical bond in the structure of the material.Overall, we observed a consistent decrease in peak sharpness for the LEOexposed samples (Figure 3e-i).This phenomenon is commonly associated with chain scission and overall material degradation.However, it is important to note that despite the decrease in peak sharpness, all the characteristic peaks of the samples were still present, indicating that the chemical composition of the samples did not undergo drastic changes.
Interestingly, only the 3D printed samples exhibited a new peak associated with silica (Figure 3h,i).We attribute this to their higher surface roughness, which is more likely to retain silica particles present in LEO.The presence of this silica peak suggests that the 3D printed samples experienced a greater level of contamination from LEO environment compared to the other samples and highlights the influence of surface roughness on the retention of LEO contaminants.

XPS
To further characterize the chemical changes of the material surface after exposure to LEO, we performed X-ray photoelectron spectroscopy (XPS) analysis.Figure 4a shows a representative XPS spectrum of the exposed area of a sample (D1), highlighting the main peaks of oxygen (O 2 ), carbon (C), and silicon (Si).The relative amounts of O 2 , C, and Si are shown in Figure 4b-d, respectively, comparing the ground control, non-exposed area of flown samples, and exposed area.All samples exhibited the presence of O 2 and C, while the flown samples also showed the presence of Si in both the non-exposed and exposed areas.Some samples also contained traces of nitrogen (N) and/or calcium (Ca).Interestingly, even the ground control of the molded samples (D1, D2, and C1) exhibited traces of Si (Figure 4d), which can be considered advantageous as previous studies have reported that Si-containing materials develop a protective surface layer of silica (SiO 2 ) when exposed to AO fluence. [13]In all samples, the exposed area had a higher percentage of O 2 and Si compared to the non-exposed area.For C1, we observed less pronounced differences in Si relative content between the exposed and unexposed samples, likely due to the presence of glass fiber already in the composite manufacturing.The increase in Si content in the exposed area is attributed to contamination from siloxanes, common pollutants in the LEO environment.After exposure to UV irradiation and AO fluence, the atomic concentration of carbon decreased, while the concentration of oxygen increased for all samples.The increase in oxygen content is consistent with previous findings, indicating the oxidation effects of LEO exposure.6a,14] Figure 4e presents a representative high-resolution C1s XPS spectrum (D1) after deconvolution, comparing the non-exposed and exposed areas of the sample.The prevalent carbon bonds were quantified and presented in detail in Figure 4f-h.In the exposed area, all samples exhibited broader peaks and higher percentages of peaks associated with carbon-oxygen bonding environments compared to the non-exposed area.This suggests that exposure to outer space resulted in oxidation or degradation of the carbon fiber surface.Notably, while the C─C/C─H and C─O bonds remained similar between the ground control and non-exposed areas, the exposed areas consistently exhibited higher percentages of C═O/O═C─O bonds compared to the ground control.It is interesting to observe that the AM2 sample showed no increase in C─O bonds, despite an increase in oxygen composition.This suggests that temperature fluctuations can induce oxidation and chemical changes in the samples, but the energetic effects of AO impingement are necessary to break the polymer and create new covalent C─O bonds.In AM2 3D printed composite, the use of glass fiber as a reinforcement material may also stabilize the C─O bonds due to a lower carbon content available for oxygenation.On the other hand, the AM1 3D printed and other compression molded composites contain the carbon fiber as main reinforcement material, making more carbon available for oxygenation.In summary, the XPS analysis revealed changes in the elemental composition and chemical bonding environments on the material surface after LEO exposure.The exposed areas exhibited increased oxygen and silicon concentrations, along with higher percentages of carbon-oxygen bonding environments.These findings support the occurrence of oxidation and degradation processes induced by the LEO environment, while also highlighting the influence of temperature fluctuations on surface oxidation.

TOF-SIMS
Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) was performed to provide spatial resolution of elements and contaminants on the surface.ToF-SIMS is inherently non quantitative, however qualitative comparisons can be made between the relative intensities of peaks across a sample or between different samples.
The ToF-SIMS spectra for almost all samples showed dominant peaks for Na, K, and PDMS (polydimethylsiloxane).Hydrocarbon (CxHy) and Ca peaks were also observed.The hydrocarbon and Ca peak images appeared fairly uniform in most cases.Figure 5a,b displays representative ToF-SIMS images of Na and K, respectively, for sample D1.The relative intensities of Na and K were significantly higher in the non-exposed regions compared to the exposed regions, indicating contamination from the gaskets used to hold the samples in place.It's worth noting that Na and K were not detected by XPS, which is not uncommon as they have high ionization efficiencies and show strong signals in SIMS even when present in low amounts (not detectable by XPS).The Si signal was generally higher in the exposed area compared to the non-exposed area (Figure 5c), consistent with the results obtained from EDX, XPS, and FTIR, indicating contamination from the LEO environment.The PDMS peak signals were similar between the exposed and non-exposed areas (Figure 5d).
To compare the ground control samples with the samples exposed in space, selected peak areas were integrated and plotted.Regions of interest were created for the exposed and non-exposed samples based on the previous analysis.Bar plots were prepared to show the normalized relative intensity of the selected peaks for each sample type (Figure 5e-h).It's important to note that since the data shown is only from one area of each sample, no conclusions can be drawn about the statistical significance of the data.Both Na (Figure 5e) and K (Figure 5f) exhibited the highest relative intensities in the non-exposed areas, indicating contamination from the gaskets, as observed before.As previously seen, Si showed the highest concentration in the exposed areas for all samples, except for C1.Interestingly, C1 exhibited the highest PDMS relative concentration in the ground control.Except for C1, the relative intensity of PDMS on the ground control samples was lower than that observed on the exposed and non-exposed sample areas, while the intensities between the non-exposed and exposed areas were similar.Figure 5i-l presents the relative intensities of the most prevalent hydrocarbon groups.Several hydrocarbons showed lower concentrations in the samples flown to LEO compared to the ground control samples.However, C1 displayed the opposite trend, with higher intensities of hydrocarbons in the exposed areas.
In summary, ToF-SIMS analysis provided insights into the elemental composition and contamination on the surface of the samples.The presence of Na, K, and PDMS was prominent, with contamination from the gaskets observed.Silicon contamination was consistent with the results obtained from other analytical techniques.The relative intensities of hydrocarbons varied between the samples, with some samples showing decreased intensities in the exposed areas, while C1 exhibited higher hydrocarbon intensities in the exposed areas.One reason for the unusual trends of C1 sample could be due to the presence of Diuron/DCMU accelerator in this composite that may be triggering the formation of additional hydrocarbon bonds in the exposed region.

Differential Scanning Calorimetry
To investigate the effect of temperature fluctuations in LEO on the polymeric/epoxy-based matrix of the CFRP samples and its potential impact on mechanical properties, differential scanning calorimetry (DSC) studies were conducted in the temperature range from 25 to 220 °C.The DSC analysis aimed to understand any changes in the glass transition temperature (T g ) and enthalpy (∆H) of the samples.Figure 6a presents a representative DSC curve for the flown sample AM2.Two heating cycles were performed, and the relevant parameters were calculated for each curve, including T g (Figure 6b), T g onset (Figure 6c), and ∆H (Figure 6d).The same parameters were calculated for the second heating cycle and compared between a ground control sample and a flown sample that was not directly exposed to UV radiation and AO fluence (Figure 6b-g).An exposed sample was not   used for this comparison because the UV and AO exposure during the experiment's duration was expected to primarily affect the surface of the sample, while large temperature fluctuations were more likely to impact the mechanical properties of the epoxy matrix throughout the sample.Due to the limited availability of exposed samples, a flown but non-exposed sample was chosen as the ideal candidate.No significant differences were observed for any of the samples or calculated parameters, except for the ∆H value in AM2.However, it should be noted that the limited number of available samples prevents statistical significance from being determined.The lower enthalpy value observed for the flown sample (Figure 6d,g) could be related to polymeric scission, as shorter polymeric chains require less energy to melt.However, it cannot be confirmed as a decrease in molecular weight is typically correlated with a decrease in T g , which was not observed in this case (Figure 6b,e). [15]Another possibility for lower enthalpy of flown sample could be due to the lack of major carbon fiber reinforcement in AM2 composite as compared to other materials.The specific cause of the observed enthalpy decrease remains uncertain.
Overall, the DSC studies did not reveal major differences in the glass transition temperature or enthalpy between the ground control and flown samples, except for a potential decrease in enthalpy in the AM2 sample.However, due to the limited number of samples, statistical significance could not be determined, and further investigations may be required to elucidate the underlying mechanisms behind the observed changes.

Three-Point Bending
To assess the mechanical changes in the exposed samples, their flexural strength was measured using a three-point bending setup.The test setup and a representative sample in position are illustrated in Figure 7a,b, respectively.A flexural strength curve for sample C1 is shown in Figure 7c as an example.The compression-molded prepreg samples exhibited an average flexural strength of 550 MPa, which is comparable to the average value for stainless steel.On the other hand, the 3D printed samples demonstrated an average flexural strength of 158 MPa, comparable to aluminum alloys.The 3D printed samples displayed slightly lower mechanical strength than the compression molded samples that can be ascribed to i) significantly higher surface roughness inherent to the 3D printing layer stacking; ii) physical van der Waals bonds in 3D printed thermoplastics, which are weaker than the chemical bonds obtained via cross-linking in thermoset epoxies.The flexural strengths obtained from the curve were compared between the exposed and ground control samples (Figure 7d), and no statistically significant difference was observed.One limitation of the structural 3-point bending test performed is that the size of the samples is relatively short compared to the standardized samples size for this test, due to limitation of performing experiments in LEO.Therefore, the flexural strength values obtained should be compared to the ones available in literature with some caution as they might correlate better with interlaminar shear strength than standard flexural strength.Nevertheless, in this experimental setup all samples had the same size thus their results could be compared within the groups.Tabulated values of both calculated flexural strength and flexural modulus are reported in Table S2 (Supporting Information).
Additionally, we performed indentation tests on both the surface and the cross section of the samples to assess whether the local mechanical properties of the samples may have been altered.Figure 7e shows the experimental setup where we used a 2 mm diameter ruby spherical probe to indent the materials surface.Force, displacement graph of the indentation tests are reported in Figure S3 (Supporting Information).Figure 7f,g shows the calculated surface hardness or the surface and cross-section of the materials, respectively.We found no statistically significant changes in surface hardness between the exposed samples compared to their respective ground controls.
These findings indicate that the tested CFRP samples experienced minor oxidation and contamination that was limited to the surface of the sample leaving the bulk of the sample unchanged and thus the samples did not exhibit any embrittlement or structural changes.To further stress this point, we provided high resolution cross-section images in Figure S1 (Supporting Information) showing no visible signs of degradation toward the bulk of the material.Therefore, they remain suitable for aerospace applications in LEO.The prepreg compounds have the potential to replace stainless steel structures at a fraction of the weight, while the 3D printed materials can be utilized for smaller replacement fixtures or smaller satellites.Overall, the flexural strength tests confirm the structural integrity of the CFRP samples after exposure to LEO, further supporting their viability for aerospace applications.

Effect of Composition and Fabrication Method on Materials Properties
Fatigue and wear resistance are essential for long-term durability of 3D printed and CFRP parts.For mechanical components, a high surface roughness is associated with irregularities that may lead to nucleation sites for cracks that could grow over time under repeated mechanical or thermal stress.An overall increase in surface roughness could be an indicator of materials erosion and wear that can lead to worsening of mechanical properties.Surface topography analysis comparing the exposed and nonexposed samples showed that for both 3D printed and compression molded CFRP there was no significant wear or deterioration.Further, surface chemical analysis revealed a thin outer layer of silica deposited on the surface due to LEO contamination [8,12] that can provide an additional protective layer to the materials.Moreover, the carbon fiber and glass reinforcement not only provide mechanical strength, but it prevents propagation of surface wear by inhibiting material embrittlement.
Radiation and thermal resistance of CFRP composite structures also determines their physical resilience on long space missions.The presence of nitrogen atoms in constituent bulk material of compression molded and 3D printed samples, as evident from EDX data, could prevent the UV degradation in CFRP composites. [16]Although FTIR showed some oxidation and carbon-chain scission on the exposed area of the samples, the impact on material mechanical performance is limited due to the prevalence of rigid C═O bonds highlighted by the XPS data.From a thermal perspective, our samples withstood the temperature conditions of low-earth orbit varying from −157 to 121 °C.The compression-molded thermoset CFRP have an intrinsic high T g (166-197 °C) due to the irreversible networks and heat-resistant reinforcement fibers ensuring high thermal stability in the LEO conditions.Interestingly, for the 3D printed thermoplastic CFRP composites we observed an increase of T g during the second heating cycle, which is directly correlated with the enthalpy decrease from the first to the second cycle.This phenomenon suggests a heat-based post-curing of the material that results in a higher T g and thus improved thermal stability.
In terms of mechanical properties, the base resin in epoxies enhanced performance due to the higher functionalization number of epoxide groups per molecule.Carbon/glass fibers and quartz/Kevlar fabrics anchored the matrix and reinforced the composite structures for D1, D2, and C1.Further, the aminehardeners and curing accelerator assisted with chemical crosslinking of epoxy resins leading to a higher flexural strength.For the 3D printed materials, the nylon thermoplastic provided stiffness and the carbon/glass fibers reinforced the composites AM1 and AM2.More importantly, the manufacturing process played an important role in determining the mechanical property.In fact, the compression molding process generates a dense packing of epoxy matrix and CF reinforcement, yielding high flexural strengths.Even though 3D printed materials present slightly lower flexural strengths, due to a sparser structure and subsequent layer stacking, they are still a valuable component in applications that require lower loads.Further improvement of mechanical properties of 3D printed materials could be achieved by combining the 3D printed architectures with irreversible crosslinked coatings via heat or chemical processes.

Simulated Versus in Space LEO Exposure
Ground-based facilities attempt to replicate some of the environmental conditions experienced in space, but complex calibrations and cautious interpretation of the results are required as historically, discrepancies have been found between simulated versus actual LEO exposure.Parameters that are often simulated are high vacuum, AO fluence, UV light exposure and temperature fluctuations.To predict long-term material properties the tests are often performed on a shorter timescale and accelerated employing under conditions more severe than those inherent to the intended operational environment. [17]For example, to accelerate UV exposure, the materials are exposed to shorter wavelength and elevated temperatures.Specifically for polymeric based materials, properties are highly dependent on temperature, thus long-term predictions are often performed using the timetemperature superposition principle (TTSP), where experiments are carried out at a higher temperature and within a shorter time. [18]owever, the Hubble MLI Failure Review Board highlighted that accelerated durability tests conducted on Earth can significantly underestimate the damage that occurs in space.A clear example are the fluorinated ethylene propylene (FEP) surfaces of Hubble Space Telescope (HST) components. [5]Even after subjecting these surfaces to highly accelerated electron and proton exposure equivalent to 100 years, the degradation observed was not as severe as what was actually seen in samples retrieved from the HST after only 9.7 years. [5]Dever et.al. [5b] found an evident synergistic effect of radiation exposure (electron, proton, and xray) and thermal cycling.However, irradiation and thermal cycling with comparable HST exposure conditions still did not produce the degradation observed in the FEP material retrieved from the sample in LEO.Although the synergistic effects of radiation and thermal cycling are evident the origin of the severe degradation observed in LEO samples is still unclear.A similar disparity was noted by Stambler et al., [19] who compared the atomic oxygen (AO) erosion values (E y ) of 39 polymers in a ground-based plasma asher with the E y values for the same polymers flown on the Materials International Space Station Experiment-2 (MISSE-2) Polymers Experiment conducted by Glenn Research Center. [20]he asher to in-space E y ratios varied from 1.0 to 37.1, indicating significant differences in erosion rates between the ground-based facility and the space environment. [20]hese discrepancies observed can be attributed to the complex and incompletely simulated nature of space environmental exposure in ground-based facilities.Further, the extreme differences in exposure rates in space and ground tests may also play a major role as slower in-space exposure rates may enable free radicals greater opportunity to degrade the polymers. [9]Another crucial factor is the influence of temperature fluctuations on polymer responsiveness to external stimuli, as the combined effect of various conditions may not simply be linearly superimposed, leading to synergistic effects. [21]Even in simulated test where radiation exposure and thermal cycling are employed, it is unclear whether their simultaneous or subsequent application may lead to differences.In these simulations, the rate of temperature sweeps may also play an important role as especially for polymeric materials, their mechanical characteristics are strictly related to temperature.Despite the increased cost and constraints of LEO exposure compared to ground simulation, conducting some of these experiments in space is still of paramount importance.

Conclusion
In conclusion, compared to the forged and conventional compression molded CFRPs, the uncoated 3D printed materials demonstrated remarkable durability and minimal degradation after more than 8 months of exposure in LEO.While minor surface chemical changes were observed primarily due to UV exposure, there were no major effects on surface erosion or detectable changes in mechanical strength.These findings highlight the potential of 3D printed CFRPs as high-performance materials for terrestrial and aerospace applications complex aerospace structures and components such as drones, satellites and other spacecrafts.The minimization of raw material, ease of manufacturing and fast turnaround time between conception and application makes 3D printed CFRPs highly advantageous.Of relevance, 3D printed CFRPs are compatible with in-space manufacturing, potentially reducing deployment costs and mitigating risks associated with payload launches.While further research is necessary to completely understand the full potential and limitations of these materials in the space setting, CFRPs are expected to gain further relevance into the aerospace industry.Translationally, our findings using the commercially available materials will acceler-ate the development of 3D printed durable structures for space deployment and could also be relevant to in situ manufacturing of space settlements.

Experimental Section
Carbon Fiber Reinforced Polymer Materials and Manufacturing: D1, D2, and C1 were manufactured using a particular kind of pre-preg compression molding (PCM) called Carbon Fiber Sheet Mold Compound (CF-SMC) where sheets of thermoset pre-impregnated materials (pre-preg) were stacked (laminated) together to obtain the required part dimensions.Then, rectangular plates with specific layup molds were created with defined values of temperature, pressure, and curing period that are material depending, typical values are 130-140 °C, 70-140 bar, 5-20 min.After a period of cooling down, the plates were cut with a waterjet machine to get the required sample size (27 mm × 27 mm).
AM1 and AM2 were manufactured using a Markforged 3D printer that used a fused filament fabrication technique capable of joining a matrix of thermoplastic filament with a reinforcement of a continuous single filament of different fibers.Both materials have an Onyx (Markforged) composite base made of carbon fiber-filled nylon.
LEO Exposure Experiment: The samples were attached onto the Space Development Acceleration Capability (SDAC) Flight Test Platform (FTP) developed by Craig Technologies.The SDAC FTP integrates into the NanoRacks External Platform (NREP) to be positioned on the outside of the international space station (ISS).The SDAC facility was flown to the ISS and the crew installed it on the NREP.The NREP passed through the Japanese Experiment Module (JEM) Airlock and was placed on the JEM Experimental Facility (EF) with the Japanese robotic arm.The NREP stayed outside of the ISS in Low Earth Orbit (LEO) environment on the JEM EF for 8 months with the samples facing wake direction.After completion of the experiment, the Japanese robotic arm removed the NREP from the JEM EF and placed it in the JEM Airlock.After passing through the JEM Airlock, the NREP was disassembled by the ISS crew and the SDAC facility was removed and stowed for return to Earth via the SpaceX Dragon capsule.
Scanning Electron Microscopy (SEM) and Energy-Dispersive X-ray (EDX) Spectroscopy: CFRP samples were imaged using FEI Nova NanoSEM 230 ultra-high resolution.All images were obtained with 5.0 kV acceleration voltage.For each sample, images at different magnifications in the range 100X to 500X were obtained.The interface between exposed and nonexposed areas were imaged, and took higher magnification of the nonexposed area, exposed area, and cross-section.EDX was performed in tandem with SEM imaging for both the exposed and non-exposed area.
Profilometry: Surface topography images were acquired using a Keyence VK-X3000 profilometer.The acquired area was centered across the exposed/non-exposed interface.The acquired images were analyzed using the software Keyence VK-X3000 MultiFileAnalyzer 3.3.1.For each sample the non-exposed and exposed area were separated and used to calculate differences in average step height (Figure 2p), peak curvature (Figure 2q) and developed interfacial area ratio (Figure 2r).Surface parameters were not calculated for the 3D printed samples (AM1 and AM2) due to the high intrinsic surface roughness.
FTIR: CFRP samples were analyzed via FTIR-ATR (NicoletTM 6700 and Smart iTRTM Attenuated Total Reflectance (ATR) sampling accessory; Thermo Fisher Scientific, Waltham, MA, USA).The samples were analyzed as is and not processed prior to the analysis.The obtained spectra are averages of 128 scans in the range of 600-4000 cm −1 anda resolution of 2 cm −1 .
XPS: All XPS spectra were taken on a Kratos Axis-Ultra DLD spectrometer.This instrument has a monochromatized Al K X-ray and a low energy electron flood gun for charge neutralization.X-ray spot size for these acquisitions was on the order of 700 × 300 μm.Pressure in the analytical chamber during spectral acquisition was less than 3 × 10 −8 Torr.Pass energy for survey spectra (composition) was 80 eV.Pass energy for the highresolution spectra was 20 eV.The take-off angle (the angle between the sample normal and the input axis of the energy analyzer) was 0°(≈100 Å sampling depth).The Kratos Vision2 software program was used to determine peak areas and to calculate the elemental compositions from peak areas.Casa XPS was used to peak fit the high-resolution spectra.For the high-resolution spectra, a Shirley background was used, and all binding energies were referenced to the C ls C─C bonds at 285.0 eV.
ToF-SIMS: ToF-SIMS spectra/images were acquired on a IONTOF ToF-SIMS 5 spectrometer using a 25 keV Bi 3 + cluster ion source in the pulsed mode.Spectra/Images were acquired for positive secondary ions over a mass range of m/z = 0 to 800.The ion source was operated with at a current of 0.22 pA.Positive secondary ions were extracted and detected using a reflectron time-of-flight mass analyzer.Spectra/images were acquired using the instruments stage raster function in order to cover both exposed and non-exposed areas.Positive ion data were calibrated using the CH 3 + , C 2 H 3 + , C 4 H 7 + and C 5 H 7 + peaks.Calibration errors were kept below 25 ppm.Mass resolution (m/Δm) for a typical spectrum was ≈4000 for m/z = 27 (pos).
The interfaces between the exposed and non-exposed areas on the MS4H, C1, and D1 samples were easy to find.For these samples a 1 cm × 1 cm stage raster was carried out using 500 μm × 500 μm tiled images.Each tile was acquired using 256 × 256 pixels and a total ion dose per tile of 3.6 × 10 11 ions cm −2 .For AM2 and AM1 samples a stage raster of 2.5 cm × 1 cm using 500 μm × 500 μm tiles was carried out from the center of the samples toward the exposed edge.Each tile was acquired using 256 × 256 pixels and a total ion dose per tile of 7.2 × 10 10 ions cm −2 .The interface between the exposed and non-exposed areas was identified using the Na + and Ca + images.
Differential Scanning Calorimetry (DSC): DSC was performed following the standard test method for transition temperatures and enthalpies of fusion and crystallization of polymers (ASTM D3418-15).A Perkin Elmer DSC8500 was used to perform the experiment, data was analyzed using the software Pyris from Perkin Elmer.The DSC cycle was performed under steady 30 mL min Three-Point Bending: The three-point bending test was performed on a custom-made setup due to the limited size of the samples comprising a 250 kN Instron 5985 Universal testing machine and a I3540-025M-ST deflectometer with a 25 mm travel.The samples were held on two support rollers with a 3 mm diameter, while the loading roller had a diameter of 6 mm.Test speed was set at 1 mm min −1 .The support span was set at 13.5 mm for D1 and AM1 samples and 15.5 mm for all other samples.To determine the maximum flexural strength of the samples "Procedure A" of the "ASTM D7264/D7264M-07″ standard was followed and used the following formula: where P max is the load measured during the test, L the support span (mm), b the sample width (mm), h the thickness of the sample.Flexural modulus was calculated using the following formula: where F is the applied force, and d is the deflection.Indentation Test: The indentation test was performed on a custommade setup featuring a 2 mm ruby spherical indenter from RENISHAW, implemented on the Univert mechanical testing platform by CellScale.The zero-point reference was established upon the initial contact of the indenter with the sample, with the force set to 0 N. Subsequently, a load up to 100 N was applied by moving the probe at a constant velocity of 5.556 μm sec −1 .
The surface hardness was determined from the slope of the forcedisplacement curve in the linear region: S = dP d (3) where P represents the force measured during the test, and  is the displacement of the indenter.The hardness value S is correlated with the material's Young's modulus through the formula: Where Er is the reduced modulus and A the tip-sample contact area.
Statistical Analysis: Data are represented as mean ± SD.Statistical significance was defined as two-tailed p <0.05 for all tests.All statistical analyses were performed with GraphPad Prism 9 (version 9.3.1;GraphPad Software, Inc., La Jolla, CA).

Table 1 .•
List of CFRP composite samples and their constituent materials along with the purpose and effects on composite property.Improved mechanical strength p-(2,3-epoxypropoxy)-N,N-bis(2,3epoxypropyl)aniline low-viscosity epoxy resin (tri-epoxide functional group)

Figure 1 .
Figure 1.Carbon fiber reinforced polymer samples.a) Picture of the International Space Station (ISS) highlighting the location of the samples.b) Japanese experiment module (JEM) with the location of the Nanoracks External Platform (NREP).c) NREP with a highlight on the samples exposed to LEO. d) Images of the CFRP samples upon return to earth: first row, optical pictures obtained with a digital camera; second row, optical picture obtained with stereo microscope; third row, SEM images at the interface between the exposed (left) and unexposed area (right).

Figure 2 .
Figure 2. Surface topography.Surface topography of CFRP at the interface between the non-exposed and the exposed regions.a-e) Images show laser depth field, f-j) color map depth field and k-o) 3D depth reconstruction.a-o) Vertical dashed lines represent approximate separation between nonexposed and exposed areas.Surface topography descriptors for the prepreg compression molded samples: p) step height (exposed -non-exposed); q) peak curvature; r) developed interfacial area.( *** = p <0.001).

Figure 3 .
Figure 3. a) EDX and FTIR.Elemental composition of CFRP samples obtained via SEM-EDX b) with details on relative percentage of Carbon, c) Nitrogen and d) Oxygen.e-i) FTIR spectrum of CFRP samples comparing ground controls, non-exposed and exposed samples.h-i) Insets show magnification of spectrum ≈1000 cm −1 and highlight the peak indicating presence of silicon.

Figure 4 .
Figure 4. a,b) XPS.Representative XPS spectra for CFRP sample D1 showing the deconvolution in C─C/C─H, C─O, C═O and O═C─O bonds.b) Elemental analysis of surface composition showing oxygen, c) carbon and d) silicon.e-h) Details and comparison of common bonds among all CFRP samples.

Figure 5 .
Figure 5. ToF-SIMS.Representative images of CFRP D1 sample showing the signal intensity (black to yellow) of different elements on the surface: sodium (Na, a), potassium (K, b), silicon (Si, c), polydimethylsiloxane (PDMS, d).Vertical dashed lines (a-d) represent approximate location of separation between exposed and non-exposed areas.Relative intensity of elemental peaks among the ground control, non-exposed and exposed samples: sodium (Na, e), potassium (K, f), silicon (Si, g), polydimethylsiloxane (PDMS, h).Relative intensity of hydrocarbons in the different groups (i-l).

Figure 7 .
Figure 7. Mechanical testing.Three-point bending experimental setup (a) with close-up view at the sample under test (b).c) Representative flexural strength curve (C1).d) Flexural strength comparison between exposed/non-exposed samples for each CFRP.e) Experimental setup of indentation testing.f) Surface harness measured on the surface of the sample and g) surface harness measured on the cross-section of the sample.(N = 3).