Can quasicrystals survive in planetary collisions?

We investigated the compressional behavior of i-AlCuFe quasicrystal using diamond anvil cell under quasi-hydrostatic conditions by in situ angle-dispersive X-ray powder diffraction measurements (in both compression and decompression) up to 76 GPa at ambient temperature using neon as pressure medium. These data were compared with those collected up to 104 GPa using KCl as pressure medium available in literature. In general, both sets of data indicate that individual d-spacing shows a continuous decrease with pressure with no drastic changes associated to structural phase transformations or amorphization. The d/d0, where d0 is the d-spacing at ambient pressure, showed a general isotropic compression behavior. The zero-pressure bulk modulus and its pressure derivative were calculated fitting the volume data to both the Murnaghan- and Birch-Murnaghan equation of state models. Results from this study extend our knowledge on the stability of icosahedrite at very high pressure and reinforce the evidence that natural quasicrystals formed during a shock event in asteroidal collisions and survived for eons in the history of the Solar System.


Introduction
Textural and chemical studies of shocked veins and mineral assemblages observed in meteorites over the last decades have been widely interpreted in light of dynamic (e.g., shock experiments) and static (e.g., multi anvil and diamond anvil cell experiments) compression investigations of diverse minerals and their stability at high (H) pressure (P) and temperature (T) (Tomioka and Miyahara 2017). Some of the early studies were pioneer in order to link the coexistence of various minerals with shockimpact events that have led to the formation of HP polymorphs in chondritic meteorites either by solid-solid reaction (Gillet et al. 2000;Beck et al. 2004;El Goresy et al. 2008, 2010Bindi et al. 2017Bindi et al. , 2020 or by crystallization of melts in shock veins (Miyahara et al. 2008(Miyahara et al. , 2009). In some cases, static experiments were preferred to shock experiments to account for collisions of large asteroidal bodies or the passage of multiple shock waves through such bodies during multi-stage collisional events (Chen et al. 1996).
Among the several studies reporting the observation of high-pressure minerals in meteorites (e.g., El Goresy et al. , 2001aGoresy et al. , b, 2008Goresy et al. , 2010Tomioka and Miyahara 2017;Bindi et al. 2017Bindi et al. , 2020Tomioka et al. 2021), the discovery of metallic alloys has been of particular interest as these provide (1) direct information on differently O 2 -depriveted portions of the Solar Nebula where alloys condensated, (2) a snapshot of the cooling history of the pristine chondritic material, and (3) a valid comparative model of core formation and chemical composition (McDonough and Sun 1995). Due to the different chemical compositions of the observed metallic particles found in meteorites, their relation with the bulk chemistry of the hosting meteorite along with textural observations has required a different interpretation. For instance, El Chao (1976, 1977) reported the discovery of Fe-Ni-Cr particles in the basement rocks of Ries crater (Germany) that were attributed to a condensation process after vaporization of the impacting (carbonaceous) stony meteorite. This finding of metal particles was anticipated by the description of Co-, Cu-, and Ni-rich spheroids within impactites from the Barringer Meteorite Crater in Arizona by Kelly et al. (1974) explained on the basis of extensive chemical reactions with the impacted target as well as Fe removal by oxidation and vaporization processes when meteorites pass through the terrestrial atmosphere. Wlotzka (1976) andEl Goresy et al. (1978) reported the finding of a metal particle from the Ca, Al-rich inclusion of the Allende and Leoville meteorites markedly enriched in platinum group elements (PGE) and explained as result of condensation process from the Solar Nebula when not representative of pre-solar grains (i.e., Fremdlinge). Several studies have focused on the different Fe/Ni ratios of kamacite and taenite from stony meteorites referred to the different cooling histories (Reed 1964). More recently, the finding of metallic Fe coexisting with Fe-rich bridgmanite (hiroseite) in a shock vein of Suizhou meteorite  has been reported as the first direct evidence in nature of redox-driven disproportionation reactions (Frost et al. 2004). Such discovery highlights the fact that (HP) mineral assemblages found in shocked meteorites can either have inherited the redox conditions of the portion of Solar Nebula where chondritic material formed (Stagno and Aulbach 2021) or have reached equilibrium within 10 −1 -10 0 seconds during the impact event (Ohtani et al. 2007;Beck et al. 2005;Xie and Sharp 2007) whose P, T, and chemical composition are influenced by the locally buffered oxygen fugacity (fo 2 ). On the other hand, studies conducted through molecular dynamic simulations have established a relation between the heliocentric distance and the redox conditions of the accreting bodies with the latter being more oxidized beyond 1.1-1.7 AU (Rubie et al. 2015).
Over this last decade, fundamental questions have been raised by the finding of metallic inclusions in the Khatyrka meteorite, a CV3 carbonaceous chondrite discovered in the Koryak mountains of Chukotka in Eastern Russia (Bindi et al. 2012;MacPherson et al. 2013). These particles were found to be Al-Cu(Ni)-Fe alloys with a quasiperiodic atomic arrangement, materials well studied for the last 35 years in the laboratory but exceedingly rare in nature (Shechtman et al. 1984;Levine and Steinhardt 1984;Bindi et al. 2009). The natural quasicrystals were then named icosahedrite (Al 63 Cu 24 Fe 13 ) and decagonite (Al 71 Ni 24 Fe 5 ; Bindi et al. 2011Bindi et al. , 2015b, and recently, a first model of formation has been reported on the basis of their trace element contents (Tommasini et al. 2021).
The presence of stishovite, ahrensite, and a spinelloid phase reported in Khatyrka fragments (Hollister et al. 2014) as well as the presence of (redox) reaction rims (Lin et al. 2017) in the proximity of these particles were interpreted as evidence that the meteorite formed as result of a multi-stage process that started with the (1) condensation of Al-Cu-Fe from reduced portions of the Solar Nebula about 4.56 Gya and (2) experienced shock metamorphism induced by a hypervelocity impact with pressures exceeding at least 10 GPa and temperatures up to 1500°C about < 600 Ma before (3) being ejected from its parent (K-type) asteroid 2-3 Ma ago (Meier et al. 2018). Moreover, the recent discovery of a COtype chondritic spherule from the Nubian Desert, Sudan, containing the same assemblage of aluminum, iron, and copper and with a morphology remarkably similar to Khatyrka, provided further support and independent evidence that these samples were formed in outer space (Suttle et al. 2019). Figure 1 is a cartoon summarizing the state of art for the metallic inclusions in Khatyrka meteorite. Several experimental studies were conducted at the aim both to reproduce shock-impact events to verify whether extraterrestrial quasicrystals would form and to test the stability of Al-Cu-Fe quasicrystal as function of pressure and temperature (see Table 1).
The first set of dynamic compression experiments successfully resulted in the formation of both icosahedral Al-Cu-Fe QC (Asimow et al. 2016;Oppenheim et al. 2017a;Hu et al. 2020) and decagonal QC (Oppenheim et al. 2017b) by performing shock experiments.
Studies aimed to test the stability of icosahedrite were carried out by Stagno et al. (2014Stagno et al. ( , 2015Stagno et al. ( , 2017 using both in situ multi anvil (MA) and diamond anvil cell (DAC) techniques assisted by synchrotron radiation X-ray diffraction (XRD) at room/high temperature. The results from these studies showed that (1) synthetic icosahedrite remains stable without any transformation to approximant phases over a wide P range from 3 to 42 GPa and T up to 1600°C, (2) it exhibits a negligible effect of pressure on the volumetric thermal expansion properties, and (3) the quasiperiodic order of icosahedrite is retained up to 51.3 GPa. Importantly, it has been shown that pressure acts to stabilize the icosahedral symmetry at temperatures much higher than previously reported and that direct solidification of AlCuFe QCs from an unusually Al-Cu-rich melt is possible but it is limited to a narrow temperature range (Stagno et al. 2017). However, the possibility that icosahedrite could form in nature and survived at elevated pressures and temperature conditions cannot be excluded as this relies on the experimental investigation of its structural behavior and potential decomposition products at high pressures.
As motivation of this study, there is evidence that heavily shocked ordinary chondrites, classified into shock stage S6, have experienced pressures up to~75 GPa as confirmed by the presence of high-pressure minerals (Stöffler et al. 2018). Hence, the extension of pressure ranges up tõ 70 GPa and beyond is important to understand the icosahedrite stability in shocked meteorites.
Therefore, we report in situ angle-dispersive X-ray powder diffraction (XRD) measurements on synthetic icosahedrite to extend the investigation of the interplanar distances at pressures up to about 75 GPa at ambient temperature. Our results are integrated with data by Takagi et al. (2015) on the compression behavior of icosahedral Al 65 Cu 22 Fe 13 to better understand the possible evolution of icosahedrite at megabar pressures adopting a univocal compression model.

Materials and methods
The synthetic quasicrystal used as starting material in this study is the same as that used by Stagno et al. (2015), and it was previously characterized by SEM and XRD measurements and showed to have the formula Al 63 Cu 24 Fe 13 (Bancel 1999) plus minor amounts of cupalite, (Cu, Fe)Al. The advantage of using this starting material is that its composition can be considered the analog of icosahedrite (Bindi et al. 2011). After extracting a small fragment from the synthetic single crystal, a small chip was crushed to form a platelet. A small platelet with a diameter of approximately 60 μm was loaded into the sample chamber of a symmetric diamond anvil cell with flat anvils of 300-μm size culet and rhenium gasket as sample chamber with a 150-μm diameter hole. In situ angle-dispersive X-ray diffraction (ADXRD) measurements were performed at high pressure at the 13ID- D beamline (GSECARS) of the Advanced Photon Source in Argonne (Illinois). Neon was loaded as the pressure medium, and the DAC was mounted on a motor-driven stage with the tungsten carbide (WC) seat on the downstream side and the cubic boron nitride (c-BN) seat on the upstream side. A focused monochromatic 30-keV Xray beam with a wavelength of 0.4133 Å and a sampleto-detector distance of~296 mm was used. Diffraction patterns were collected on a MarCCD-165 detector with exposure time of 15 s. Pressure was measured using the thermal equation of state (EoS) of Ne used as pressure medium (Hemley et al. 1989). The collected data were processed using FIT2D software (Hammersley 1998), and the d-spacing relative to each reflection was accurately determined using PeakFit software.

Results and discussion
The collected most representative diffraction data of synthetic icosahedrite are shown in Fig. 2 as function of pressure, while Table 2 shows the d-spacing at different pressures using the two-integer indices (Janot 1994;Lu et al. 2001) along with the estimated pressures from the equation of state of Ne. The peaks were indexed according to Bindi et al. (2011), and the d-spacing was   Mao et al. (1990) EDXD stays for energy-dispersive X-ray diffraction, while ADXD stays for angle-dispersive X-ray diffraction Stagno et al. Progress in Earth and Planetary Science (2021) 8:27 determined by fitting each diffraction peak using a Gaussian formula within an uncertainty of less than 0.01 Å. The effect of pressure on the d-spacing is shown in Fig. 3 with P determined by the EoS of Ne. According to our results, seven out of eight interplanar distances show a similar linear compression behavior, unlike (8,4) that appears more compressible over the P range of investigation. This is more evident in Fig. 4 where the pressure dependence of the lattice parameter a 6D is plotted up to the maximum pressure of~76 GPa defined as, where d is the d-spacing in Å, N and M are the Cahn indices for which the d-spacing is experimentally determined (i.e., (8,4) in this case), and τ is the golden ratio, (1+√5)/2 (Steurer and Deloudi 2009). The sixdimensional lattice parameter, with respect to the zero-P parameter by Bindi et al. (2011), is shown to gradually decrease with increasing pressure (Fig. 4) well in agreement with data by Stagno et al. (2015) providing, therefore, a significant evidence of reproducibility of the obtained results that reflects also the accuracy of the performed experiments in both compression and decompression. The estimated reduction of the lattice parameter from the ambient pressure value of 12.64 Å to 75.8 GPa is about 10.4% implying a constant compressibility behavior as previously observed up to 50 GPa. No clear evidences of either structural phase transitions or amorphization were observed within the investigated pressure range as it can be also seen in Fig. 2. Figure 4 also shows data by Takagi et al. (2015) collected by in situ ADXRD but on synthetic Al 65 Cu 22 Fe 13 . In their study, these authors reported no substantial changes in the XRD profile relative to five characteristic peaks up to 104 GPa. These are the (24,36), (28,44), (32,48), (72,116), and (80,128) for which the least anisotropy was also observed in our measurements in case of the synthetic analog of icosahedrite, Al 63 Cu 24 Fe 13 . No information was derived for the most anisotropic (8,4) reflection, which is likely the reason of the little difference with respect to the study by Stagno et al. (2015).
The data were fitted to both Murnaghan and Birch-Murnaghan EoS (Angel et al. 2014). The former model allows a direct comparison with the previous experimental data from literature (see Table 1) and results in a bulk modulus of 121(8) GPa ( B 0 0 of 3.68 ± 4), which is close with the 113.7 GPa (B 0 0 of 4.22) obtained for the same QC compressed to 50 GPa (Stagno et al. 2015). As a consequence of the observed consistency between the present data and those by Stagno et al. (2015), we can fit successfully all of them to the same EoS and obtain a B 0 of 120(2) GPa and B  Sadoc et al. (1994), 155 GPa (B 0 0 of 2.0) by Lefebvre et al. (1995), and close to 120 GPa (B 0 0 of 5) for decagonal Al 72 Ni 20 Co 8 (Hasegawa et al. 1999). A fit of our data to the third-order Birch-Murnaghan EoS produces little negligible differences (B 0 116 ± 3, B 0 0 of 4.39 ± 17). Noteworthy, we refer to the decagonal phase since, at present, this is the only available data on the physical properties of decagonal QC with composition very close to that found in nature, Al 71 Ni 24 Fe 5 , for which a HP origin has been proposed (Bindi et al. 2015a, b).
Interestingly, Takagi et al. (2015) noted a slightly volume shift between 72 and 75 GPa beyond which they claimed that the QC would transform to an approximant phase. This consideration is supported, in their opinion, by the appearance of small unknown peaks at 2θ of~8.5°and 10.6°for (λ = 0.41441 Å). Therefore, their data were fitted to 72 GPa using a third-order Birch-Murnaghan EoS to give a bulk modulus of 131(7) GPa (B 0 0 of 4 as fixed value). A second EoS fit for volume-P data collected from 74 to 104 GPa provided a higher bulk modulus like 170(40) GPa (B 0 0 of 4 as fixed value). Such difference was proposed to be caused by the different compressional behavior of i-QC and an approximant phase, although no sufficient evidence exists that such phase transformation could occur. For instance, it cannot be excluded that the peaks reported as unknown could be representative of weak reflections of the i-QC that appears visible as result of preferred orientation upon compression (similar peaks also were observed within a lower P range for similar compositions by Lefebvre et al. 1995). The unknown peak that Takagi et al. (2015) noted at 10.6°=~2.24 Å would correspond to the reflection (56,84) of the i-QC (Bindi et al. 2009(Bindi et al. , 2011. Further, it is not surprising that the compressibility reported by Stagno et al. (2015) is slightly higher than that obtained by Takagi et al. (2015) as the latter determined the EoS from the least compressible d-spacings of i-QC; on the contrary, Stagno et al. (2015) performed a highresolution investigation of the icosahedral structure by monitoring the behavior in the high d-spacing region that included peaks (12,16) at 5.53 Å and (8,4) at 8.94 Å. The d-spacing referred to this latter reflection manifested a more evident anisotropic compressional behavior resulting, therefore, in a lower bulk modulus. On the other hand, first-order phase transitions would appear visible through the variation of the d-spacing as function of P. Figure 3 shows a continuous decrease of d-spacing as P increases up to 75.8 GPa. Therefore, we fit the data by Stagno et al. (2015) up to 72 GPa along with those by Takagi et al. (2015) up to 104 GPa with only a unique EoS. The obtained bulk modulus of 121(3) GPa ( B 0 0 of 3.87 ± 12) using a Murnaghan EoS fitting model is consistent with the EoS parameters used to fit our data up to 75 GPa. Fitted to a third-order Birch-Murnaghan EoS, we obtain a bulk modulus of 117(3) GPa ( B 0 0 of 4.45 ± 15), which is again consistent with the previous fit by Stagno et al. (2015). Both fitting models are showed in Fig. 5 with an imperceptible difference between them. Data from literature are also shown by comparison. In particular, the compressibility of icosahedral Al 62 Cu 25.5 Fe 12.5 determined by Sadoc et al. (1994) and Lefebvre et al. (1995) has been extrapolated up to 104 GPa assuming that no phase transition occurs for this composition by analogy with those investigated here. Both curves are quite consistent with the experimental measurements although a little shift is likely due to, again, the fact that the lattice parameter was determined by the less compressible d-spacing. In addition, loss of hydrostaticity could have occurred Fig. 4 Variation of the lattice parameter as function of pressure compared with similar studies on Al 63 Cu 24 Fe 13 (Stagno et al. 2015) and Al 65 Cu 22 Fe 13 (Takagi et al. 2015). The zero-P volume is from Bindi et al. (2011) Stagno et al. Progress in Earth andPlanetary Science (2021) 8:27 because of the use of silicon oil as pressure medium. Importantly, the extrapolated V/V 0 is shown also for the rhombohedral and pentagonal approximant phases by Lefebvre et al. (1995) where periodic approximant refers to a crystalline solid with similar chemical composition to a quasicrystal, but whose atomic arrangement is slightly distorted so that the symmetry conforms to the conventional laws of three-dimensional crystallography. These phases, although described to have a slightly different chemical composition, resulted to have the same physical behavior at high P reflected by the same EoS parameters. Interestingly, these approximant phases are shown to undergo a volumetric reduction under compression that is apparently lower than the QC phase. Therefore, we are confident that no phase transition occurred at P > 72 GPa within the Al-Cu-Fe system investigated to date. This evidence raises important questions on the stability also of approximant phases on extraterrestrial bodies that experienced shock-impact events. The recent findings of a quasicrystal with 10-fold symmetry, decagonite, with composition Al 71 Ni 24 Fe 5 (Bindi et al. 2015a, b) and its approximant, proxidecagonite , in the same Khatyrka meteorite suggest the possibility that P-T conditions might play a key role in determining the most energetically stable atomic configuration. To date, however, no approximant to icosahedrite has been found that support the possibility that this phase transition might require extremely high P-T conditions never explored so far in experiments.
At the present, both dynamic and static compression experiments have revealed the structural stability of i-QC. Both icosahedral Al-Cu-Fe quasicrystals (Asimow et al. 2016;Oppenheim et al. 2017a;Hu et al. 2020) and decagonal quasicrystals (Oppenheim et al. 2017b) were reported as run products of shock experiments during which shock pressures up to 35 GPa were generated with T as high as 400°C within less than 1 μs. These experiments succeeded in reproducing both the textural contacts between the quasicrystalline phases and the Al-Cu-Fe alloys as well as their chemical composition as in the Khatyrka meteorite. More importantly, the finding of quasicrystals incorporating additional elements like Cr and Ni to the known Al-Cu-Fe, as well as the variability of the Al/ Cu and Cu/Fe atomic ratios, represent an important evidence that quasicrystals do form during impact and keep their structural stability under more complex chemical environments than previously thought.
Static experiments carried out in the diamond anvil cell, in particular, have extended our knowledge of the behavior of synthetic icosahedrite up to megabar pressures. The effect of temperature has been investigated through either laser heating in DAC or multi anvil experiments. While the former helped to verify the structural stability of the i-QC during heating up to~1850°C and upon cooling/quench, experiments performed with the large volume press helped to shed light on the kinetics versus thermodynamic stability of the QC. Stagno et al. (2017) showed experimentally that  Sadoc et al. (1994) and Lefebvre et al. (1995) both for icosahedral Al 62 Cu 25.5 Fe 12.5 and for one approximant phase with composition Al 64 Cu 24 Fe 12 (Lefebvre et al. 1995). The curves obtained from these previous studies are extrapolated up to 104 GPa assuming that no structural phase transformation occurs Stagno et al. Progress in Earth and Planetary Science (2021)  icosahedrite can be stable for hours under hydrostatic compression at HT in the presence of additional minerals like cupalite, khatyrkite, and stolperite, similarly to what are found in nature. The experimental observation of the coexistence of synthetic icosahedrite along with an Al-Cu-Fe melt (plus additional phases) proves that icosahedrite can survive at those high P-T conditions at which shock glassy veins containing high-P polymorphs were proposed to form. We are confident that important steps forward have been done through experimental studies that demonstrate both the kinetics and thermodynamic stability of Al-Cu-Fe QCs, expanding, therefore, their potential occurrence through space and time.

Conclusions
We presented the results of experiments conducted at P up to~76 GPa using the DAC technique that confirms the stability of icosahedrite at high P. Our results appear consistent with previous data collected up to 50 GPa and show a linear decrease of the d-spacing with the exception for the (8,4) reflection, as never observed to the present, for which the calculated volume reduction is about 10% with respect to the V 0 . Our data were, therefore, integrated with those previously published by Takagi et al. (2015), who claimed for the transition of the i-QC to an approximant phase at 72-75 GPa. Based on a re-fit of these data along with those by Stagno et al. (2015) to a unique EoS fit model, we propose that icosahedrite can retain its structure up to megabar pressures. This extended P range, further accompanied by the stability of i-QC at elevated T representative of shock impacts, rules out the possibility to find Al-Cu-Fe approximant phases in the Khatyrka meteorite.