betaFeedback

WorksPhysics and Astronomy

X-ray photo-induced atomic motion in Phase Change Materials and conventional covalent chalcogenide glasses

Abstract

X-ray Photon Correlation Spectroscopy (XPCS) enables direct access to atomic-scale dynamics in disordered materials, revealing both spontaneous and X-ray-induced relaxation processes. Here, we study two compositionally similar alloy glasses near their glass transition temperatures: the phase change material (PCM) Ge15Sb85 and the non-PCM alloy Ge15Te85. Both exhibit X-ray induced atomic motion, yet with markedly different responses. Ge15Sb85 undergoes an immediate transition to a photo-induced yielding state, characterised by stationary dynamics governed solely by the absorbed dose. In contrast, Ge15Te85 shows a progressive slowing-down of the relaxation process, accompanied by a crossover from compressed to stretched exponential decay in the density autocorrelation functions. This behaviour is consistent with the emergence of liquid-like collective motion as supported by de Gennes narrowing in the wave-vector dependence of the dynamics at length scales comparable with the first sharp diffraction peak. Unlike Ge15Sb85, this alloy does not reach a stationary regime within experimental timescales, implying that the yielding transition occurs only after thousands of seconds with the available dose rate. Its response is also temperature dependent: at lower temperatures, the dynamics reflects intrinsic stress relaxation processes, whereas at higher temperatures becomes dose-controlled. These findings demonstrate that the dynamical response to X-ray excitation is not determined solely by chemical composition or bonding character, but results from the interplay between irradiation effects and structural relaxation pathways.

Introduction

X-rays are a fundamental tool for studying the intrinsic properties of a material [1]. Over the past decades, improvements in the brilliance and coherence of X-ray beams delivered in modern synchrotron sources have led to the development of powerful techniques to investigate the dynamical properties of matter, one of the most significant being X-ray Photon Correlation Spectroscopy (XPCS) [2]. This method provides information on the collective microscopic rearrangements that occur in disordered materials at nanometric and atomic scales [2]. In XPCS, the dynamics is investigated by analysing the temporal fluctuations in the intensity of speckles, which are interference patterns generated by coherent X-rays scattered from the material. By recording series of speckle patterns, XPCS captures these fluctuations, offering a direct measure of the particle dynamics. Studies on the atomic motion in hard materials have shown that the response of these systems to the high fluxes of X-ray beams employed in XPCS can vary significantly depending on the material family [3, 4, 5, 6, 7, 8, 9, 10].

While in crystalline [3] and glassy [4] alloys XPCS provides information on the atomic diffusion and collective motion in the material, unexpected results have been reported in vitreous silica and germania [4]. In these materials, the incident flux was found to induce atomic dynamics within the samples, likely due to radiolysis [4]. This process occurs when an X-ray photon is absorbed, creating an electron-hole pair that leads to atomic displacement and the formation of a long-lived exciton. The latter gives rise to bond breaking and further atomic displacement. Subsequent studies have expanded this investigation to a variety of materials, including borate glasses [5, 6, 7, 10], sodium silicates [10], lithium-borates [6], and chalcogenide glasses [8, 11]. In particular, the studies on the oxide glass LiBO2\mathrm{LiBO}_{2} [12] and the chalcogenide glass GeSe3\mathrm{GeSe}_{3} [11] show that after a certain absorbed dose, the system reaches a stationary state of yielding, which is photo-induced by the X-rays. The specific absorbed dose depends on the physicochemical characteristics of the material.

In general, these studies collectively demonstrate that X-rays can be employed not only to probe spontaneous electron density fluctuations but also to drive them, granting access to material properties that would otherwise be inaccessible using conventional techniques. The response of a glass to X-ray irradiation contains information on the structure and the microscopic relaxation processes active in the material [4]. Hence, X-rays have emerged as a powerful probe to study spontaneous and induced dynamical fluctuations in complex systems in response to other physical parameters, such as temperature and pressure [11].

Despite extensive studies on metallic and oxide glasses, no XPCS studies have been performed on glassy phase change materials (PCM). These materials (such as Ge15Sb85\mathrm{Ge}_{15}\mathrm{Sb}_{85}, Ge2Sb2Te5\mathrm{Ge}_{2}\mathrm{Sb}_{2}\mathrm{Te}_{5}, GeTe, and Ag-In-doped Sb2Te\mathrm{Sb}_{2}\mathrm{Te}) are a class of functional materials relevant for electronic and optical switching devices [13, 14]. PCMs can be rapidly and reversibly switched between amorphous and crystalline phases by Joule heating or optical pulses. Their strong property contrast between the two phases can be used for encoding data or manipulating optical beams for functional devices. Although both PCMs and conventional chalcogenide-based non-PCM alloys (e.g. Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85}, GeSe\mathrm{GeSe}, and GeSe2\mathrm{GeSe}_{2}) are commonly considered as covalent glasses, they differ remarkably in physical properties such as liquid fragility, structural relaxations, crystallisation kinetics, and glass forming abilities [13, 14].

In this work, we characterize the evolution of the collective atomic motion induced by intense X-ray radiation in two compositionally similar glassy alloys, a PCM Ge15_{15}Sb85_{85} and a non-PCM Ge15_{15}Te85_{85}, investigated in the vicinity of their glass transition temperature Tg. Ge15_{15}Sb85_{85} is a metallic-like alloy classified as a phase change material (PCM) due to its rapid and reversible glass-crystal switching properties, which are fundamental for non-volatile memory applications [15, 16, 17]. Ge15_{15}Te85_{85}, on the other hand, is a conventional chalcogenide glass with high amorphous stability and slow crystallization kinetics, extensively studied in both liquid and glassy phases [16, 18, 19, 20, 21, 22, 23, 24, 25, 26]. The strong interest in this composition is driven by its favourable properties for selector devices, such as a high crystallisation temperature that ensures a wider operating range, a fast threshold switching speed enabling rapid off-to-on transitions, and a low threshold voltage for efficient switching [22].

From the dynamical point of view, recent studies performed by some of the authors showed that PCMs, including Ge15_{15}Sb85_{85}, shows a pronounced secondary β-relaxation in the glassy state, faster that the main α-relaxation process, implying the presence of local fast atomic motion and loosely bonded regions in the material [27]. By contrast, when Sb is replaced by Te, the similar composition Ge15_{15}Te85_{85} shows virtually no β-relaxation, implying a rigid bonding network without local fast dynamics. The striking difference in relaxation behaviours is demonstrated to correlate with their markedly different glass forming abilities, thermal stabilities, and crystallisation kinetics [27]. While Ge15_{15}Te85_{85} is a good glass former with a relatively stable supercooled liquid region, Ge15_{15}Sb85_{85} can easily crystallize on approaching the glass transition temperature.

As we show below, this distinction also emerges from a different dynamical response to the X-ray beams. Specifically, Ge15_{15}Sb85_{85} exhibits an immediate transition to a yielding state, similar to GeSe3_3 [11], whereas Ge15_{15}Te85_{85} is more resistant, requiring two to three orders of magnitude higher absorbed doses to reach the same state.

Experimental methods

Samples

Glassy samples of Ge15_{15}Sb85_{85} and Ge15_{15}Te85_{85} were prepared through magnetron sputtering deposition of stoichiometric targets synthesised following the same procedure as described in a previous study [16]. A thick uniform layer of samples was deposited onto a 4-inch Si (100) wafer using DC 10 W, 10 sccm Ar flow, and 5×10−35 \times10^{-3} mbar Ar pressure. A steep-edge profilometer determines the sputtering rates of the target compositions to estimate the sputtering time needed for the desired layer thickness LL (3-8 micrometers) to match the absorption length μ\mu of the samples. The amorphous layers were separated from the Si substrate in the form of pieces or flakes, where single relatively large pieces (with a diameter of several millimetres) are selected for XPCS measurements. The nominal sample compositions were confirmed (within instrumental error ∼ 3%) twice using x-ray fluorescence (XRF) with a Rigaku NEX CG, first for the sputtering target and then for the as-deposited layers. The main properties of the two samples are summarised in (Table I) [13, 21, 27].

TgT_{\mathrm{g}} [K]FSDP [nm−1^{-1}]ρ\rho [g/cm3^{3}]LL [μ\mum]μ\mu [μ\mum]
Ge15_{15}Sb85_{85}50020.55.32338.5
Ge15_{15}Te85_{85}41219.65.5386.9

Table I.

Table 1. Calorimetric glass transition temperature TgT_{\mathrm{g}}, wavevector of the first sharp diffraction peak (FSDP) maximum, mass density ρ\rho, sample thickness, and X-ray absorption length at the energies used in the XPCS experiments for Ge15_{15}Te85_{85} and Ge15_{15}Sb85_{85} [13, 21, 27].

X-ray Photon Correlation Spectroscopy

XPCS experiments were carried out during two separate beamtimes at the beamline P10 of PETRA III synchrotron in Hamburg, Germany. In the first experiment, we measured the Ge15_{15}Te85_{85} composition. An X-ray beam of 8 keV was focused onto a spot size of full width at half maximum (FWHM) of 2.8×2.2 μm22.8 \times2.2\ \mu\mathrm{m}^{2} (H×\timesV) with a photon flux of ϕ0∼4.5×1010\phi_{0} \sim4.5 \times10^{10} ph/s, whereas in the second experiment, we measured the Ge15_{15}Sb85_{85} using a beam energy of 8.1 keV, with a spot size of 3.4×2.0 μm23.4 \times2.0\ \mu\mathrm{m}^{2} and a photon flux of ϕ0∼8×1010\phi_{0} \sim8 \times10^{10} ph/s. The samples were placed under vacuum, to prevent oxidation, on a Cu sample holder, whose temperature was controlled by a resistive heater and a Lakeshore temperature controller. Measurements were performed at a wave vector qq corresponding to the position of the maximum of the FSDP for both compositions. Ge15_{15}Te85_{85} was studied at temperatures T=[295,325]T = [295, 325] K (corresponding to T≃[0.7,0.8]TgT \simeq[0.7, 0.8]T_{\mathrm{g}}, see values in Table I), while Ge15_{15}Sb85_{85} was measured at T=[295,350,400]T = [295, 350, 400] K (T≃[0.6,0.7,0.8]TgT \simeq[0.6, 0.7, 0.8]T_{\mathrm{g}}). Additionally, the Ge15_{15}Te85_{85} composition was investigated as a function of the scattering vector qq in a range [2.18;27.37] nm−1[2.18; 27.37]\ \mathrm{nm}^{-1} at 295 K.

To explore the effect of the X-rays on the dynamics, the full beam intensity ϕ0\phi_{0} was attenuated through Si attenuators (att). The fractions of the transmitted beam intensity (i.e. the transmission TaT_{\mathrm{a}}) and the relative beam fluxes ϕ\phi used in the experiments are summarised in Table II.

Ge15_{15}Sb85_{85}Ge15_{15}Te85_{85}
attTaT_{\mathrm{a}}Φ\Phi [×1010\times10^{10} ph/s]attTaT_{\mathrm{a}}Φ\Phi [×1010\times10^{10} ph/s]
att = 0100%8att = 0100%4.5
att = 245%3.58att = 245%1.96
att = 422%1.73att = 422%1.26
att = 610%0.77att = 610%0.91
att = 85%0.4att = 85%0.39

Table II. Transmission percentages for the silicon attenuators used in the XPCS experiments for each material.

Each XPCS measurement was performed on a fresh sample spot to avoid any cumulative radiation effects. Different series of speckle patterns were collected in transmission geometry using an EIGER 4M detector positioned 1.8 m from the sample stage. Fluctuating speckle patterns are obtained when an amorphous system scatters a coherent X-ray, and they reflect the internal dynamics of the sample. Thus, the autocorrelation of the scattered intensity allows for the calculation of the microscopic dynamics, being

Cq(t1,t2)=⟨Ip(t1)Ip(t2)⟩p⟨Ip(t1)⟩p⟨Ip(t2)⟩p(1)C_q(t_1,t_2) = \frac{\langle I_p(t_1) I_p(t_2) \rangle_p}{\langle I_p(t_1) \rangle_p \langle I_p(t_2) \rangle_p} \tag*{(1)}

This function is called two-times correlation function (TTCF). Ip(t)I_p(t) is the intensity measured at pixel pp and time tt, and ⟨⋯ ⟩p\langle\cdots\rangle_p denotes an average over detector pixels at a given momentum transfer qq. Examples of TTCFs are presented in Figure 2c-Figure 2d). The main diagonal of the TTCFs corresponds to t1=t2t_1 = t_2 and represents the correlation of a frame at a given time with itself, while all the off-diagonals are the correlation of a frame at time t1t_1 (or t2t_2) with a subsequent frame at time t2t_2 (or t1t_1).

Temporal evolution of the dynamics in Ge15Sb85 and Ge15Te85

Figure 2. Temporal evolution of the dynamics at 0.7Tg0.7T_g for a fixed flux of the X-ray beam of 3.58×10103.58 \times10^{10} ph/s in Ge15_{15}Sb85_{85} a) and of 1.96×10101.96 \times10^{10} ph/s in Ge15_{15}Te85_{85} b). The dynamics remains stationary within the probed time window for the Ge15_{15}Sb85_{85} a), while physical aging in the Ge15_{15}Te85_{85} is signalled by the shift of the curves toward longer times with increasingly twt_w b). Two-times correlation functions measured in Ge15_{15}Sb85_{85} c) and Ge15_{15}Te85_{85} d) showing stable dynamics and aging, respectively.

From Cq(t1,t2)C_q(t_1,t_2), the autocorrelation function g2(t)g_2(t) of the intensity fluctuations can be extracted:

g2(q,t)=⟨Cq(t0,t0+t)⟩(2)g_2(q,t) = \langle C_q(t_0,t_0+t)\rangle \tag*{(2)}

where t0=t1t_0=t_1, t=t2−t1t=t_2-t_1, and ⟨⋯ ⟩\langle\cdots\rangle represents the average over time. This means that the TTCFs are a map of the time evolution of the microscopic dynamics. The g2(q,t)g_2(q,t) provides information on the microscopic dynamics through the determination of the intermediate scattering function (ISF) F(q,t)F(q,t) via the Siegert relation g2(q,t)=1+Ce∣F(q,t)∣2g_2(q,t)=1+C_e|F(q,t)|^2, where CeC_e is the experimental contrast related to the degree of coherence of the X-rays [28]. The ISF is the time-dependent density-density correlation function and quantifies the extent to which the atomic configuration at a given moment resembles its configuration at a later time tt.

Results

We model the g2(q,t)g_2(q,t) with a Kohlrausch-Williams-Watts (KWW) function [29]

g2(q,t)=y0+Ce−2(tτ)β(3)g_2(q,t) = y_0 + C e^{-2\left(\frac{t}{\tau}\right)^\beta} \tag*{(3)}

where C=Ce∣fq∣2C=C_e|f_q|^2, with fqf_q being the non-ergodicity factor and CeC_e the experimental contrast, τ\tau is the relaxation time, β\beta the degree of exponentiality of the decay, and y0y_0 accounts for the baseline [28]. All the parameters depend on the exchanged wavevector qq, the temperature TT, and the absorbed dose DD. The total radiation dose absorbed by the sample during an exposure time Δt\Delta t is calculated as:

D=ϵmFΔt(4)D = \frac{\epsilon}{m}F\Delta t \tag*{(4)}

where ϵ\epsilon is the X-ray energy and FF is the fraction of the incident beam, ϕ\phi, absorbed by the sample given by F=ϕ(1−Tr)F=\phi(1-T_r), with TrT_r being the sample transmission. As anticipated in the experimental section, the value of the incident flux ϕ\phi was varied by attenuating the full beam ϕ0\phi_0 through Si attenuators with transmissions TaT_a reported in (Table II). The incident fluxes employed in the experiments result in ϕ=ϕ0Ta\phi=\phi_0T_a. The dose also depends on the mass mm within the scattering volume VsV_s: m=ρVsm=\rho V_s, where Vs=ALV_s=AL (with AA as the focal spot area and LL the sample thickness) and ρ\rho is the mass density.

Consistent with previous studies on other chalcogenide glasses [8, 11], our data show that the probed atomic motion is also triggered by the X-ray irradiation in the studied compositions. This effect is observed at all investigated temperatures and at the probed qq no appreciable structural damage is detected (see Figure (SM1)). However, within the limits of our experimental resolution, we cannot determine whether this apparent stability reflects a genuine resistance to structural damage, as reported for other glassy systems [4], or whether minor structural damage occurs below our detection threshold or in a different qq range.

An example of the probed dynamics is shown in Figure 1 where we report auto-correlation functions measured with XPCS at different values of the X-ray flux at T=0.7TgT=0.7T_g, for both Ge15Sb85\mathrm{Ge}_{15}\mathrm{Sb}_{85} (1a) and Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} (1b). The long time-decay of the curves corresponds to the characteristic time associated with the probed dynamics. In both cases, this decay shifts to larger times with decreasing flux (from blue to green curves), indicating that the dynamics can be tuned by varying the intensity of the X-ray beam. The black dashed lines are the best fit to the experimental data according to equation (3). To ensure consistency between data collected at different X-ray fluxes, the comparison of the curves was performed by evaluating the g2(t)g_2(t) considering a fixed absorbed dose of 38 GGy for each composition.

Auto-correlation functions of the intensity fluctuations measured with XPCS at 0.7$T_g$ in a) Ge$_{15}$Sb$_{85}$ and b) Ge$_{15}$Te$_{85}$ at varying beam attenuations

Figure 1. Auto-correlation functions of the intensity fluctuations measured with XPCS at 0.7TgT_g in a) Ge15_{15}Sb85_{85} and b) Ge15_{15}Te85_{85} at varying beam attenuations. The curves of each sample are obtained by ensuring a total absorbed dose of 38 GGy for both samples. The dashed black lines represent the best fits to the KWW model. Each curve has been measured on a fresh sample position.

Although the dynamics of both glasses clearly depends on the X-ray flux, the two materials exhibit distinct responses to X-ray exposure. This is particularly evident from the temporal evolution of their relaxation processes at a fixed beam intensity as shown in Figure 2. Here, the curves were evaluated by averaging the TTCFs according to Eq. (2) over time windows Δt\Delta t, centred at waiting time tw=Δt(n+1/2)t_{\mathrm{w}} = \Delta t(n + 1/2) with n∈[0,3]n \in[0, 3], and Δt=900\Delta t = 900 s for the Ge15_{15}Sb85_{85} and Δt=700\Delta t = 700 s for the Ge15_{15}Te85_{85}. In this context, the waiting time twt_{\mathrm{w}} reflects the time elapsed since the onset of each measurement. As depicted in Figure 2a, g2(q,t)g_2(q,t) curves acquired at different waiting times overlap each other, indicating that the dynamics of the Ge15_{15}Sb85_{85} remains stationary over time. This is also evident from the TTCF, shown in Figure 2c, where a constant intensity profile is observed along the main diagonal from the bottom left corner to the upper right corner. In contrast, Ge15_{15}Te85_{85} exhibits a progressive slowing down of its dynamics over time. As shown in Figure 2b, the g2(q,t)g_2(q,t) curves shift towards longer decay times as the waiting time progresses from 350 s to 2450 s, highlighting this effect. The presence of this process is also confirmed by the broadening of the intensity along the main diagonal of the TTCF of Ge15_{15}Te85_{85} reported in Figure 2d.

The slowing down of the dynamics is a spontaneous process in glasses held at temperatures close to TgT_{\mathrm{g}}, known as physical aging [30]. As out-of-equilibrium systems, glasses naturally evolve over time towards more energetically stable configurations, leading to an increase in relaxation time. Under standard conditions, this aging process is thermally driven, as the system progressively explores its energy landscape. In our experiments, however, the observed slowing down is not solely spontaneous but is induced by the X-ray beam itself. The beam not only probes and stimulates atomic dynamics but also promotes a temporal evolution of the system. This effect may be linked to stress release occurring under irradiation or to local annealing processes initiated by the beam. These aspects will be discussed in greater detail later in this section.

The contrasting temporal evolutions of the dynamics in the two compositions imply a different response of the two systems to the accumulated dose DD. To investigate this aspect, we express the relaxation time τ\tau, obtained from XPCS measurements and KWW fits (3), in terms of the absorbed dose DτD_{\tau} required to induce the observed dynamics:

Dτ=ϵmϕ0Ta(1−Tr)τ(5)D_{\tau} = \frac{\epsilon}{m}\phi_{0}T_{a}(1-T_{r})\tau \tag*{(5)}

This formulation accounts for the interaction between the material and the X-rays. In the case of a process purely induced by irradiation [4], the relaxation time τ\tau is expected to scale inversely with the incident X-ray flux ϕ\phi, meaning that halving the beam flux (Ta=50%T_a = 50\%) doubles τ\tau. Consequently, the dynamics depend only on the absorbed dose, independent of flux variations. This expression also allows the comparison of the relaxation dynamics between different samples compositions and experimental parameters. To explore the distinct responses of Ge15_{15}Sb85_{85} and Ge15_{15}Te85_{85} to X-ray exposure, Figure 3 and Figure 4 show the absorbed dose in relaxation time DτD_{\tau} and the KWW exponent β\beta as functions of the accumulated dose DD, for different attenuator transmissions Ta,iT_{a,i}. This representation enables us to assess the dependence of the dynamical parameters on the absorbed dose and provides insight into the nature of the interaction, allowing us to distinguish between dynamics that are purely beam-induced and those influenced by additional mechanisms.

Evolution of the dynamics in Ge15Sb85 as a function of the absorbed dose

Figure 3. Evolution of the dynamics in Ge15_{15}Sb85_{85} as a function of the absorbed dose. a-b) At 0.7TgT_{\mathrm{g}} and c-d) at 0.8TgT_{\mathrm{g}}, the DτD_{\tau} and the β\beta tend to collapse into a master curve when plotted versus Dose. Both parameters are almost independent of the applied dose.

Evolution of the dynamics in Ge15Te85

Figure 4. a-b) Evolution of the dynamics in Ge15_{15}Te85_{85} as a function of the waiting time twt_w, at 0.7Tg0.7T_g. DτD_{\tau} and β\beta collapse into a master curve when plotted versus twt_w, suggesting the dynamics is primarily influenced by intrinsic relaxations active in the material. While the beam does not irradiate the sample spot, both DτD_{\tau} and β\beta tend to recover the initial state (insets of a-b)). c-d) At 0.8Tg0.8T_g, DτD_{\tau} and β\beta collapse in a master curve by plotting them versus the Dose. a-c) As the waiting time or the deposited dose increases, the dynamics slows down, suggesting a final plateau. b-d) With increasing the absorbed dose, the shape parameter β\beta decreases from β>1\beta> 1, signature of compressed correlation functions, to β<1\beta< 1, typical of stretched exponential decays.

Let’s first focus on the dynamical behaviour of Ge15_{15}Sb85_{85}. The dependence of the XPCS parameters τ\tau and β\beta on the accumulated dose DD is presented in Figure 3 at two temperatures, 0.7Tg0.7T_g or 0.8Tg0.8T_g. Panels a) and c) show the dose DτD_{\tau} accumulated in a relaxation time τ\tau at the two temperatures, while the dose evolution of the β\beta parameter is plotted in panels b) and d). Different symbols and colours refer to different beam fluxes. The probed dynamics is beam-induced, since the DτD_{\tau} parameter is independent of the beam flux for almost all the datasets, implying that the decorrelation time τ\tau is inversely proportional to the beam flux. Overall, a good collapse of points corresponding to different beam fluxes is observed in all panels. There is, however, one dataset that follows a different trend. The blue circles at T=0.7TgT = 0.7T_{\mathrm{g}}, corresponding to full beam on the sample, exhibit a lower DτD_{\tau} and a higher β\beta compared to the other points, likely due to local beam-induced heating that slightly increases the effective temperature of the probed spot. Apart from this dataset, the other points suggest that the system is in a yielded state at all the investigated doses, because the DτD_{\tau} parameter is almost constant during the measurement and the stretching parameter β\beta is close to one, as in the case of LiBO2_2 [12] and GeSe3_3 [11].

We now turn to the dynamical properties of Ge15_{15}Te85_{85}. As discussed before, a purely beam-induced dynamics is expected to scale with the accumulated absorbed dose DD: if the dynamics is entirely governed by the X-rays, parameters such as the absorbed dose in XPCS relaxation times, DτD_{\tau}, and the shape parameter β\beta should collapse onto universal curves when plotted as a function of DD. However, for Ge15_{15}Te85_{85} at 0.7TgT_{\mathrm{g}}, this universal scaling is not observed (see Supplementary Figure (S2)). Although the beam induces the dynamics, DτD_{\tau} and β\beta exhibit significant variations with the incident flux when plotted against DD, suggesting that the absorbed dose alone does not fully describe the system’s evolution.

To further clarify this behaviour, we plot DτD_{\tau} and β\beta as functions of the experimental waiting time twt_w, defined as the time elapsed since the start of exposure. Interestingly, when expressed in terms of twt_w, the data collapse onto master curves across all fluxes (Figure 4a-Figure 4b)), revealing a universal, time-dependent evolution of the dynamics. To our knowledge, such a collapse with respect to the waiting time, rather than the absorbed dose, has not been previously reported in studies of beam-induced dynamics. This observation indicates that, while X-ray irradiation initiates atomic motion, the evolution at 0.7Tg0.7T_g remains governed primarily by intrinsic relaxation processes with timescales dictated by the material itself, rather than solely by the accumulated dose. This phenomenon can be rationalised by considering that 0.7Tg0.7T_g for Ge15_{15}Te85_{85} corresponds to room temperature. It is therefore highly likely that the sample retains residual stresses from its preparation, which are progressively released under X-ray exposure. The stress relaxation drives the evolution of the dynamics, dominating the beam-induced effect.

Further evidence supporting this interpretation comes from “beam-of” experiments (Figure 4 insets), where the X-ray beam was switched off for extended periods (see SM and Figure S3 for further details). For short interruptions (10 minutes), stress relaxation continues, reflecting intrinsic structural relaxation mechanisms triggered by the X-rays. Differently, for longer interruptions (1 hour), both DτD_{\tau} and β\beta reverted toward their initial values, indicating a recovery of the original state in the absence of irradiation. Such a reversible response upon switching off the beam has never been observed in non-stationary beam-induced dynamics so far, and supports the absence of structural damage within the experimental doses.

Dynamics evolution plots for Ge15Te85 with the beam switched off

Figure S3. Dynamics evolution at T=0.7TgT = 0.7T_{\mathrm{g}} of Ge15_{15}Te85_{85} switching off the beam for 600 s and for 3600 s. a) The relaxation time τ\tau increases in time, but while the beam is off it tends to go back to the initial values. b) the shape parameter β\beta decreases in time, but while the beam is off it tends to go back to the initial values.

This hypothesis of room temperature stress-release is confirmed by the fact that at 0.8Tg0.8T_{\mathrm{g}}, both DτD_{\tau} and β\beta exhibit a universal scaling with the absorbed dose DD, regardless of the incident flux (Figure 4c-Figure 4d)). This indicates that, at 0.8Tg0.8T_{\mathrm{g}}, the dynamics depends solely on the X-ray flux and is thus purely beam-induced. Remarkably, a modest 10% increase in temperature is sufficient to eliminate the stress-driven contributions observed at 0.7Tg0.7T_{\mathrm{g}}, placing the system entirely under the control of irradiation-driven dynamics.

Next, we examine how the dynamical parameters of the two compositions depend on the absorbed dose DD (or on the waiting time twt_{\mathrm{w}}) at both 0.7Tg0.7T_{\mathrm{g}} and 0.8Tg0.8T_{\mathrm{g}}. To do that, we refer to Figure 3 and (4) for Ge15Sb85Ge_{15}Sb_{85} and Ge15Te85Ge_{15}Te_{85}, respectively. As anticipated, the Ge15Sb85Ge_{15}Sb_{85} glass exhibits stationary dynamics and a KWW exponent β≲1\beta\lesssim1 at both investigated temperatures. These features correspond to a local state of photo-induced yielding by the X-rays, resembling the previous studies on other chalcogenide and oxide glasses [8, 11, 12]. In materials science, yielding refers to the transition from an elastic response, where deformation is reversible, to a plastic state, where the material undergoes irreversible atomic rearrangements. In glasses, yielding is particularly subtle since plastic events appear at very small scales and are associated with localised atomic rearrangements known as shear transformation zones [12]. Martinelli et. al. [12] showed that X-ray irradiation acts like an external force, generating tiny defects that trigger these hidden structural rearrangements, effectively pushing the material into a plastic, irreversible state at the microscopic scale. In the case of the chalcogenide glass GeSe3GeSe_{3}, the yielding is induced within 100 s of exposure at ambient temperature [11], while the Ge15Sb85Ge_{15}Sb_{85} composition reaches this stationary state within the exposure time of 1 s at the probed temperatures of 350 K and 400 K (0.7Tg0.7T_{\mathrm{g}} and 0.8Tg0.8T_{\mathrm{g}}, respectively). This difference in time scales is likely the consequence of the larger thermal fluctuations at the probed temperatures, which facilitate the occurrence of local yielding at a lower dose. In fact, in agreement with previous works, we also observe a transitory state with a weak aging at ambient temperature (0.6Tg0.6T_{\mathrm{g}}) in Ge15Sb85Ge_{15}Sb_{85} accompanied by a transition in the KWW parameter β\beta from compressed (≥1\geq1) to stretched (≤1\leq1) and the reaching of a stationary regime around 40 GGy (see S4). As discussed in detail in Ref. [12], these are the typical characteristics of the induced yielding transition.

On the other hand, the Ge15Te85Ge_{15}Te_{85} exhibits a strong dependence on the waiting time and absorbed dose at both temperatures, as evident from Figure 4. We see that neither the absorbed dose associated with the relaxation time DτD_{\tau} nor the β\beta parameter reaches a complete plateau. Only the DτD_{\tau} measured at Ta,6T_{\mathrm{a,6}} and the β\beta at Ta≥28%T_{\mathrm{a}} \geq28\% at 0.7Tg0.7T_{\mathrm{g}} suggest the proximity to a stationary regime. To understand whether this behaviour corresponds to the onset of the photo-induced yielding transition, we measured the wave-vector q\mathrm{q} dependence of the dynamics (and thus of the parameters DτD_{\tau} and β\beta) as a function of increasing absorbed dose (see Figure 5). Following Ref. [12], during the transition to local yielding, one should also observe a crossover from a 1/q1/q wave-vector dependence of τ\tau and an almost q-independent evolution of β\beta, to a liquid-like behaviour.

Wave vector dependence of relaxation time and KWW beta parameter

Figure 5. Wave vector dependence of τ\tau a) and β\beta b) at 0.7Tg0.7T_{\mathrm{g}} of Ge15_{15}Te85_{85} for three selected doses: D=7D = 7 GGy (blue circles), D=17D = 17 GGy (red squares) and D=74D = 74 GGy (yellow triangles). The vertical dashed and dashed-dotted black lines represent, respectively, the position of the pre-peak and the main peak. a) The inset shows the dependence on the dose of the power-law exponent α\alpha describing the linear dependence of τ\tau at low qq-values.

At low qq, a linear trend emerges in the log-log plot of DτD_{\tau} versus qq (Figure 5a), indicating a power-law relationship between the relaxation time and the wavevector. By analysing this dependence in the range [2, 3, 4, 5, 6, 7] nm−1^{-1}, we can access the nature of the particle dynamics and how it evolves with absorbed dose. Fitting the data with the expression Dτ(q)=(bqα)−1D_{\tau}(q) = (bq^{\alpha})^{-1}, where the exponent α\alpha characterizes the type of motion, we obtain α=1.7±0.4\alpha= 1.7 \pm0.4, independently of dose (see inset of Figure 5a). This value is larger than the one reported for LiBO2_2 glass, where α\alpha decreases from ∼ 1.1 to ∼ 0.3 with increasing dose. This trend was interpreted as a transition in the nature of collective motion, from ballistic-like dynamics to a regime where defect creation and annihilation are balanced, similar to a steady-state condition observed in liquids [12]. Differently, in the Ge15_{15}Te85_{85}, the larger value of α\alpha suggests the presence of diffusive motion at low qq-s, with a negligible dose dependence. Additional data at lower qq values, well separated from structural contributions, are required to confirm this behaviour.

At high qq (greater than 7 nm−1^{-1}), we observe an overall increase in DτD_{\tau} and a concurrent decrease in β\beta with increasing dose DD. At all doses, DτD_{\tau} presents a peak at ∼ 10.5 nm−1^{-1}, which mirrors the corresponding prepeak of the static structure factor S(q)S(q) [21]. At high doses, a peak arises at ∼ 19.6 nm−1^{-1}, corresponding to the main first sharp diffraction peak (FSDP). Even the KWW β\beta parameter exhibits features resembling the evolution of the S(q)S(q), less evident at low doses but pronounced at higher doses. The resemblance between the qq-dependence of τ\tau and β\beta, and that of the S(q)S(q) is reminiscent of the phenomenon known as de Gennes narrowing, which takes the name from the narrowing at qq-values corresponding to the FSDP of the width of the quasi-elastic contribution in neutron scattering experiments on liquids [31]. Given that the half-width at half maximum of the quasi-elastic peak is inversely proportional to structural relaxation time, Γ(q)∝1/τ\Gamma(q) \propto1/\tau, this results in an increase in τ\tau near structure factor peaks, implying that the most probable structural configuration is also the most stable. Therefore, the observed qq-dependence in Ge15_{15}Te85_{85} suggests that the sample volume irradiated by the X-rays displays a liquid-like behaviour and thus a local yielding as in the LiBO2_2. This interpretation is supported by the transition of β\beta from values greater than 1, typical of a stress-driven dynamics of glasses well below TgT_{\mathrm{g}} [32], to values lower than 1, fingerprint of liquid-like heterogeneous relaxation dynamics [33, 34].

We note also that at low doses (below 0.5 GGy), the de Gennes narrowing is entirely absent for LiBO2_2, in both τ\tau and β\beta [12]. In our case, instead, the de Gennes narrowing effect is already present at the 10.5 nm−1^{-1} pre-peak in τ\tau (DτD_{\tau}), even in the lowest probed dose regime (below 7 GGy). In particular, at low doses only the signature of the pre-peak around ∼\sim 10.5 nm−1^{-1} is observed in the evolution of DτD_{\tau} with qq, while at high doses, the signature of the main peak around ∼\sim 19.6 nm−1^{-1} arises as well. Such a strong dynamical signature corresponding to the weak pre-peaks in structure factors is a general feature of glass-forming liquids, which has been associated with the influence of dynamic heterogeneity phenomena at the single and many particle level [35, 36].

Discussion and conclusion

Our results reveal a strong correlation between the local atomic structure and the dynamical response to intense X-ray fluxes in two compositionally similar glasses, Ge15_{15}Sb85_{85} and Ge15_{15}Te85_{85}. Consistent with previous findings in other oxides and chalcogenides [8, 11, 12], the X-rays can induce a local state of yielding in both compositions, where the amorphous solids present liquid-like features in the irradiated zone. The propensity of this phenomenon, however, strongly depends on the chemical and physical details of the material.

In the PCM Ge15_{15}Sb85_{85}, our data exhibit an almost instantaneous photo-induced yielding transition at 0.7TgT_{\mathrm{g}} and 0.8TgT_{\mathrm{g}}. This rapid response is favoured by thermal motion, since at room temperature the transition requires ∼\sim 400 s of continuous irradiation to occur (see SM). This is in agreement with what has been observed in GeSe3_3 [11] at ambient temperature, where yielding occurs in ∼\sim 100 s. The faster transition at room temperature in GeSe3_3 is likely due to the Se being lighter than Sb, a fact supported by a faster relaxation time of τ∼2\tau\sim2 s in GeSe3_3 in comparison to τ∼100\tau\sim100 s in Ge15_{15}Sb85_{85} for similar doses. Increasing the temperature to 350 K (0.7TgT_{\mathrm{g}}) enhances atomic mobility within Ge15_{15}Sb85_{85}, enabling the system to reach the yielding state almost immediately.

Unlike Ge15_{15}Sb85_{85} and GeSe3_3, Ge15_{15}Te85_{85} does not completely reach the yielding state within the experimental timeframe, neither at 0.7TgT_{\mathrm{g}} (which corresponds to room temperature) nor 0.8TgT_{\mathrm{g}}. This can be explained by the fact that Te atoms are heavier than Se or Sb atoms, reducing atomic mobility and resulting in a greater resistance to structural change upon irradiation with X-ray beams. This is consistent with the absence of visible structural changes on the FSDP, which are instead clear in GeSe3_3. Moreover, Ge15_{15}Te85_{85} exhibits recovery effects upon beam switching off: once the X-ray beam is turned off, the system begins to recover toward its initial dynamical state and the longer the absence of irradiation, the closer the dynamical parameters return to their original values (see insets of Figure 4a-Figure 4b)). In contrast, the GeSe3_3, once it has reached the yielding state, shows a stable dynamics: during the absence of the beam, the system remains in that state. It is important to emphasise that these differing responses cannot be explained solely by considering the type of chemical bonding. Both GeSe3_3 and Ge15_{15}Te85_{85} are covalent glasses with good glass-forming ability and slow crystallisation kinetics. Nevertheless, they show distinct X-ray-induced dynamics, suggesting that factors beyond bonding, such as atomic mass, network topology, and local structural heterogeneity, play a crucial role in determining the onset and nature of yielding under irradiation.

The faster photo-induced atomic motion in the PCM Ge15_{15}Sb85_{85} and the higher sensitivity to the X-rays induced yielding is likely the consequence of its locally loosely bonded microscopic structure and the presence of a pronounced secondary β\beta-relaxation which promotes local mobility at the atomic scale [27]. This mobility is also confirmed by the fact that the dynamics does not scale completely with the dose DD, strengthening the idea of the presence of an intrinsic dynamic process already active at the probed temperatures. By contrast, the rigid covalently bonded network in Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} is less prone to changes of the local atomic environment, which reflects in longer relaxation times at the atomic scale and a pure beam-induced dynamics but at ambient temperature. At this temperature, the scaling law is not fulfilled, probably due to the presence of some additional stresses related to the sample preparation, which sum up with the irradiation-induced yielding transition.

Interestingly, we note that the Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} exhibits a dependence on the absorbed dose similar to that reported in the oxide glass LiBO2\mathrm{LiBO}_{2} [12]. In both systems, increasing the absorbed dose leads to a transition in the KWW β\beta shape parameter from a compressed (β≥1\beta\ge1) to a stretched (β≤1\beta\le1) regime, which is accompanied by the emergence of a liquid-like wave-vector dependence of the dynamics at large doses. As previously discussed, this behaviour can be attributed to the transition from a purely beam-induced stress-driven motion to a local yielding transition of the material under X-ray irradiation [12].

It is, however, important to underline the main differences between LiBO2\mathrm{LiBO}_{2} and Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85}. One crucial factor influencing beam-induced dynamics is the X-ray absorption. Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} consists of heavy atoms, which absorb X-rays more efficiently than the lighter elements present in LiBO2\mathrm{LiBO}_{2}. This difference is reflected in the experimental conditions: the thickness of the Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} sample was comparable to its attenuation length at 8 keV (∼8 μm\sim8\ \mu\mathrm{m}), ensuring substantial absorption, whereas LiBO2\mathrm{LiBO}_{2} was significantly thinner (125 μm\mu\mathrm{m}) relative to its attenuation length at 8.4 keV (680 μm\mu\mathrm{m}). Consequently, the dose range explored for Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} reached tens of GGy, while for LiBO2\mathrm{LiBO}_{2}, it remained at a few GGy. Despite this large difference in the absorbed dose, both materials exhibit a similar evolution of the KWW β\beta parameter and τ(q)\tau(q) as a function of the absorbed dose. This suggests that atomic rearrangements in Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} require a significantly higher number of electron-hole pairs, likely due to its heavier atomic composition. This interpretation aligns with the slower dynamics observed in Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} (on the order of hundreds of seconds) compared to LiBO2\mathrm{LiBO}_{2} (∼20 s\sim20\ \mathrm{s}), despite its higher absorbed dose.

In conclusion, our XPCS study reveals distinct beam induced dynamical responses in glassy Ge15Sb85\mathrm{Ge}_{15}\mathrm{Sb}_{85} and Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} near their glass transition temperatures. Both systems are influenced by X-ray irradiation, yet their responses differ markedly despite their similar composition. Ge15Sb85\mathrm{Ge}_{15}\mathrm{Sb}_{85} undergoes a rapid transition to a stationary yielding regime within the exposure time, characterized by a KWW parameter β≲1\beta\lesssim1. By contrast, Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} exhibits strong dose-dependent evolution, including aging and a progressive shift of the shape parameter β\beta from compressed (β>1\beta> 1) to stretched (β<1\beta< 1). This transition is consistent with the emerging of liquid-like collective motion, as supported by de Gennes narrowing observed in the wavevector qq-dependence of the dynamics. This glass does not reach a stationary regime within experimental timescales, indicating that local yielding requires thousands of seconds. This resistance to photoinduced yielding in Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} is consistent with its heavier atomic constituents, which enhance structural rigidity and demand much higher deposited doses to induce local transformations. These findings align with its higher amorphous thermal stability and slow crystallisation kinetics [22], in contrast to Ge15Sb85\mathrm{Ge}_{15}\mathrm{Sb}_{85} which exhibits rapid crystallisation kinetics and less covalent bonding [16]. Overall, neither chemical composition alone nor bonding character alone dictates the X-ray induced response of the non-PCM Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} and PCM Ge15Sb85\mathrm{Ge}_{15}\mathrm{Sb}_{85}; instead, it emerges from the interplay between irradiation effects and intrinsic relaxation pathways. Finally, we note that such beam-induced dynamics may provide a novel means of probing β\beta-relaxations in PCMs below TgT_\mathrm{g}. Previous studies have shown that PCMs such as Ge2Sb2Te5\mathrm{Ge}_2\mathrm{Sb}_2\mathrm{Te}_5, GeTe, and AIST exhibit pronounced excess wings in the loss modulus associated with β\beta-relaxations, in stark contrast to those chalcogenide non-PCMs whose excess wings are vanishingly small [37].

Acknowledgements

We gratefully acknowledge PETRA III synchrotron (Hamburg, Germany) for providing beamtime, including experiments carried out at P10 under the LTP project II-20230020 EC. This project received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Grant Agreement No 948780).

Author Contributions

G.B. and B.R. conceived and designed the project. G.B., A.C., I.F., T.F., J.M., A.R., B.R., M.S., S.W., and F.W. performed the experiments. J.M., R.E., A.P. and R.E. carried out sample fabrication. I.F. analysed the data. G.B., B.R. and I.F. wrote the manuscript with contributions from all the authors. All the authors contributed to the discussion of the results, reviewed the data and provided feedback on the manuscript.

Supplementary information to: X-ray photo-induced atomic motion in Phase Change Materials and conventional covalent chalcogenide glasses

I. Festi; A. Cornet; T. Fujita; J. Moesggard; A. Ronca; J. Shen; M. Sprung; S. Wei; F. Westermeier; R. Escalier; A. Piarristeguy; G. Baldi; B. Ruta

Methods

Absorbed dose calculation

The dose rate of X-ray absorption from a material is

Drate=ϵemϕ0Ta(1−Tr),(6)D_{rate} = \frac{\epsilon e}{m}\phi_0 T_a(1-T_r), \tag*{(6)}

where ϵ\epsilon and ϕ0\phi_0 are respectively the energy and the flux of the X-rays, ee is the electron charge, and TaT_a is the transmission of the Si attenuators employed to attenuate the beam intensity. The dose rate also depends on the absorbed fraction of the beam

1−Tr=1−e−Lμ,(7)1-T_r = 1-e^{-\frac{L}{\mu}}, \tag*{(7)}

where LL is the sample thickness, while μ\mu is the absorption length of the material at the energy ϵ\epsilon. Finally, the dose rate depends on the mass mm within the scattering volume VsV_s, according to m=ρVsm=\rho V_s, being ρ\rho the mass density of the sample. The volume is found considering an elliptical beam cross section, whose semiaxes correspond to the FWHM/2ln⁡2\sqrt{2\ln2} values.

The absorbed dose is found considering the exposure amount of time Δt\Delta t of the sample spot to the beam

D=DrateΔt(8)D = D_{rate}\Delta t \tag*{(8)}

Data treatment of τ\tau and β\beta vs qq

The qq-dependence of the dynamics of Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85} shown in Figure 5 of the main article was measured at 0.7Tg0.7T_{\mathrm{g}} with attenuator 2. Due to the lower statistics at high qq-values, we applied a two-point averaging to the β\beta parameter to improve data clarity and reduce statistical fluctuations. This averaging was not applied to the τ\tau relaxation time, as it is a significantly more robust parameter and did not require smoothing.

Beam damage

The scattered intensity from the samples showed no significant variation over the duration of the XPCS measurements (Figure S1), suggesting that either the underlying structure remained largely unchanged by the X-ray beam, indicating no appreciable structural damage, or that any structural modifications occurring were not detectable through the analysis of the mean scattered intensity alone. This observation implies that the observed dynamical changes are likely not a direct consequence of gross structural degradation of the amorphous network within the probed timeframe.

Average scattered intensity collected by the detector for each frame

Figure S1. The average scattered intensity collected by the detector for each frame, normalized to the relative beam intensity at various attenuation levels. As time passes, no appreciable structural damage is occurring.

Ge15Te85\mathrm{Ge}_{15}\mathrm{Te}_{85}

T=0.7TgT=0.7T_{\mathrm{g}} versus absorbed dose DD

At 0.7Tg0.7T_{\mathrm{g}}, if we plot DτD\tau and β\beta versus the absorbed dose DD we observe a failure in the collapse into a universal master curve, differently from what was observed at 0.8Tg0.8T_{\mathrm{g}}.

Dynamics evolution plots for Ge15Te85 at T = 0.7T_g

Figure S2. Dynamics evolution at T=0.7TgT = 0.7T_{\mathrm{g}} of Ge15_{15}Te85_{85}. a) DτD_{\tau} and b) β\beta fail to collapse into a master curve when plotted versus DD.

Recovery of dynamics after beam interruption

To further investigate the nature of the beam-induced dynamics in Ge15_{15}Te85_{85}, we conducted experiments where the X-ray beam was switched off after an initial measurement on a single sample spot, and then turned back on for a subsequent measurement on the same location. Our findings revealed that during the period when the beam was off, the dynamics of the material tended to revert towards its initial state (Figure S3).

Ge15_{15}Sb85_{85}

T=0.6Tg_{\mathrm{g}}

While Ge15_{15}Sb85_{85} generally exhibits stationary dynamics at 0.7Tg_{\mathrm{g}} and 0.8Tg_{\mathrm{g}}, measurements at 0.6Tg_{\mathrm{g}} reveal some hints of aging, specifically at the highest beam intensities. At lower fluxes, the relaxation time τ\tau remains stable. At 0.6Tg_{\mathrm{g}}, we also observe a decrease in the stretching exponent β\beta from approximately 1.1 to 0.9, a trend reminiscent of Ge15_{15}Te85_{85} but significantly less pronounced (where β\beta shifts from ∼1.3\sim1.3 to ∼0.6\sim0.6). This transient behaviour in Ge15_{15}Sb85_{85} at room temperature (0.6Tg0.6T_{\mathrm{g}}) is likely due to residual stresses introduced during sample preparation. These stresses may make the material more susceptible to the X-ray beam, a phenomenon far more pronounced in Ge15_{15}Te85_{85}, before dissipating at higher temperatures where the intrinsic stability of the glass dominates.

Dynamics evolution at T = 0.6Tg of Ge15Sb85

Figure 6. Dynamics evolution at T = 0.6Tg0.6T_{\mathrm{g}} of Ge15_{15}Sb85_{85}. a) the relaxation time τ\tau shows a constant behaviour, except for very high beam intensities. b) β\beta decreases from ∼1.1\sim1.1 to ∼0.9\sim0.9.

References

  1. [1]José Baruchel, Jean-Louis Hodeau, and Mogens S. Lehmann. Neutron and synchrotron radiation for condensed matter studies. Volume I, Theory, instruments and methods. Springer, 1993.
  2. [2]Anders Madsen, Andrei Fluераsu, and Beatrice Ruta. Structural dynamics of materials probed by x-ray photon correlation spectroscopy. In Eberhard Jaeschke, Shaukat Khan, Jochen R. Schneider, and Jerome B. Hastings, editors, Synchrotron Light Sources and Free-Electron Lasers: Accelerator Physics, Instrumentation and Science Applications, pages 1–21. Springer International Publishing, 2015.
  3. [3]Michael Leitner, Markus Stana, Manuel Ross, and Bogdan Sepiol. Acceleration of atomic dynamics due to localized energy depositions under x-ray irradiation, arXiv, arXiv:1510.01918 2015.arxiv.org/abs/1510.01918
  4. [4]B. Ruta, F. Zontone, Y. Chushkin, G. Baldi, G. Pintori, G. Monaco, B. Rufflé, and W. Kob. Hard x-rays as pump and probe of atomic motion in oxide glasses. Scientific Reports, 7(1):3962, Jun 2017.DOI
  5. [5]G. Pintori, G. Baldi, B. Ruta, and G. Monaco. Relaxation dynamics induced in glasses by absorption of hard x-ray photons. Phys. Rev. B, 99:224206, Jun 2019.DOI
  6. [6]Dallari, Pintori, Baldi, Martinelli, Ruta, Sprung, and Monaco. X-rays induced atomic dynamics in a lithium borate glass. Condensed Matter Physics, 22(4):43606, November 2019.
  7. [7]Katharina Holzweber, Christoph Tietz, Tobias Michael Fritz, Bogdan Sepiol, and Michael Leitner. Beam-induced atomic motion in alkali borate glasses. Phys. Rev. B, 100:214305, Dec 2019.
  8. [8]Jianheng Li, Meera Madhavi, Spencer Jeppson, Louie Zhong, Eric M. Dufresne, Bruce Aitken, Sabyasachi Sen, and Roopali Kukreja. Observation of collective molecular dynamics in a chalcogenide glass: Results from x-ray photon correlation spectroscopy. The Journal of Physical Chemistry B, 126(28):5320–5325, 2022. PMID: 35730616.DOI
  9. [9]Francesco Dallari, Alessandro Martinelli, Federico Caporaletti, Michael Sprung, Giacomo Baldi, and Giulio Monaco. Stochastic atomic acceleration during the x ray-induced fluidization of a silica glass. Proceedings of the National Academy of Sciences, 120(2):e2213182120, 2023.
  10. [10]E. Alfinelli, F. Caporaletti, F. Dallari, A. Martinelli, G. Monaco, B. Ruta, M. Sprung, M. Zanatta, and G. Baldi. Amorphous-amorphous transformation induced in glasses by intense x-ray beams. Phys. Rev. B, 107:054202, Feb 2023.DOI
  11. [11]Jacopo Baglioni, Alessandro Martinelli, Peihao Sun, Francesco Dallari, Fabian Westermeier, Michael Sprung, Gerhard Grübel, and Giulio Monaco. Uniqueness of glasses prepared via x-ray induced yielding. Reports on Progress in Physics, 87(12):120503, nov 2024.
  12. [12]Alessandro Martinelli, Federico Caporaletti, Francesco Dallari, Michael Sprung, Fabian Westermeier, Giacomo Baldi, and Giulio Monaco. Reaching the yield point of a glass during x-ray irradiation. Phys. Rev. X, 13:041031, Nov 2023.DOI
  13. [13]Shuai Wei, Pierre Lucas, and Charles Angell. Phase change materials: The view from the liquid phase and the metallicity parameter. MRS Bulletin, 44:691–698, 09 2019.DOI
  14. [14]Piarristeguy, Andrea, Noé, Pierre, and Hippert, Françoise. Verres de chalcogénures pour le stockage de l’information. Reflets phys., (74):58–63, 2022.DOI
  15. [15]Shuai Wei, Pierre Lucas, and C. Austen Angell. Phase change materials: The view from the liquid phase and the metallicity parameter. MRS Bulletin, 44(9):691–698, 2019.DOI
  16. [16]Jens Moesgaard, Tomoki Fujita, and Shuai Wei. Unveiling the boson peaks in amorphous phase-change materials. Journal of Physics: Condensed Matter, 37(2):025101, oct 2024.arxiv.org/abs/2404.17226
  17. [17]Jean Yves Raty, Wei Zhang, Jennifer Luckas, Chao Chen, Riccardo Mazzarello, Christophe Bichara, and Matthias Wuttig. Aging mechanisms in amorphous phase-change materials. Nature Communications, 6(1):7467, Jun 2015.DOI
  18. [18]M. Chen, K. A. Rubin, and R. W. Barton. Compound materials for reversible, phase-change optical data storage. Applied Physics Letters, 49(9):502–504, 09 1986.DOI
  19. [19]Shuai Wei, Pierre Lucas, and Charles Angell. Phase change alloy viscosities down to T₍g₎ using Adam-Gibbs equation fittings to excess entropy data: A fragile-to-strong transition. Journal of Applied Physics, 118:034903, 07 2015.DOI
  20. [20]Hans Weber, Jiri Orava, Ivan Kaban, Julian Pries, and A. Lindsay Greer. Correlating ultrafast calorimetry, viscosity, and structural measurements in liquid GeTe and Ge₁₅Te₈₅. Phys. Rev. Mater., 2:093405, Sep 2018.
  21. [21]Shuai Wei, Moritz Stolpe, Oliver Gross, William Hembree, Simon Hechler, Jozef Bednarčik, Ralf Busch, and Pierre Lucas. Structural evolution on medium-range order during the fragile-strong transition in Ge₁₅Te₈₅. Acta Materialia, 129:259–267, 2017.DOI
  22. [22]Rathinavelu Sengottaiyan, Nishant Saxena, Krishna Dayal Shukla, and Anbarasu Manivannan. Temperature dependent structural evolution and crystallization properties of thin Ge₁₅Te₈₅ film revealed by in situ resistance, x-ray diffraction and scanning electron microscopic studies. Journal of Physics D: Applied Physics, 53(2):025108, oct 2019.
  23. [23]Pierre Lucas, Shuai Wei, and Charles Angell. Liquid liquid phase transitions in glass-forming systems and their implications for memory technology. International Journal of Applied Glass Science, 11, 02 2020.DOI
  24. [24]Julian Pries, Shuai Wei, Felix Hoff, Pierre Lucas, and Matthias Wuttig. Control of effective cooling rate upon magnetron sputter deposition of glassy Ge₁₅Te₈₅. Scripta Materialia, 178:223–226, 2020.
  25. [25]Julian Pries, Yuan Yu, Peter Kerres, Maria Häser, Simon Steinberg, Fabian Gladisch, Shuai Wei, Pierre Lucas, and Matthias Wuttig. Approaching the glass transition temperature of GeTe by crystallizing Ge₁₅Te₈₅. physica status solidi (RRL) – Rapid Research Letters, 15(3):2000478, 2021.
  26. [26]Tomoki Fujita, Yuhan Chen, Yoshio Kono, Seiya Takahashi, Hidetaka Kasai, Davide Campi, Marco Bernasconi, Koji Ohara, Hirokatsu Yumoto, Takahisa Koyama, Hiroshi Yamazaki, Yasunori Senba, Haruhiko Ohashi, Ichiro Inoue, Yujiro Hayashi, Makina Yabashi, Eiji Nishibori, Riccardo Mazzarello, and Shuai Wei. Pressure induced reversal of Peierls-like distortions elicits the polymorphic transition in GeTe and GeSe. Nature Communications, 14, 12 2023.DOI
  27. [27]Yudong Cheng, Qun Yang, Jiangjing Wang, Theodoros Dimitriadis, Matthias Schumacher, Huiru Zhang, Maximilian J. Müller, Narges Amini, Fan Yang, Alexander Schoekel, Julian Pries, Riccardo Mazzarello, Matthias Wuttig, Hai-Bin Yu, and Shuai Wei. Highly tunable β-relaxation enables the tailoring of crystallization in phase-change materials. Nature Communications, 13(1):7352, Nov 2022.DOI
  28. [28]Gerhard Grübel and Federico Zontone. Correlation spectroscopy with coherent x-rays. Journal of Alloys and Compounds, 362(1):3–11, 2004. Proceedings of the Sixth International School and Symposium on Synchrotron Radiation in Natural Science (ISSRNS).
  29. [29]J C Phillips. Stretched exponential relaxation in molecular and electronic glasses. Reports on Progress in Physics, 59(9):1133, sep 1996.DOI
  30. [30]Zach Evenson, Beatrice Ruta, Simon Hechler, M. Stolpe, Eloi Pineda, Isabella Gallino, and Ralf Busch. X-ray photon correlation spectroscopy reveals intermittent aging dynamics in a metallic glass. Physical Review Letters, 115:175701, 10 2015.DOI
  31. [31]P.G. De Gennes. Liquid dynamics and inelastic scattering of neutrons. Physica, 25(7):825–839, 1959.DOI
  32. [32]Maximilian Frey, Nico Neuber, Sascha Riegler, Antoine Cornet, Yuriy Chushkin, Federico Zontone, Lucas Ruschel, Bastian Adam, Mehran Nabahat, Fan Yang, Jie Shen, Fabian Westermeier, Michael Sprung, Daniele Cangialosi, Valerio Di Lisio, Isabella Gallino, Ralf Busch, Beatrice Ruta, and Eloi Pineda. Liquid-like versus stress-driven dynamics in a metallic glass former observed by temperature scanning x-ray photon correlation spectroscopy. Nature Communications, 16, 05 2025.
  33. [33]B. Ruta, S. Hechler, N. Neuber, D. Orsi, L. Cristofolini, O. Gross, B. Bochtler, M. Frey, A. Kuball, S. S. Riegler, M. Stolpe, Z. Evenson, C. Gutt, F. Westermeier, R. Busch, and I. Gallino. Wave-vector dependence of the dynamics in supercooled metallic liquids. Phys. Rev. Lett., 125:055701, Jul 2020.
  34. [34]Nico Neuber, Oliver Gross, Maximilian Frey, Benedikt Bochtler, Alexander Kuball, Simon Hechler, Fan Yang, Eloi Pineda, Fabian Westermeier, Michael Sprung, Isabella Gallino, Ralf Busch, and Beatrice Ruta. Disentangling structural and kinetic components of the α-relaxation in supercooled metallic liquids, Communications Physics, 5:316, 2022.
  35. [35]Vladimir Novikov, K Schweizer, and Alexei Sokolov. Coherent neutron scattering and collective dynamics on mesoscale. The Journal of chemical physics, 138:164508, 04 2013.
  36. [36]Beatrice Ruta, Giacomo Baldi, Y Chushkin, Benoit Rufflé, Luigi Cristofolini, A Fontana, Marco Zanatta, and Francesco Nazzani. Revealing the fast atomic motion of network glasses. Nature communications, 5:3939, 05 2014.
  37. [37]Si-Xu Peng, Yudong Cheng, Julian Pries, Shuai Wei, Hai-Bin Yu, and Matthias Wuttig. Uncovering β-relaxations in amorphous phase-change materials. Science Advances, 6(2):eaay6726, 2020.DOI

Paper details

vlnc:W_k26vfdnx
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 7

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents