# Fileset

[GaS_Brillouin_Resbmission.pdf](https://mdr.nims.go.jp/filesets/b0ed6e77-2512-4980-840d-c933b4a1dab8/download)

## Creator

Daniel Ramesh Paulo-Wach, Ethan Chen, [Masaru Nakamura](https://orcid.org/0000-0001-9729-847X), Oscar D. Dubon, Kristie J. Koski

## Rights

This article may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. This article appeared in Daniel Paulo-Wach, Ethan Chen, Masaru Nakamura, Oscar D. Dubon, Kristie J. Koski; Large elastic anisotropy in Brillouin scattering of copper-intercalated GaS and GaSe. Appl. Phys. Lett. 29 December 2025; 127 (26): 262202 and may be found at https://doi.org/10.1063/5.0300066.[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Large Elastic Anisotropy in Two-dimensional Layered GaS and GaSe Modulated by Copper Intercalation](https://mdr.nims.go.jp/datasets/e1f547d3-7519-4083-a5ca-339dae23e2b6)

## Fulltext

Large Elastic Anisotropy in Brillouin Scattering of Copper-Intercalated GaS and GaSeDaniel Ramesh Paulo-Wach,1, 2 Ethan Chen,3 Masaru Nakamura,4 Oscar D. Dubon,1, 2 and Kristie J. Koski3, ∗1Department of Materials Science and Engineering,University of California Berkeley California 94720, USA2Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA3Department of Chemistry, University of California, Davis, California 95616, USA4National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan(Dated: December 9, 2025)Brillouin scattering of bulk layered, hexagonal gallium sulfide (GaS) and gallium selenide (GaSe)were measured both as pristine materials and intercalated with copper and silver. The soundvelocities, refractive index, and four of the five independent elastic stiffnesses (cij) were determined.Values of the c11 elastic stiffness in thin crystals are higher than in previous measurements, bringingthe elastic anisotropy (c11/c33) into a range typical of other layered materials. Copper intercalationshowed a larger effect on acoustic phonons in GaS than GaSe.The processing of materials can have a significant ef-fect on their physical properties.[1] Gallium sulfide (GaS)and gallium selenide (GaSe) are two-dimensional lay-ered materials with a hexagonal structure that can begrown as either p-type or n-type.[1] Both materials showunique opto-electronic properties with promise in pho-todetectors, lasers, and elasto-optical applications.[2–4]Additionally, the properties of these materials are highlysensitive to dopants and impurities.[1] GaS and GaSehave been the subject of many previous mechanical, Bril-louin and ultrasonic investigations.[5–13] Previously re-ported Brillouin scattering measurements showed a sur-prisingly low elastic anisotropy (c11/c33) for a layered ma-terial, where the elastic stiffness in the plane was foundto only be about three times larger than that in thestacking direction.[6–8, 11] Conversely, most other lay-ered materials have significantly larger reported elasticanisotropy.[14]Brillouin scattering is a laser-based technique thatmeasures acoustic phonons to extract elastic stiffnessesand sound velocities of materials. The interaction be-tween the incident laser and the sample leads to theStokes creation or anti-Stokes annihilation of an acous-tic phonon with a wavevector q⃗, visible as a peak in theBrillouin spectra. In a Brillouin scattering experiment,measurements taken at different geometries of incidentlight, scattered light, and sample angles yield represen-tative spectra peaks of acoustic phonons in different di-rections. Brillouin scattering data typically shows onestrong elastic scattered peak surrounded by Stokes/anti-Stokes pairs of a longitudinal (or quasi-longitudinal) andtwo transverse (or quasi-transverse) phonon peaks.We investigate Brillouin scattering of GaS and GaSeand use intercalation of copper to modulate the elas-tic properties, motivated by the large impact of dopantsand impurities on these materials.[1] We also intercalateGaS with silver. While we find values of the out-of-plane c33 elastic stiffness are similar to previous mea-∗ koski@ucdavis.edusurements, we find significantly different values of the in-plane c11 stiffness yielding a larger, but more expected,elastic anisotropy than previous measurements.[6–8, 11]We show that intercalation of silver and copper has a sig-nificant impact on the acoustic phonons of GaS but notmuch impact in GaSe.Bulk GaS crystals were grown via the Bridgmanmethod following Nakamura et al.[15], and bulk GaSecrystals are legacy crystals from Cradley Crystals Corpo-ration. All samples were exfoliated to thinner dimensions(<∼ 0.3 mm) using a razor blade to separate the layers.Crystals were intercalated with copper and silver usingsimilar processes detailed in Koski et al.[16, 17] Sam-ples were suspended in free space for measurement. Bril-louin scattering was collected with a scanning, tandem,multi-pass Sandercock TFP-1 with a Coherent Verdi V6NdYVO4 laser at λo = 532 nm with a 20X SLWD Mitu-toyo objective (N.A. = 0.42) for collection with laser pow-ers of ∼ 7 mW on the sample for GaS and ∼ 3 mW on thesample for GaSe. In 90◦ geometries, a 10x Mitotoyo wasused for focusing. Finesse was approximately 100-300.Acquisition times were on the order of 5-10 minutes forGaS (1100 counts) and several hours for GaSe. Regions ofhigher intercalant concentration were measured by look-ing for higher signal intensity, which is theorized to orig-inate from some form of plasmonic enhancement.[18, 19]XRD was collected on a Bruker Eco Advance with a CuKα source. Intercalant concentration was determinedusing X-ray photoelectron spectroscopy (XPS) with aKratos Axis Supra with an Al anode and scanning elec-tron microscope energy dispersive x-ray spectra (SEM-EDX) on a FEI Scios Dual Beam FIB/SEM with an Ox-ford X-MaxN EDX detector. Raman spectra were col-lected from a home-built Raman system with a Prince-ton Instruments SCT320, 1800 groove/mm grating, andPixis CCD, with a 532 nm Coherent Sapphire SF laserat 15 mW.Figure 1 shows characterization of Cu-intercalated GaSand GaSe and Ag-intercalated GaS. GaS and GaSe bothhave hexagonal crystal structures.[1] XRD (Fig 1a,b) wasacquired on the flakes used for Brillouin scattering. Thethin single-crystal GaS and GaSe flakes show the (00l)210 20 30 40 50 60 70 80Intensity (arb. u.)Wavenumber (cm-1)2 (002) (004)(006)(008)(002)(004) (0010)(105)(106) (00 12)A11gA11gA21gA21gE1gE2gE2gGaSAg-GaSCu-GaSGaSAg-GaSAg-GaSCu-GaSGaSeGaSeCu-GaSeCu-GaSeCu-GaSeCu-GaS200 400 600 800Wavenumber (cm-1)20 30 40 50 602Θ Intensity (arb. u.)200 300cm-102 Δcm-1200 300cm-10246 Δcm-1200 400 600abcde gf0 105Energy (keV)Intensity0400800Binding Energy (eV)0400800Ag GaGaSGaGa 2pCu2pO 1sS 2pSe 3dO 1sCu2pCCGa 2pFIG. 1. (a) XRD of GaS pristine and intercalated with Agand Cu, (b) XRD of GaSe and intercalated with Cu, (c) Ra-man scattering of GaS and intercalated with Ag and Cu Inset:Shift of Cu- and Ag-GaS raman peaks relative to GaS. Un-certainties are within the data point. (d) Raman scatteringof GaSe pristine and intercalated with Cu. Cu intercalationsuppresses the GaSe photoluminescence edge. Inset: Shift ofCu-GaSe Raman peaks relative to GaSe. Uncertainties arewithin the data point. (e) SEM-EDX of Ag-intercalated GaS,(f) XPS wide spectra of Cu-GaS and (g) Cu-GaSe.peaks, limiting structural determination to the c latticeconstant which is in the direction of the layer stacking.The c-axis of GaS (c = 15.49 ± 0.01 Å) shows negligiblechange with intercalation of Cu-GaS (c = 15.50 ± 0.01Å) and Ag-GaS (c = 15.50 ± 0.01 Å). GaSe (c = 15.96 ±0.01 Å) also shows negligible change upon intercalationof Cu-GaSe (c = 15.95 ± 0.01 Å). Intercalation generallyleads to an initial contraction in the host followed by anexpansion[20]. Using SEM-EDX, the intercalant concen-trations were silver at 0.1 atm % in GaS, Cu at 0.6 atm% in GaS, and Cu at 0.2 atm % in GaSe.As noted by previous studies, wavenumber shifts in Ra-man scattering spectra of intercalated materials are oftencomplex, demonstrating either optical phonon stiffeningor softening as a result of expansion or contraction of thehost, local polarizability changes from the intercalant,and the guest/host donor/acceptor nature.[19, 21, 22]Figure 1c,d show the Raman spectra of intercalated GaSand GaSe, respectively. An inset is provided showingthe change in wavenumber shift from the unintercalated-40 -20 0 20 40Frequency (GHz)Intensity (arb. u.)-30 -20 -10 0 10 20 30Frequency (GHz)Intensity (arb. u.)LLLL LLT TTTLGaSAg-GaSCu-GaSGaSAg-GaSCu-GaSHVHHHVHHHVHHkiksqkiksqcacaabL32 34FIG. 2. Brillouin scattering of GaS pristine and intercalatedwith Ag (blue) and Cu (red) in a (a) 90a geometry usingHH and HV polarizers to identify longitudinal and transverseacoustic modes and (b) 180 geometry. Inset shows the stiff-ening of the longitudinal acoustic phonon with intercalation.k⃗i is incident light; k⃗s is scattered light; and q⃗ is the probedwavevectorhost. In GaS (Fig 1c), Cu and Ag intercalation in GaSleads to an increase in the Raman shifts while interca-lation of Cu in GaSe increases some Raman shifts anddecreases others. The Raman spectra of GaSe (Fig 1d)show a large broad peak at higher wavelengths due tothe strong photoluminescence in GaSe.[23] Intercalationof Cu in GaSe quenches or shifts the photoluminescenceas seen with the decrease of the strong edge of this peak.As hexagonal crystals, both GaSe and GaS have anelastic stiffness tensor with five independent stiffnesses(cij) given by Eq. 1,cij =c11 c12 c13 0 0 0c12 c11 c13 0 0 0c13 c13 c33 0 0 00 0 0 c44 0 00 0 0 0 c44 00 0 0 0 0 c66 (1)where c12 = c11 − 2c66.[24] Several scattering geome-tries, and various tilts or rotations from those scatteringgeometries, were used to extract as many elastic stiffnesstensor elements as possible. Diagrams of the primaryscattering geometries [25] can be found in Figure 2 andFigure 3. In a 90a geometry (Fig 2a), the input light andscattered light are placed at ∼90 degrees to each other.3The sample is rotated between the input and scatteredlight such that the q⃗ is in the ab-plane of the crystalsprobing in that direction yielding the c11 stiffness value.In a backscattering geometry (Fig 2b), the input andscattered light are oriented normal to the surface of thesample. This orientation of input and scattered light onthe sample measures Brillouin scattering in the stackingdirection and gives the c33 stiffness value. In a 90r geom-etry, the crystal is placed such that the q⃗ probes downthe c-axis correcting for index of refraction differences inthe sample (Fig 3b) with respect to the incoming beamand scattered light such that q⃗ is along the c-axis. Thisprobes the same direction as backscattering but allowsappearance of transverse modes, which are eliminated bysymmetry in backscattering. Mode assignment was madeusing cross (HV gives T) and parallel (HH gives L) polar-izers in a 90a scattering geometry (Fig 2a). L identifieslongitudinal or quasi-longitudinal modes and T identifiestransverse or quasi-transverse modes. Two longitudinalacoustic modes were found in GaS at ∼ 31 GHz and ∼ 22GHz. The mode at 31 GHz also shows up in the Brillouinbackscattering measurement (Fig 2b) and can be as-signed as a spontaneous backscattering mode,[26] leavingthe 22 GHz mode identified as a genuine 90a longitudinalmode. This type of spontaneous Brillouin backscatteringhas been known to occur in thin samples.[26] Two trans-verse modes are identified in GaS under HV polarization.In Ag-intercalated GaS, a third longitudinal mode is ob-served, possibly due to intermediate structures and in-complete staging similar to that suggested in 20 whichcould also be caused by localized islanding of the inter-calant in the sample with regions of localized expansion;intercalant levels are low enough for large separation inisland regions. The second transverse mode in 90a in Ag-GaS vanishes; however, we are able to extract the soundvelocity as that transverse mode also appears in 90r scat-tering. Brillouin backscattering (Fig 2b, inset) showsonly minor (∼ ± 0.4 GHz) acoustic phonon stiffeningin the stacking direction with intercalation. The acous-tic phonons in the 90a scattering direction correspond toc11 (longitudinal), c44 (transverse), and c66 (transverse)and can be used to find c12. Thus, these two primaryscattering geometries (90a, 180) are complementary andallow measurement of the acoustic phonons down the a(Fig 2a) and c axes (Fig 2b) and determination of allthe elastic stiffnesses except c13. Acquisition of Brillouinscattering in GaSe required significant time (hours todays) due to very low signal (Fig 3). Just as in GaS,three primary scattering geometries were used (90a, 180,and 90r).The sound velocity, V , can be calculated from the Bril-louin frequency shift, ∆ν, the refractive index, n, thelaser wavelength λo, and the scattering angle Θ as mea-sured inside the material (Eq 2).V =|∆ν|λo2n sin(Θ2) (2)Frequency (GHz)-20 0 20 -20 0 20-20 0 20Intensitykiksq kiksqcacakiksqcaa b c90r 18090aGaSeCu-GaSeGaSe GaSeCu-GaSeCu-GaSeLLLL LTTFIG. 3. Brillouin scattering of pristine (black) and Cu-intercalated (red) GaSe in a (a) 90a (b) 90r, and (c) 180scattering geometry.For geometries used in this study, this formula evalu-ates to:V90a =|∆ν|λo√2; V180 =|∆ν|λo2n; V90r =|∆ν|λo√4n2 − 2(3)The stiffness tensor element is calculated from thesound velocity, given a density ρ by Eq 4. A den-sity of 3860 kg/m3 is used for GaS and 5030 kg/m3 forGaSe.[7, 9, 27] Liquid gradient density techniques can beused to find the density of a material,[19] however GaSand GaSe have densities of 3860-5040 kg/m3 which isfar above what that method can measure. As concen-trations of intercalant are low, we assume small to nodensity change.cij = V 2ρ (4)The longitudinal sound velocity V is the same for 180and 90r measurements, as both measure in the stackingdirection (c), and allows determination of the refractiveindex n, but this technique is prone to large error Wefound that GaS n = 2.5 ± 0.1; AgGaS n = 2.6 ± 0.1;CuGaS n = 2.5 ± 0.1; GaSe n = 2.5 ± 0.1; CuGaSe n =2.5± 0.1. The wavelength dependent refractive index ofGaS has been carefully measured by Yael et al.[28] which,interpolated, yields a refractive index of 2.65 at 532 nm,which was used for sound velocity calculations here, and2.6 for GaSe from Wasscher et al.[5] Wide variability isfound in reported values of refractive index.[5, 10, 28] The90r geometry is also used to extract c44 in Ag-GaS wherethe corresponding transverse mode appeared in the 90rmeasurement but not the 90a.The crystal is rotated in a 90a geometry (Fig 4a) toensure that the q⃗ is measured in the plane. Since thecrystals are hexagonal, there should be no deviation infrequency shift throughout a 180 degree rotation. Fig-ures 4a through 4c show the frequency shifts of the acous-tic phonons in pristine and intercalated GaS; a line is4-20 0 20Intensity (arb. u.)-20 0 20-40o-30o-20o-10o0o10o20o30o40o45o-40o-30o-20o-10o0o10o20o30o40o45o-20 0 20Frequency (GHz)-40o-30o-20o-10o0o10o20o30o40o45o-40-20 0 20 40Tilt Angle-40-20 0 20 40Tilt Angle0102030GHz-40-20 0 20 40Tilt AnglecaGaS AgGaS CuGaSGaS AgGaS CuGaSd e fg h i40 80010141822GHzGaS40 800in-plane rotation (o)  AgGaSLTTLTTLLTca0 40 80 120CuGaSa b cFIG. 4. Brillouin frequency shifts of (a) GaS, (b) Ag- and(c) Cu-intercalated GaS rotated in a 90a geometry to verifythat acoustic phonons measured are all in the ab-plane withmodes showing almost no variation. Backscattering spectra asa function of sample tilt of (d) GaS, (e) Ag-intercalated GaS,(f) Cu-intercalated GaS and (g) Brillouin frequency shifts ofGaS, (h) Ag-intercalated GaS, and (i) Cu-GaS.drawn for the frequency shifts of pristine GaS. Intercala-tion of silver results in two longitudinal modes with onegreater frequency than GaS and one at lower frequency.Intercalation of Cu notably decreases the Brillouin fre-quency shifts of all acoustic phonon modes in the plane.Table I presents the sound velocities from Brillouinscattering: the longitudinal velocity in the ab-plane (VLa,giving c11) and along the stacking direction (VLc, givingc33), as well as the pure-transverse velocities for shear inthe ab-plane (VTa1, giving c66), and in the ac-plane (VTa2,giving c44). Values of the sound speed match well withrecent measurements which found VLc = 3140±20m/s inGaS.[12] Intercalation of metals into GaS and GaSe doesnot affect the sound velocity in the stacking directionas found in Brillouin scattering investigations of otherintercalated layered materials.[19, 21, 22, 29]GaS was tilted in the backscattering geometry to breaksymmetry and access the transverse acoustic phonon inthe ac-plane which would give VTa2yielding c44 in the ex-trapolated limit of small tilt angles. Fig 4 shows the tiltof the sample in 180 backscattering along with peak posi-tions for GaS (Fig 4a,b), AgGaS (Fig 4c,d), and CuGaS(Fig 4 e,f). In principle, the variation with tilt angle (-50◦ to 50◦; sin−1(sin 50◦/n) ≈ 18◦) provides informationabout c13, but due to the fairly large refractive index, theTABLE I. Sound velocity of GaS, GaSe and metal intercalatedGaS and GaSe in m/s (error in parenthesis).VLa VLc VTa1 VTa2GaS 8069(31) 3284(26) 3517(46) 1877(10)Ag-GaS 8992(76); 7837(72) 3186(158) 3244(39) 1740(12)Cu-GaS 6867(23) 3280(7) 3319(21) 1499(12)GaSe 7058(48) 3223(1) 4180(38) 2496(102)Cu-GaSe 6901(376) 3214(10) 4357(56) 2569(10)TABLE II. Elastic stiffnesses, cij in GaS and GaSe in GPafrom compared to Polian et al.[8] and Chiang et al.[6]c11 c33 c44 c66 c12 c11/c33GaS 251(2) 41.6(7) 13.6(1) 47.7(1) 156(4) 6.0Ag-GaS 312(5) 39(4) 11.7(2) 41(1) 231(7) 8.0Cu-GaS 182(1) 41.5(2) 8.7(1) 42.5(5) 96.9(2) 4.4GaS [8] 122 38 9.8 44.3 33.4 3.2GaSe 251(3) 37.8(1) 12.2(1) 31(3) 188(9) 6.6Cu-GaSe 240(26) 37.6(2) 11.8(0.2) 33(1) 173(27) 6.4GaSe [6] 105 35.1 10.4 36.3 32.5 3.0thinness of the crystal, and the uncertainties in the otherstiffness tensor elements, it was not possible to extracta meaningful c13 value. Cu-GaSe (Fig 4f) shows noisepeaks from excess laser light entering the TFP-1 at andnear normal incidence.Table II gives the elastic stiffnesses, cij , for pristineand intercalated GaS and GaSe. Values of the out-of-plane longitudinal elastic stiffness (c33) match wellwith recent measurements which found c33 = 38.1 ±0.5 GPa.[12] Effects of intercalation on the acousticphonons of layered crystals are complicated.[22] Simi-lar to almost all other measurements of Brillouin scat-tering of intercalated layered materials, the intercalationaffects the in-plane stiffness more than the plane-normalstiffness.[19, 21, 22, 29]We find significantly larger in-plane sound velocity andelastic stiffness (c11) yielding larger elastic anisotropy(c11/c33) than previous measurements, displayed in TableII. Values for c11 derived from previous measurements[8]predict a longitudinal mode at 15 GHz in Figure 2; in-stead, we find a longitudinal mode much higher at ∼20GHz. There are several explanations possible: (i) samplegrowth differences coupled with impurities, (ii) overlap-ping longitudinal modes from successive orders in oldertechniques, (iii) issues with scattering geometries thatdo not measure in the correct direction, or (iv) failure toproperly identify mode assignments such as with usingpolarizers. Previous measurements had disagreement inelastic stiffness values using different techniques.[8, 13]It is possible that the first published measurements mayhave biased later results.[9, 10, 13, 27, 30]Previous experiments by Polian et al.[8] used ”up to5 passes with the Tropel Inc Model FP-100 Fabry-Perot5-40 0 40Intensity a b c40GHz0 20 40 60Frequency (GHz)40GHz0 20 40 60ca kiksq59 GHzFIG. 5. (a) Schematic illustration of collection using a Sander-cock TFP. (b) Mode misassignment possible with older inter-ferometry techniques. Modes can be misaligned in frequencyshift because of overlapping frequency ranges (yellow) includ-ing being hidden by the elastic peak (red). (c) Geometry usedby Chiang et al.[6]interferometer.” This is a multi-pass transmission instru-ment that yields inelastic peaks between two successiveelastic mode; it is possible to mistakenly reference a Bril-louin peak to the wrong elastic mode, or have overlap ofan inelastic mode within the elastic peak giving an in-correct frequency shift. The advances from the Sander-cock design of the tandem multi-pass Fabry-Perot inter-ferometer (TFP-1 and TFP-2) eliminate overlapping or-der modes so only a single elastic peak is seen with minor,nearly invisible doublet ghost peaks from adjacent orders.Honma et al.[9] used a similar system and published theraw data that make it possible to observe the overlappingorder issue and possible peak misassignment.[9] Figure5a shows an example of the kind of spectra that wouldbe seen in the TFP-1 Sandercock design which makes iteasy to identify the single strong elastic peak and theinelastic peaks referenced to it. Figure 5b shows anexample of the kind of spectra that might be observedusing a Tropel Model FP-100, with two elastic peaks ap-pearing at 0 and 60 GHz, each of which serves as a zeroreference for its own set of inelastic peaks. Such spec-tra could easily be misinterpreted by referencing eachinelastic peak to the wrong elastic peak (yielding an in-correct measurement of 20 GHz in this example, whenthe correct Brillouin shift should be 40 GHz). Honma etal.[9] observed other peaks at high frequency that wentunassigned, which could belong to a mode with a soundvelocity of 8069 m/s, as we measure.[9] Also, it is pos-sible that a limited free spectral range used in previousstudies would result in a longitudinal mode overlappingthe elastic mode, obscuring observation of a mode witha large sound speed in the ab-plane.[9] Overlap of ordersand misassignments of longitudinal and transverse peaksis something that has long contaminated Brillouin resultsand is especially prevalent in systems that use single ordouble non-scanning etalons or VIPAs.[25]Another issue could be measuring down an incorrectaxis. Figure 5 shows a representative schematic that issimilar to the drawing of Chiang et al. used in measuringGaSe in a 90a geometry.[6] In this geometry, the inputbeam is at an angle to the c-axis of the crystal and thescattered light is out the a-normal face of the crystal.The q⃗-vector ends up being between a and c, thus givinga much lower sound velocity in that direction. In thisstudy, we rotated the crystal in a 90a geometry to ensurethe q⃗-vector is in the plane and often found we sometimeshad to redo measurements many times. Previous studiesdid not have a technique at the time to ensure that the q⃗-vector was measuring only in the plane. Finally, previousmeasurements may not have used a polarizer to identifymode assignments, leading to misassigned modes.There are other issues that could account for the dif-ferences in measured anisotropy, such as differences ingrowth technique, sample handling, dopants, or contam-ination. The significant effect that impurities and samplequality have on sound velocity and stiffness in GaS wasnoted by Gatulle et al.[13] in ultrasonic measurements,with reported sound velocities ranging almost ± 1000m/s across different samples. In contrast, we found asimilar longitudinal sound velocity across different sam-ples, but there was clearly some minor sample-to-samplevariation. No major impurities were detected with EDXin GaS or GaSe. GaS grown via iodine vapor transport asin Polian et al.[7, 8] can have an impurity concentrationof up to 10 ppm, significantly affecting properties as re-vealed by Brillouin scattering.[1, 8] Bridgman-grown GaScrystals in this work show broader Raman modes in pris-tine versus intercalated, suggesting impurities or defects,which are removed by chemical treatment in the interca-lation process, narrow observed Raman modes. Interca-lation at very small amounts radically affects the elasticstiffness of GaS, demonstrating that even the smallest ofimpurities has a strong effect on elastic properties in thismaterial.In terms of previous crystal preparation,[6–8] the crys-tal was measured at a polished edge to access phononwavevectors located in the ab-plane of GaS needles. Weattempted to polish the edges to access these additionalangles as well. We found that polishing the edges frayedthe crystal at the layer edge, such that the acoustic modeswe detected were still often in the c-direction or somecombination, thereof, not accessing pure ab-plane modes.In conclusion, this study finds a large elastic anisotropy(c11/c33) in GaS and GaSe crystals, which differs frompreviously reported values for GaS and GaSe but is con-sistent with the strong anisotropy observed in other two-dimensional layered materials. Intercalation has a signif-icant, but complicated, effect on the sound velocity andelasticity in these materials, similar to the large effect onother properties of GaS and GaSe.[1]DATA AVAILABILITYData available on request from the authors.ACKNOWLEDGMENTSThe authors acknowledge funding from the NationalScience Foundation NSF-DMR-2202472 and NSF-DMR-2500503. DPW acknowledges funding support from the6National GEM Consortium. MN acknowledges GaSgrowth support by JSPS KAKENHI Grant Number24K08274. The authors thank B. W. Reed for usefulconversations.[1] R. M. A. Lieth, Preparation and crystal growth of mate-rials with layered structures, Vol. 1 (Springer Science &Business Media, 1977).[2] T. Barker, A. Gray, M. Weir, J. Sharp, A. Kenton,Z. Kudrynskyi, H. Rostami, and A. Patané, npj FlexibleElectronics 9, 2 (2025).[3] Y. Sun, Z. Ren, Z. Zhao, F. Zhang, and F. Xing, ACSApplied Nano Materials 8, 8712 (2025).[4] S. Ahmed, P. K. Cheng, J. Qiao, W. Gao, A. M. Saleque,M. N. Al Subri Ivan, T. Wang, T. I. Alam, S. U. Hani,Z. L. Guo, et al., ACS nano 16, 12390 (2022).[5] J. Wasscher and J. Dieleman, Physics Letters A 39, 279(1972).[6] T. Chiang, J. Dumas, and Y. Shen, Solid State Commu-nications 28, 173 (1978).[7] A. Polian, J. M. Besson, M. Grimsditch, and H. Vogt,Applied Physics Letters 38, 334 (1981).[8] A. Polian, J. Besson, M. Grimsditch, and H. Vogt, Phys-ical Review B 25, 2767 (1982).[9] Y. Honma, M. Yamada, K. Yamamoto, and K. Abe,Journal of the Physical Society of Japan 52, 2777 (1983).[10] N. C. Fernelius, Properties and bibliography of GaSe,Tech. Rep. (1994).[11] M. Fischer, A. Polian, A. Chevy, and J. Chervin, Solidstate communications 56, 311 (1985).[12] W. Al-Basheer, C. Viernes, R. Zheng, S. Netzke,K. Pichugin, and G. Sciaini, ACS omega 9, 47475 (2024).[13] M. Gatulle, M. Fischer, and A. Chevy, physica statussolidi (b) 119, 327 (1983).[14] M. Ren, J. Z. Liu, L. Wang, and Q. Zheng, arXivpreprint arXiv:1405.4086 (2014).[15] M. Nakamura, H. Nakamura, K. Shimamura, andN. Ohashi, Journal of Crystal Growth 573, 126303(2021).[16] K. J. Koski, J. J. Cha, B. W. Reed, C. D. Wessells,D. Kong, and Y. Cui, Journal of the American ChemicalSociety 134, 7584 (2012).[17] K. J. Koski, C. D. Wessells, B. W. Reed, J. J. Cha,D. Kong, and Y. Cui, Journal of the American ChemicalSociety 134, 13773 (2012).[18] W. Robertson, A. Moretti, and R. Bray, Physical ReviewB 35, 8919 (1987).[19] B. W. Reed, C. Tran, and K. J. Koski, Physical ReviewMaterials 7, 044003 (2023).[20] A. Powell, Annual Reports Section” C”(Physical Chem-istry) 90, 177 (1993).[21] B. W. Reed, V. Huynh, C. Tran, and K. J. Koski, Phys-ical Review B 102, 054109 (2020).[22] B. W. Reed, E. Chen, and K. J. Koski, Nano Letters 24,11954 (2024).[23] S. Quan, Y. Wang, Y. Liang, J. Jiang, B. Zhong, K. Yu,H. Zhang, and G. Kan, The Journal of Physical Chem-istry C 124, 10185 (2020).[24] J. F. Nye, Physical properties of crystals: their represen-tation by tensors and matrices (Oxford university press,1985).[25] P. Bouvet, C. Bevilacqua, Y. Ambekar, G. Antonacci,J. Au, S. Caponi, S. Chagnon-Lessard, J. Czarske, T. De-houx, D. Fioretto, et al., Nature Photonics 19, 681(2025).[26] Y. Takagi and R. W. Gammon, Journal of applied physics61, 2030 (1987).[27] B. Powell, S. Jandl, J. Brebner, and F. Levy, Journal ofPhysics C: Solid State Physics 10, 3039 (1977).[28] Y. Gutierrez, D. Juan, S. Dicorato, G. Santos, M. Duwe,P. H. Thisen, M. Giangregorio, F. Palumbo, K. HIngerl,C. Cober, et al., Dieletric function crystalline 2H-GaS(2022).[29] B. W. Reed, D. R. Williams, B. P. Moser, and K. J.Koski, Nano letters 19, 4406 (2019).[30] G. Belen’kĭi, E. Y. Salaev, and R. Sulĕimanov, SovietPhysics Uspekhi 31, 434 (1988).