Twin Induced Reduction of Seismic Anisotropy in Lawsonite Blueschist

: Lawsonite is an important mineral for understanding seismic anisotropy in subducting oceanic crust due to its large elastic anisotropy and prevalence in cold subduction zones. However, there is insufﬁcient knowledge of how lawsonite twinning affects seismic anisotropy, despite previous studies demonstrating the presence of twins in lawsonite. This study investigated the effect of lawsonite twinning on the crystal preferred orientation (CPO), CPO strength, and seismic anisotropy using lawsonite blueschists from Alpine Corsica (France) and the Sivrihisar Massif (Turkey). The CPOs of the minerals are measured with an electron backscatter diffraction instrument attached to a scanning electron microscope. The electron backscatter diffraction analyses of lawsonite reveal that the {110} twin in lawsonite is developed, the [001] axes are strongly aligned subnormal to the foliation, and both the [100] and [010] axes are aligned subparallel to the foliation. It is concluded that the existence of twins in lawsonite could induce substantial seismic anisotropy reduction, particularly for the maximum S-wave anisotropy in lawsonite and whole rocks by up to 3.67% and 1.46%, respectively. Lawsonite twinning needs to be considered when determining seismic anisotropy in the subducting oceanic crust in cold subduction zones.


Introduction
Seismic anisotropy, which is useful for studying tectonic fabric in the Earth, has been widely observed in subduction zones [1][2][3][4], including subducting oceanic slabs [5,6]. It can be caused by the crystal preferred orientation (CPO) of elastically anisotropic minerals [7,8]. Since lawsonite is an elastically anisotropic mineral in blueschist and eclogite facies rocks at the top of the subducting slab in cold subduction zones, the development of the CPO of lawsonite could cause substantial seismic anisotropy. The CPO of lawsonite has been used to interpret various seismic anisotropies observed in cold subduction zones from naturally and experimentally deformed lawsonite-bearing rocks such as blueschist and eclogite [9][10][11][12][13][14].
Several studies performed in the last decade have shown the existence of twins in lawsonite from the Sivrihisar Massif (Turkey) [10,15,16], South Motagua (Guatemala) [16,17], Southern New England Orogen (Australia) [18], Alpine Corsica (France) [19], and experimentally deformed lawsonites [14]. However, there is an insufficient understanding of how lawsonite twinning affects seismic anisotropy. In this study, the effect of lawsonite twinning on the CPO, CPO strength, and seismic anisotropy of lawsonite was investigated utilizing natural lawsonite blueschists collected from Alpine Corsica (France) and the Sivrihisar Massif in the Tavşanlı Zone (Turkey). The main focus of this study is the direct effect of twins in lawsonite on the seismic anisotropy. We compare the CPO and seismic anisotropy of lawsonite in four natural samples with those of the modelled samples where  [37]. (b) Simplified tectonic map of western and central Turkey with sample locations indicated by red star (samples 2023 and 2021), modified from Cao and Jung [10] and Davis and Whitney [34]. CACC

Methods
Foliation is determined by parallel alignment of glaucophane, lawsonite, and white mica, and compositional layering of minerals. The lineation of the samples is defined by the shape preferred orientations of the lawsonite and glaucophane [38], and thin sections were made in the x-z plane (x: parallel to the lineation; z: normal to the foliation). The thin sections were polished with diamond paste and colloidal silica (0.06 μm) and coated with carbon to avoid charging in the scanning electron microscope. Electron backscatter diffraction (EBSD) analysis, utilized to measure the CPOs of lawsonite and glaucophane, was performed in the x-z plane of the samples utilizing Aztec software (Version 4.3, Oxford Instruments, Abingdon, UK) with a symmetry detector attached to a field emission scanning electron microscope (JSM 7100F, JEOL, Tokyo, Japan) at the School of Earth and Environmental Sciences at Seoul National University, Korea. The system was operated at a 20 kV acceleration voltage and a 25 mm working distance. The lawsonite and glaucophane EBSD patterns were automatically indexed by mapping the sample with step sizes ranging from ~1.2-3.0 μm, which were at least 10 times smaller than the average grain size of the samples. To prevent misinterpretation of the original data, the raw EBSD data was

Methods
Foliation is determined by parallel alignment of glaucophane, lawsonite, and white mica, and compositional layering of minerals. The lineation of the samples is defined by the shape preferred orientations of the lawsonite and glaucophane [38], and thin sections were made in the x-z plane (x: parallel to the lineation; z: normal to the foliation). The thin sections were polished with diamond paste and colloidal silica (0.06 µm) and coated with carbon to avoid charging in the scanning electron microscope. Electron backscatter diffraction (EBSD) analysis, utilized to measure the CPOs of lawsonite and glaucophane, was performed in the x-z plane of the samples utilizing Aztec software (Version 4.3, Oxford Instruments, Abingdon, UK) with a symmetry detector attached to a field emission scanning electron microscope (JSM 7100F, JEOL, Tokyo, Japan) at the School of Earth and Environmental Sciences at Seoul National University, Korea. The system was operated at a 20 kV acceleration voltage and a 25 mm working distance. The lawsonite and glaucophane EBSD patterns were automatically indexed by mapping the sample with step sizes ranging from~1.2-3.0 µm, which were at least 10 times smaller than the average grain size of the samples. To prevent misinterpretation of the original data, the raw EBSD data was postcleaned utilizing Aztec software (Oxford Instruments) in three steps: (1) wild spikes with a Minerals 2021, 11, 399 4 of 17 pixel size were removed, (2) zero solution pixels that neighbor six pixels with solutions were eliminated, and (3) Step 1 was repeated.
The procedure to remove twins in lawsonite for plotting the CPOs is described as follows: Twin boundaries having a [001] rotation axis with a 67 • misorientation angle and a 5 • deviation were automatically disregarded as grain boundaries in EBSD orientation maps using Channel 5 software (Version 5.12.74.0). Then, the crystallographic orientation representing each grain was determined by the mean orientation of the lawsonite grains. This procedure was reiterated consistently for all grains in all samples. The lawsonite CPO with and without twins and glaucophane CPO were plotted in a pole figure as one point per grain to avoid introducing a bias by the large grains [39]. The misorientation index [40] was calculated to determine the CPO strength of the minerals.
The seismic velocity and anisotropy of the P-and S-waves of lawsonite were calculated utilizing the CPO, crystal density, and elastic constants of lawsonite [41], with a modified crystallographic reference [9]. They were calculated for glaucophane [42] utilizing a FORTRAN software program [43]. The seismic velocity and anisotropy of lawsonite and whole rocks were calculated for samples with and without twins. To calculate the seismic velocity and anisotropy in the absence of twins in lawsonite, narrow twin areas in lawsonite grains were removed manually in EBSD orientation maps using HKL Channel 5 Tango software. The choice of narrow twin removal in lawsonite was consistent for all samples. All point-per-grain crystallographic orientation data of lawsonite and glaucophane were used for accurate seismic properties with and without twins [39]. The seismic properties of the whole rocks were calculated utilizing the normalized volume fraction of the major minerals as lawsonite and glaucophane ( Table 1). The normalized volume proportions of lawsonite and glaucophane were based on the mineral fractions from the large area EBSD mapping. Minor minerals such as garnet, omphacite, titanite, quartz, phengite, and epidote were ignored in the calculations.

Microstructures
Four representative lawsonite blueschist samples collected from Alpine Corsica at Monte Pinatelle and the Halilbagı area in the Tavşanlı Zone were studied. The major minerals in the blueschists were lawsonite and glaucophane, and the minor minerals were garnet, omphacite, titanite, quartz, phengite, and epidote. Blueschists from Alpine Corsica showed higher phengite and omphacite content than those from the Halilbagı area. All samples showed high volume proportions of lawsonite (Table 1) in the 37-53% range.

CPO of Glaucophane
The glaucophane CPOs are presented in Figure 6. The maxima of the glaucophane [001] axes were aligned subparallel to the lineation, and the maxima of the (110) poles and [100] axes were aligned subnormal to the foliation (Figure 6a,b,d). The (010) poles showed a high concentration subnormal to the lineation near the center of the pole figure, except for sample 2023, which showed weak girdle distributions of the (110) and (010) poles subnormal to the lineation (Figure 6c).

CPO of Glaucophane
The glaucophane CPOs are presented in Figure 6. The maxima of the glaucophane [001] axes were aligned subparallel to the lineation, and the maxima of the (110) poles and [100] axes were aligned subnormal to the foliation (Figure 6a,b,d). The (010) poles showed a high concentration subnormal to the lineation near the center of the pole figure, except for sample 2023, which showed weak girdle distributions of the (110) and (010) poles subnormal to the lineation (Figure 6c).

Seismic Velocity and Anisotropy of Minerals and Whole Rocks
The seismic velocities and anisotropies of lawsonite aggregates with and without twins are presented in Figure 7. For the lawsonite aggregate with twins, the P-wave seismic anisotropy (AVp) ranged from 5.1% to 12.6%, and the maximum S-wave seismic anisotropy (max AVs) ranged from 10.15% to 20.87% (Figure 7a-d). For the lawsonite aggregate without twins, AVp and max AVs ranged from 5.1% to 13.0% and from 10.12% to 22.64%, respectively (Figure 7e-h). Therefore, AVp and max AVs are lower in lawsonite aggregates with twins, except for sample 2021, which shows a very low frequency of lawsonite twinning (Figure 7d,h). The direction of the P-wave velocity is subnormal to the foliation regardless of the existence of twins.

Seismic Velocity and Anisotropy of Minerals and Whole Rocks
The seismic velocities and anisotropies of lawsonite aggregates with and without twins are presented in Figure 7. For the lawsonite aggregate with twins, the P-wave seismic anisotropy (AVp) ranged from 5.1% to 12.6%, and the maximum S-wave seismic anisotropy (max AVs) ranged from 10.15% to 20.87% (Figure 7a-d). For the lawsonite aggregate without twins, AVp and max AVs ranged from 5.1% to 13.0% and from 10.12% to 22.64%, respectively (Figure 7e-h). Therefore, AVp and max AVs are lower in lawsonite aggregates with twins, except for sample 2021, which shows a very low frequency of lawsonite twinning (Figure 7d,h). The direction of the P-wave velocity is subnormal to the foliation regardless of the existence of twins.  The seismic anisotropy of the glaucophane aggregates is presented in Figure 8. The AVp and max AVs of glaucophane range from 13.0% to 25.1% and from 6.33% to 14.48%, respectively, and the direction of the P-wave velocity is aligned subparallel to the lineation (Figure 8a-d).
Minerals 2021, 11, 399 11 of 17 The seismic anisotropy of the glaucophane aggregates is presented in Figure 8. The AVp and max AVs of glaucophane range from 13.0% to 25.1% and from 6.33% to 14.48%, respectively, and the direction of the P-wave velocity is aligned subparallel to the lineation (Figure 8a-d). The seismic anisotropies of the whole rocks are presented in Figure 9. For whole rocks with lawsonite twins, AVp and max AVs vary from 4.3% to 9.5% and from 4.54% to 7.63%, respectively (Figure 9a-d). For whole rocks without lawsonite twins, AVp and max AVs ranged from 4.3% to 9.5% and from 6.00% to 8.13%, respectively (Figure 9e-h). Generally, AVp and max AVs increased in the samples without lawsonite twins, and the direction of the P-wave velocity was similar to that in glaucophane regardless of the existence of twins. The seismic anisotropies of the whole rocks are presented in Figure 9. For whole rocks with lawsonite twins, AVp and max AVs vary from 4.3% to 9.5% and from 4.54% to 7.63%, respectively (Figure 9a-d). For whole rocks without lawsonite twins, AVp and max AVs ranged from 4.3% to 9.5% and from 6.00% to 8.13%, respectively (Figure 9e-h). Generally, AVp and max AVs increased in the samples without lawsonite twins, and the direction of the P-wave velocity was similar to that in glaucophane regardless of the existence of twins.

CPO of Lawsonite and Effect of Twins on the CPO Strength of Lawsonite
All the CPOs of lawsonite in deformed blueschists from Alpine Corsica and the Sivrihisar Massif are characterized by the maxima of the [001] axes aligned subnormal to the foliation regardless of twin existence and by [100] and [010] axes aligned subparallel to the foliation ( Figure 5). These features are consistent with those of previous studies on blueschists and eclogites from the North Qilian suture zone in China [9], the Sivrihisar Massif in Turkey [10], the Diablo Range in California, USA [11], the southern Motagua fault zone in Guatemala [12], and the Kurosegawa Belt in Japan [45]. However, this study reveals that the CPO strength of lawsonite is reduced when considering the existence of twins in lawsonite. The CPO strength of lawsonite with twins is lower (M = 0.078-0.23) than that of lawsonite without twins (M = 0.080-0.24), except in sample 2023 ( Figure 5). Sample 2023 shows a higher CPO strength of lawsonite with twins than without twins (M = 0.12 and 0.09, respectively). This may be related to the scattered distribution of the [001] axes that are partly aligned subparallel to the lineation.  (Figure 5a,b,e,f). Therefore, lawsonite twinning could be one of the mechanisms that weakens the CPO strength of lawsonite. However, we do not know the exact origin of twins in lawsonite. There is a lack of information on the pressure, temperature, and differential stress conditions in which lawsonite twins occur.

CPO of Lawsonite and Effect of Twins on the CPO Strength of Lawsonite
All the CPOs of lawsonite in deformed blueschists from Alpine Corsica and the Sivrihisar Massif are characterized by the maxima of the [001] axes aligned subnormal to the foliation regardless of twin existence and by [100] and [010] axes aligned subparallel to the foliation ( Figure 5). These features are consistent with those of previous studies on blueschists and eclogites from the North Qilian suture zone in China [9], the Sivrihisar Massif in Turkey [10], the Diablo Range in California, USA [11], the southern Motagua fault zone in Guatemala [12], and the Kurosegawa Belt in Japan [45]. However, this study reveals that the CPO strength of lawsonite is reduced when considering the existence of twins in lawsonite. The CPO strength of lawsonite with twins is lower (M = 0.078-0.23) than that of lawsonite without twins (M = 0.080-0.24), except in sample 2023 ( Figure 5). Sample 2023 shows a higher CPO strength of lawsonite with twins than without twins (M = 0.12 and 0.09, respectively). This may be related to the scattered distribution of the [001] axes that are partly aligned subparallel to the lineation.  (Figure 5a,b,e,f). Therefore, lawsonite twinning could be one of the mechanisms that weakens the CPO strength of lawsonite. However, we do not know the exact origin of twins in lawsonite. There is a lack of information on the pressure, temperature, and differential stress conditions in which lawsonite twins occur.

Effect of Lawsonite Twinning on Seismic Anisotropy in Subducting Oceanic Crust
The seismic anisotropy of lawsonite reduces the AVp and AVs by up to 0.60% and 3.67%, respectively, when lawsonite twinning is taken into account (Table 2). Notably, S-wave anisotropy is significantly greater than P-wave anisotropy, possibly due to the seismic anisotropy of the lawsonite single crystal [41] which shows a much larger max AVs value (65%) than the AVp value (24%) [9]. The effect of lawsonite twinning on the seismic anisotropy of whole-rock blueschist is also presented in Table 2. The seismic anisotropy of the whole rock also reduces the AVp and AVs by up to 0.30% and 1.46%, respectively, when lawsonite twinning is considered. Lawsonite twinning has only a minor effect on the P-wave anisotropy, but a relatively larger effect on the S-wave anisotropy of lawsonite and whole rock. The seismic anisotropy reduces with increasing area of twinned domain in lawsonite (Tables 1 and 2). For example, the difference of the max AVs of lawsonite with and without twins is very small for sample 2021, which has a very small area of twinned domain; however, the difference of max AVs of lawsonite is 3.67% for sample 3034, which has a large area of twinned domain. 1 Change of seismic anisotropy of lawsonite and whole rock with twin compared to that of lawsonite and whole rock without twin. Anisotropy of P-wave (AVp) = 200 × (Vp max − Vp min )/(Vp max + Vp min ), and anisotropy of S-wave (AVs) = 200 × (Vs1 − Vs2)/(Vs1 + Vs2). Max AVs: maximum anisotropy of S-wave velocity.
In Figure 10, we manually removed the blue part (narrow area) of the twin in the lawsonite grain. This is because most of the twins in lawsonite in the studied samples are observed as narrow areas (for example, see Figure 3). The choice of twin removal may have an influence on the resulting CPO and seismic anisotropy of lawsonite. According to our study, the seismic anisotropy reduces with increasing area of the twinned domain in lawsonite. If wide areas of twins are removed in lawsonite, the area of the twinned domain in the sample would become larger, resulting in a large effect of twins in lawsonite on the reduction of seismic anisotropy. Thus, the choice of narrow twin removal in this study has produced the minimum effect of twins on the reduction of seismic anisotropy.
Based on these findings, the existence of twins in lawsonite can induce the reduction of seismic anisotropy in blueschist, particularly S-wave anisotropy, which is attributed to the ability of the {110} twin to weaken the lawsonite CPO strength (i.e., weak alignment of the [100] and [010] axes) ( Figure 5). The CPO strength, along with other factors, including rock type and interactions between the CPOs of different minerals, can lead to a change in the seismic anisotropy in lawsonite blueschist [10]. This study clearly shows that lawsonite twinning can be an important factor in determining the seismic anisotropy of lawsonitebearing rocks if the area of the twinned domain in lawsonite is large.
Lawsonite twinning has also been reported in natural lawsonite eclogites [10,18]. As eclogite mainly consists of anhydrous minerals such as garnet and omphacite, it has a weaker seismic anisotropy than blueschist [46,47]. Therefore, the influence of lawsonite on the seismic anisotropy of eclogite is significant [12,13]. Lawsonite twinning can also produce seismic anisotropy reduction in lawsonite eclogite, particularly S-wave anisotropy. As lawsonite is prevalent in the subducting oceanic crust of cold subduction zones, lawsonite twinning can be important in determining seismic anisotropy.

Conclusions
The effect of lawsonite twinning on the CPO and seismic anisotropy of lawsonite and lawsonite blueschist was studied using lawsonite blueschists collected from Alpine Corsica and the Sivrihisar Massif. The results indicate that polysynthetic and deformation twins of lawsonite exist on the {110} plane with the rotation of the [001] symmetry axis by 67 • . The CPO of lawsonite with twins is weakened due to the scattered distribution of the [100] and [010] axes. Lawsonite twinning induces a substantial reduction in the maximum S-wave anisotropy (max AVs) by up to 3.67% and 1.46% for lawsonite and whole rock, respectively, and a minor reduction of the P-wave anisotropy (AVp) by up to 0.6% and 0.3% for lawsonite and whole rock, respectively. Considering the distribution of lawsonite in blueschist and eclogite in cold subduction zones, lawsonite twinning needs to be taken into account in determining seismic anisotropy.