Beating Homogeneous Nucleation and Tuning ... - ACS Publications

Subscriber access provided by READING UNIV

Communication

Beating homogeneous nucleation and tuning atomic ordering in glass-forming metals by nanocalorimetry Bingge Zhao, Bin Yang, Alexander S Abyzov, Jürn W. P. Schmelzer, Javier Rodriguez-Viejo, Qijie Zhai, Christoph Schick, and Yulai Gao Nano Lett., Just Accepted Manuscript • Publication Date (Web): 07 Nov 2017 Downloaded from http://pubs.acs.org on November 7, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Beating homogeneous nucleation and tuning atomic ordering in glass-forming metals by nanocalorimetry Bingge Zhao†, §, Bin Yang¶, Alexander S. Abyzov‡, Jürn. W. P. Schmelzer¶, Javier RodríguezViejo∫, Qijie Zhai†, Christoph Schick¶, *, Yulai Gao†, §, * †

State Key Laboratory of Advanced Special Steel & Shanghai Key Laboratory of Advanced

Ferrometallurgy & School of Materials Science and Engineering, Shanghai University, Shanghai 200072, P.R. China §

Laboratory for Microstructures, Shanghai University, Shangda Road 99, Shanghai 200444, P.R. China ¶

Institute of Physics, University of Rostock, Albert-Einstein-Street 23-24, Rostock 18051, Germany

‡

National Science Center Kharkov Institute of Physics and Technology, Academician Street 1, Kharkov 61108, Ukraine ∫

Physics Department, Universitat Autònoma de Barcelona, Bellaterra 08193, Spain

ACS Paragon Plus Environment

1

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 34

ABSTRACT

In this paper, the amorphous Ce68Al10Cu20Co2 (at. %) alloy was in situ prepared by nanocalorimetry. The high cooling and heating rates accessible with this technique facilitate the suppression of crystallization on cooling and the identification of homogeneous nucleation. Different from the generally accepted notion that metallic glasses form just by avoiding crystallization, the role of nucleation and growth in the crystallization behavior of amorphous alloys is specified, allowing an access to the ideal metallic glass free of nuclei. Local atomic configurations are fundamentally significant to unravel the glass forming ability (GFA) and phase transitions in metallic glasses. By this reason, isothermal annealing near Tg from 0.001 s to 25,000 s following quenching becomes the strategy to tune local atomic configurations and facilitate an amorphous alloy, a mixed glassy-nanocrystalline state, and a crystalline sample successively. Based on the evolution of crystallization enthalpy and overall latent heat on reheating, we quantify the underlying mechanism for the isothermal nucleation and crystallization of amorphous alloys. With Johnson-Mehl-Avrami method, it is demonstrated that the coexistence of homogeneous and heterogeneous nucleation contributes to the isothermal crystallization of glass. Heterogeneous rather than homogeneous nucleation dominates the isothermal crystallization of the undercooled liquid. For the mixed glassy-nanocrystalline structure, an extraordinary kinetic stability of the residual glass is validated, which is ascribed to the denser packed interface between amorphous phase and ordered nanocrystals. Tailoring the amorphous structure by nanocalorimetry permits new insights into unraveling GFA and the mechanism that correlates local atomic configurations and phase transitions in metallic glasses.

KEYWORDS: nucleation, crystallization, nanocalorimetry, glass forming ability


2

Page 3 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Since the first discovery of Au-Si metallic glasses (MGs) by splat quenching in 1960, considerable attention has been paid on their phase transitions owing to their importance both in condensed matter physics and engineering fields1-4. The transformation from liquid to a crystalline solid generally involves nucleation and growth. Nevertheless, glass forming ability (GFA) focuses on the critical cooling rates avoiding "a detectable fraction of crystals to nucleate and grow"5, yet the nucleation and growth is less distinguished. Recently, an increasing number of reports shed new light on growth and depict a coherent picture and outlook of GFA2,6,7. The role of nucleation in glass formation, by contrast, is ambiguous at present as direct visualization of nucleation remains challenging and inconclusive8-10. In some metallic glasses, ultra-fine crystals (less than 2 nm) with an extremely high density of 1020-1022 /m3 are observed after quenching or isothermal annealing of the as-quenched sample, which is indicative of homogeneous nucleation11-13. Due to the limitation in the experimentally accessible cooling rate, these quenched-in nuclei are hard to suppress. As a result, it is difficult to achieve a truly amorphous structure. Such local ordered nuclei have an important bearing on subsequent phase transitions14,15. Nevertheless, quantitative analysis on the initiation and evolution of quenched-in nuclei and their effect on the subsequent transformations (e.g. glass transition and crystallization) currently remains a challenging issue. In common with the undercooled liquid, glass possesses a disordered structure as well, yet it behaves mechanically like a solid16. It is stated that atom diffusion in glass dramatically slows down and the decrease of temperature is confronted with a super-Arrhenius increase in relaxation time17. Below Tg, atoms are structurally "stuck" in local clusters. Both the nucleation and crystallization of glass are significantly slower as compared with that in undercooled liquid. For example, 200 years is required for the crystallization of Ce70Al10Cu20 at room temperature18.


3

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 34

Accordingly, traditional studies just address the relaxation and local atomic rearrangement rather than long-range ordering in glass19,20. Neither the nucleation mechanism below Tg nor its influence on following glass transition, crystallization and melting has been illustrated in detail. In other words, the bridge between undercooled liquid and glass is cut off artificially, bringing the intriguing scenario: whether classical nucleation theory (CNT) is still applicable in glassy state. Nanocalorimetry provides a potential solution to address these issues due to its ultrafast scanning rates and ultrahigh sensitivity21-24. The cooling rate up to 106 K/s allows the in situ preparation of amorphous phase and tuning the local atomic configurations25. Compared with conventional DSC, the ultrahigh heating rate of nanocalorimetry can elevate Tg significantly, which enhances atom mobility and accelerates the nucleation and crystallization in the glassy state on a laboratory timescale. Actually, some attractive results have been acquired using this novel technique. For instance, the ultrahigh heating rate alters both the transition temperature and diffusion kinetics in Zr-B thin films26,27. A transition from surface heterogeneous nucleation to volume heterogeneous nucleation was detected in single Sn droplets with the increase of cooling rate28. The ultrahigh sensitivity of nanocalorimetry enabled heat capacity measurements on ultrathin CoO films of thickness down to 1.5 nm to unveil the thickness dependence of the magnetic transition29. In this paper, the nucleation and crystallization of Ce68Al10Cu20Co2 (at. %) bulk metallic glass (BMG) are systematically demonstrated by DFSC both under isothermal and non-isothermal conditions. Ce68Al10Cu20Co2 is not a popular metallic glass in contrast to Zr- and Pd-based ones but it is used here mainly for two reasons: (i) This alloy has low melting temperature. It can fully melt on heating and form amorphous structure on following quenching without damaging the sensor used in nanocalorimetric measurements. (ii) Many of the phase


4

Page 5 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

transitions in this alloy remain inconclusive due to the atom arrangements at ambient temperature. Nanocalorimetry with ultrahigh cooling rate and very short stabilization times between cooling and heating favors the in situ preparation of metallic glass, overcoming the challenge caused by room-temperature relaxation and atom ordering. As the cooling rate changes from 100 K/s to 50,000 K/s, not only the critical cooling rate suppressing crystallization but the one beating homogeneous nucleation is unambiguously distinguished. The isothermal annealing treatments at temperatures near Tg after quenching create various structure, such as relaxed clusters, ordered nuclei and nanocrystals. Based on the crystallization enthalpy and overall latent heat on subsequent reheating, isothermal ordering and its effect on glass transition, crystallization and melting are quantitatively evaluated based on Johnson-Mehl-Avrami (JMA) method. All the measurements in this study were performed on one Ce68Al10Cu20Co2 sample, enabling enthalpy calculation to quantify phase transitions. The experimental details can be found in the Supporting Information. Beating homogeneous nucleation on quenching The Ce68Al10Cu20Co2 particle was first heated to 735 K to melt at a heating rate of 1,000 K/s and then went through a series of cooling and reheating treatments. We used different cooling rates from 100 K/s to 50,000 K/s to quench the melt to 200 K while a heating rate of 30,000 K/s was then adopted to reheat the sample to 735 K. Figure 1 demonstrates phase transitions on the reheating following quenching. At cooling rates lower than 1,000 K/s, no glass transition and crystallization exothermic peaks were detected (Figure 1a), which means both nucleation and growth have occurred, leading to a fully crystallized sample. When the cooling rate is above 1,000 K/s (1,000 K/s to 10,000 K/s), the reheating curves in Figure 1a show more pronounced glass transition and crystallization exothermic peaks. The cooling rates from 1,000 K/s to 10,000


5

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 34

K/s can preserve the amorphous phase and increase its volume fraction. As the cooling rate is higher than 10,000 K/s, the reheating curves are nearly independent on previous cooling rate, suggesting a stable transformation on cooling. The dependence of Tx,o and Tm,p on previous cooling rate is reflected in Figure 1b. The increase of cooling rate from 100 K/s to 50,000 K/s provokes a reduction of Tm,p from 690 K to 678 K, which is attributed to the asymmetry in crystallization behavior on quenching the equilibrium melt and heating the glass. According to Figure 1a, 1,000 K/s is the critical point at which the crystallization behavior upon quenching changes. Below that, the crystallization is triggered on cooling, leading to a fully crystallized sample. On the subsequent reheating, both the glass transition step and crystallization exotherm are absent, and thus only a melting endotherm that corresponds to the crystals formed on cooling is detected. However, the cooling rate above 1,000 K/s leads to the amorphous structure. During reheating, those amorphous states go through devitrification and melting successively. In other words, the crystals are created on cooling when the cooling rate is below 1,000 K/s while they are produced on heating when the cooling rate exceeds this value. Given the discrepancy in crystallization temperature between cooling and reheating, a variation in crystalline phases is assumed to occur, which is responsible for the asymmetry in melting endotherms. Although structure characterization is absent, the formation of melting shoulder and its evolution on reheating curve can manifest this hypothesis, which is demonstrated below. When the previous cooling rate is below 1,000 K/s, a small shoulder in the leading edge of the melting endotherm is detected on the reheating curves (see Figure 1a). Nevertheless, as previous cooling rate exceeds 1,000 K/s, the shoulder disappears, leaving an individual melting endotherm. Because of the ultrahigh reheating rate as reflected in Figure 1a, the shoulder overlaps with the main melting endotherm apparently, making it difficult to


6

Page 7 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

distinguish the phases formed on cooling and reheating. A lower heating rate of 1,000 K/s can increase the temperature resolution and separate the overlapping thermal transitions, as confirmed by Figure S4 of the Supporting Information. According to this result and Figure 1a, it is highly convincing that the asymmetry in phase formation leads to the difference in the melting endotherm. The position of melting shoulders, on the other hand, is dependent on the previous cooling rate, which is attributed to the size-effect in melting behavior. It is argued that the crystallization on heating is controlled by crystal growth while that is controlled by nucleation on cooling30,31, as schematically expressed in Figure 1c. According to the generally accepted theory32, both the nucleation rate (blue curves) and diffusion-limited growth rate (red curve) reach their maximum values between Tg and Tm. Since amorphous phases are produced as cooling rate is beyond 1,000 K/s, nanocrystals are assumed to form at cooling rate below that. A higher cooling rate in general gives rise to refined nanocrystals arisen from the enhanced nucleation rate, and a reduction of melting temperature attributing to the size-effect is thus achieved on following reheating. Nevertheless, as cooling rate is larger than 1,000 K/s, amorphous phases are acquired. Subsequent reheating of the quenched glass causes crystallization, and crystal size is not significantly changed owing to the identical reheating rate in different cycles. A stable melting temperature is thus detected even though the previous cooling rate is further enlarged. Tx,o increases from 532 K to 572 K when the cooling rates increase from 1,000 K/s to 50,000 K/s (Figure 1b). Both the glass transition and crystallization on reheating are more pronounced as previous cooling rate enlarges from 1,000 K/s to 10,000 K/s, which is indicative of the mixed glassy-crystalline structure under this condition. In general, the composition of the residual amorphous phase can be changed with the precipitation of crystals. In the present study,


7

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 34

nevertheless, both Tg (Figure S4a of the Supporting Information) and Tm (Figure 1b) are very stable despite the formation of crystals at different previous cooling rates. In those cases, the crystallization can be limited in an infinitesimal zone attributing to the ultrahigh scanning rate of the utilized nanocalorimetry. That is to say, the element diffusion can be greatly prohibited during this process. And therefore, the main residual amorphous matrix can be kept unchanged. Instead, heterogeneous nucleation should be weighed in the evolution of Tx,o. The crystals produced on cooling provide sites for nucleation, facilitating the crystallization on following reheating. The increase of cooling rate reduces the volume fraction of crystalline phase, providing less catalytic sites for heterogeneous nucleation and making it more difficult to provoke the crystallization on reheating. In this case, a higher crystallization temperature is induced. At cooling rates exceeding 10,000 K/s, Tx,o nearly keeps constant, suggesting the structure homogeneity regardless of different cooling rates. Arguably, the experimental critical cooling rate suppressing crystallization on cooling is about 10,000 K/s in this BMG. Different from the results in Figure 1, it has been reported that there was no detectable crystallization in the as-cast Ce68Al10Cu20Co2 sample33 although the cooling rate of copper-mold casting (102-103 K/s19) is far less than the one we used in DFSC measurement (e.g. 1,000-50,000 K/s). A few nanocrystals less than 5 nm have been observed in the as-cast sample using transmission electron microscopy33, suggesting the occurrence of nucleation (the early stage of crystallization) in the as-cast sample. Up to now, little information is available about such behavior. Due to the infinitesimal local structure (τc of crystallization enthalpy is therefore added to eq 1, yielding

) |+τ |

∆, = ∆ 1 − exp −2 τ + !" (ln% − ln&' ( " +τ + 1

(2)

where A2 is the secondary crystallization parameter reflecting the fraction of the secondary crystallization. At the cooling rate of 50,000 K/s, heterogeneous nucleation is inevitable, and surface crystallization takes place, as manifested by the non-zero ∆Hc and ∆Ho,heating. So another constant term concerning the enthalpy originating from heterogeneous crystallization, ∆Hc,het , when annealing is performed at 420 or 440 K, is added to eq 2, and the overall latent heat is written as

) |+τ |

∆, = ∆' 1 − exp −2 τ + !" (ln% − lnτ' ( " +τ + 1 + Δ , (3)

As indicated by the upper part of Figure 4, eq 3 is effectively describing the experimental results. The Avrami exponents are listed in Table S1 of the Supporting Information. Constant ∆Hc,het corresponds to the crystallization enthalpy that is caused by the ultrafast saturation of the heterogeneously-formed nuclei on the sample surface within 0.001 s which is beyond the time resolution of DFSC.


14

Page 15 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Homogeneous nucleation is confined on quenching at 50,000 K/s while it is speculated to occur inside the sample during the annealing, as supported by the increased ∆Hc in Figure 4. The enthalpy associating with nucleation can be described as

∆ = ∆,. 1 − exp −2 τ/

(4)

Here, ∆Hn,hom∞ is the crystallization enthalpy related to the homogeneous nucleation at infinite time, τn is the half time of nucleation which is corresponding to the time needed to acquire 50% nuclei from the amorphous phase, n is the Avrami parameter reflecting nucleation. Similarly, the enthalpy evolution reflecting heterogeneous nucleation, ∆Hn,het, is added to eq 4, yielding the modified nucleation-caused crystallization enthalpy

∆ = ∆,. 1 − exp −2 τ + Δ, /

(5)

Constant ∆Hn,het implies ultrafast saturation of heterogeneously-formed nuclei on the sample surface within 0.001 s. Equation 5 can describe the accumulation of nuclei well during annealing but fail to describe the isothermal crystallization after the saturation of nuclei. Under this condition, the crystallization on reheating is governed by the isothermal crystallization rather than nucleation. Additionally, ∆Ho,heating can reflect the isothermal crystallization, and eq 3 has been successful in fitting the crystallization-caused enthalpy on reheating. This case implies that the isothermal nucleation and crystallization contribute to the crystallization enthalpy on reheating. Therefore, the total crystallization enthalpy on reheating can be considered as the sum of isothermal nucleation- and crystallization-caused enthalpy (∆Hc,iso): ∆ ,01 = ∆ + ∆ ,2

(6)

As mentioned above, ∆Hc,iso is equal to ∆Ho,heating. Consequently, eq 6 was used to fit the crystallization enthalpy on reheating, which describes the experimental results well, as indicated


15

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 34

in the bottom part of Figure 4 (Table S2 of the Supporting Information lists the Avrami exponents). As noted, nucleation and growth occur successively, and Avrami exponents are varied by this reason. So, it seems arguable to describe these successive phase transitions just by a single value of n, and further information is being pursued. Based on the annealing at 420 K and 440 K, both homogeneous and heterogeneous nucleation take place and contribute to crystallization. At 460 K, however, the increase of crystallization enthalpy due to homogenous nucleation is absent, implying the exclusive contribution by heterogeneous nucleation. As revealed in Figure 4c, crystallization enthalpy on heating can be well fitted only by eq 1, which further points to the absence of homogeneous nucleation in the annealing of 460 K. According to CNT, homogeneous nucleation rate is dramatically confined as temperature approaches Tg. By contrast, heterogeneous nucleation occurs at a much higher rate that depends on the contact angle50. Based on the CNT and half-times of nucleation and crystallization in the present experiment, one possible activation diagram depicting the nucleation and crystallization of Ce68Al10Cu20Co2 BMG is schematically plotted in Figure 5. In a glassy state, unavoidable heterogeneous nucleation is highly feasible on the sample surface while homogeneous nucleation takes place in the internal zone, which is similar with the formation of Pd seeds in an aqueous solution61. After the saturation of the homogeneously- and heterogeneously-formed nuclei, they grow in size and result in isothermal crystallization with further annealing. Whereas at 460 K, fast crystallization arises as soon as heterogeneous nucleation occurs, implying that heterogeneous nucleation makes a major contribution to the isothermal crystallization of undercooled liquid. In spite of the isothermal crystallization, there remains amorphous phase in the sample, as proven by the coexistence of a glass transition step and low-temperature melting endotherm in


16

Page 17 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 6a. That is to say, a mixed glassy-nanocrystalline structure is formed with a substantial increase of Tg on reheating. The results summarized in Figure 6b reveal an enhanced kinetic stability of residual glass with nanocrystals in comparison with a fully amorphous glass, resembling the concept of rigid-amorphous fractions62. Generally, the composition of the residual amorphous phase is affected by the formation of nanocrystals, which affects the phase transitions on reheating. However, some recent work implies that the structure evolution may be the decisive factor in this case. Apart from the higher Tg owing to relaxation63,64, denser clusters are also reveled in the mixed glassy-crystallization state. For example, Ichitsubo obtained a partially crystallized Pd40Ni40P20 sample by sub-Tg annealing and found that the strongly bonded regions are more stable65. In another sub-Tg annealed Cu46Zr44Al8Hf2 BMG, polymorphic phase precipitation occurs without changing the composition of the amorphous phase but denser packed local configurations are also detected66. An increased Tg is suggested in these cases. Although a direct characterization on the structure after annealing at 420 K, 440 K and 460 K is inaccessible at present, a constant melting temperature regardless of annealing time is observed in Figure 6b. This scenario suggests that the composition of the residual amorphous phase is rarely affected during isothermal crystallization despite the mixed glassy-nanocrystalline state in this sample. The absence of composition variation in crystallized sample that experiences longterm annealing has been demonstrated in other cases67, which is consistent with our results. A possible mechanism is hypothesized to elucidate the extraordinary kinetic stability. At the beginning of annealing, less than 1 s at 420 K for example, the sample is amorphous albeit the enthalpy relaxation and local rearrangement of atoms (Figure 6c). As the annealing continues, nuclei form and develop into nanocrystals. As identified by the coexistence of glass transition and decreased crystallization enthalpy at reheating (see Figure 3a-c), some amorphous


17

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 34

phases remain in the sample and act as the matrix, creating the mixed glassy-crystalline structure (see the schematic in Figure 6(d)). Even for the remnant amorphous phase, a structure heterogeneity is expected. An intermediate interfaced between the parent phase and new phase during solid-solid phase transitions has been demonstrated9. This intermediate phase has a distinguished microstructure from both the parent and product phases. By an early molecular dynamics (MD) simulation, an enhanced ordering has been illustrated in the solid-liquid interface68. For the amorphous/crystalline interface, atoms are more densely packed than common amorphous phase69,70, that is to say, a particular amorphous/crystalline interface can form when nanocrystals precipitate from the amorphous phase71. On the reheating after annealing, these interfaces (dark blue) are more stable than the relaxed matrix (light blue), causing a higher Tg on reheating. As noted in Figure 6a, the shift of Tg begins from the isothermal crystallization stage in which nano-sized nuclei have saturated and crystal growth dominates the crystallization.

Attributing to

the

large number of nuclei,

the interface

between

ordered/disordered phase makes up the majority of residual glass and dominates the phase transitions on reheating. With the assumption of residual amorphous phase during further annealing (Figure 6e), the role of interface is enhanced, giving rise to higher Tg. Moreover, it is argued that the amorphous/crystalline interface can work as sinks, enabling the absorption of free volumes72, which increases Tg as well. As for bulk metallic glass composites, Tg remains constant regardless of the crystals73. Most crystals in BMG composite are several tens of microns, producing a much smaller ratio of interface to volume compared with the nanocrystals in the present study. Tg of the BMG composite is still determined by amorphous matrix and behaves nearly the same as a result.


18

Page 19 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Because of the ultrahigh cooling rate of nanocalorimetry, amorphous Ce68Al10Cu20Co2 was in situ prepared, and a state-of-the-art strategy to tune the local atomic configurations was put forward accordingly. The reheating process after annealing allows the understanding of the underlying mechanism about the isothermal nucleation and crystallization as well as their decisive effect on glass transition, crystallization and melting. Based on the crystallization enthalpy and overall latent heat, the kinetics of ordering both in glass and undercooled liquid is quantitatively elucidated. The outcome of this research raises a possibility to manipulate the local atomic configurations and phase transitions precisely through nanocalorimetry. With the increase of the quenching rate from 100 K/s to 50,000 K/s, crystalline, mixture of amorphous-crystalline and completely amorphous state can be obtained successively in Ce68Al10Cu20Co2. Due to the strong heterogeneous nucleation on the surface of the DFSC sample, the critical cooling rate avoiding crystallization is experimentally determined as 10,000 K/s which is much higher than that estimated from the as-cast sample. Additionally, homogeneous nucleation can be suppressed at cooling rates higher than 10,000 K/s, providing an ideal state to tune the local atomic configurations by isothermal annealing. Current in-depth analyses distinguish the nucleation from crystallization, which allows new insights into the GFA and the nature of crystallization. Beginning with the amorphous phase, local atomic configurations are tuned by annealing in glass and in undercooled liquid from 0.001 s to 25,000 s, in which enthalpy relaxation, nucleation and crystallization are detected continuously. Size effects are observed attributing to the formation of nano-sized clusters during annealing, producing the low-temperature melting endotherm. The kinetics of isothermal nucleation and crystallization is quantitatively revealed by analyzing the evolution of crystallization enthalpy and overall latent heat. Annealing in glassy


19

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 34

state, homogeneous and inevitable heterogeneous nucleation contribute to the isothermal crystallization. While the isothermal crystallization of undercooled liquid is dominated by the heterogeneous nucleation. A mixed glassy-nanocrystalline structure is achieved during annealing. The denser packed interface between amorphous phase and nanocrystals contributes to an enhanced kinetic stability of residual glass with a higher Tg. Last but not least, nanocalorimetry has provided impressive and insightful results on the phase transitions occurring under extremely non-equilibrium conditions. Structure characterization, on the other hand, is less involved due to the micro- or nano-sized nanocalorimetric sample. The combination of nanocalorimetry and structure characterization is deemed as the next hotspot in this field.


20

Page 21 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 1. Phase transitions on reheating as a function of previous cooling rate. (a) Reheating curves after cooling at rates from 100 K/s to 50,000 K/s. The upper one indicates the typical reheating curves with previous cooling at 500 K/s, 5,000 K/s and 50,000 K/s respectively. Other curves are given in Figure S4 of Supporting Information. The bottom one displays the spectra of all reheating curves. (b) Relationship between the melting temperature (Tm,p, the peak temperature), the crystallization temperature (Tx,o, the onset temperature) and the corresponding cooling rate. Due to the overlap between crystallization exotherm and melting endotherm, the determination of characteristic temperature is varied consequently. (c) Schematic of nucleation rate and growth rate at different temperatures. The blue solid line (indicated by the blue solid arrow) represents the evolution of homogeneous nucleation rate while the dashed-dot curve (blue dashed-dot arrow) schematically displays the heterogeneous nucleation rate. The red solid curve (red arrow) denotes the growth rate between Tg and Tm. T1 and T2 correspond to the various nucleation temperatures.


21

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 34

Figure 2. Evolution of the overall latent heat and crystallization enthalpy on reheating at 30,000 K/s versus previous cooling rate. The evolution of melting enthalpy as well as the determination of enthalpy is displayed in the Supporting Information.


22

Page 23 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 3. Phase transitions on the reheating after the isothermal annealing. (a), (b) and (c) are the spectra of reheating curves following the isothermal annealing at 420 K, 440 K and 460 K from 0.001 s to 25,000 s, respectively. Detailed reheating curves can be found in Figure S7 of the Supporting Information. (d) and (e) illustrate the evolution of enthalpy and temperature of the low-temperature melting endotherm as a function of annealing time.


23

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 34

Figure 4. Time evolution of enthalpy during annealing at (a) 420 K, (b) 440 K and (c) 460 K. Upper figures represent the overall latent heat while the bottom ones indicate the evolution of crystallization enthalpy. Here, ∆Ho,heating is calculated as the sum of crystallization enthalpy and melting enthalpy (both the low- and high-temperature melting endotherms are included). Region

α represents the enthalpy caused by the inevitable heterogeneous nucleation on cooling. β corresponds to the homogeneous nucleation during annealing which produces more nuclei. γ indicates the isothermal crystallization which reduces the number of nuclei.


24

Page 25 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 5. Schematic activation diagram depicting the nucleation and crystallization of Ce68Al10Cu20Co2. In glassy state (420 K and 440 K), both homogeneous and heterogeneous nucleation contribute to the crystallization. Heterogeneous nucleation, on the other hand, dominates the isothermal crystallization of undercooled liquid (460 K).


25

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 34

Figure 6. Enhanced kinetic stability of the residual amorphous phase under different annealing conditions. (a) Time-dependent reheating curves displaying the glass transition and lowtemperature melting endotherm. Tg is defined as the onset temperature, as indicated by the black arrows. (b) Evolution of Tg and high-temperature melting endotherm with annealing time in the mixed glassy-crystalline state. Tg increases with annealing time while the melting endotherm is nearly constant, implying a composition homogeneity of the residual amorphous phase despite the nanocrystals. (c)-(e) Schematically indicating the nucleation and crystallization during annealing. A higher densely packed interface (dark blue) is argued to exist between the nanosized ordered structure (grey spheres) and amorphous matrix (light blue). AUTHOR INFORMATION Corresponding Authors *

Christoph Schick: [email protected]


26

Page 27 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

*

Yulai Gao: [email protected]

ACKNOWLEDGMENT This work is supported by the National Natural Science Foundation of China (Grant Nos. 51671123, 51171105, 50971086, 50571057 and 50401023), Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning (Grant No. TP2014042), P.R. China, and the 085 project in Shanghai University, P.R. China. ASSOCIATED CONTENT Supporting Information Details of sample preparation, nanocalorimetry profiles, DFSC curves, melting enthalpy, the determination of enthalpy, atom rearrangements at room temperature and Avrami fitting parameters (PDF). REFERENCES 1.

Johnson, W. L.; Kaltenboeck, G.; Demetriou, M. D.; Schramm, J. P.; Liu, X.; Samwer,

K.; Kim, C. P.; Hofmann, D. C. Science 2011, 332 (6031), 828-833. 2.

Orava, J.; Greer, A. L.; Gholipour, B.; Hewak, D. W.; Smith, C. E. Nat. Mater. 2012, 11

(4), 279-283. 3.

Kumar, G.; Desai, A.; Schroers, J. Adv. Mater. 2011, 23 (4), 461-476.

4.

Nakayama, K. S.; Yokoyama, Y.; Wada, T.; Chen, N.; Inoue, A. Nano Lett. 2012, 12 (5),

2404-2407. 5.

Johnson, W. L.; Na, J. H.; Demetriou, M. D. Nat. Commun. 2016, 7, 10313.


27

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 34

6.

Greer, A. L. Nat. Mater. 2015, 14 (6), 542-546.

7.

Zhong, L.; Wang, J.; Sheng, H.; Zhang, Z.; Mao, S. X. Nature 2014, 512 (7513), 177-

180. 8.

Gasser, U.; Weeks, E. R.; Schofield, A.; Pusey, P. N.; Weitz, D. A. Science 2001, 292

(5515), 258-262. 9.

Peng, Y.; Wang, F.; Wang, Z.; Alsayed, A. M.; Zhang, Z.; Yodh, A. G.; Han, Y. Nat.

Mater. 2015, 14 (1), 101-108. 10. Xia, Y.; Xiong, Y.; Lim, B.; Skrabalak, S. E. Angew. Chem. Int. Ed. 2009, 48 (1), 60-103. 11. Shen, Y. T.; Kim, T. H.; Gangopadhyay, A. K.; Kelton, K. F. Phys. Rev. Lett. 2009, 102 (5), 57801. 12. Kelton, K. F. J. Non-Cryst. Solids 2004, 334-335, 253-258. 13. Kelton, K. F.; Croat, T. K.; Gangopadhyay, A. K.; Xing, L. Q.; Greer, A. L.; Weyland, M.; Li, X.; Rajan, K. J. Non-Cryst. Solids 2003, 317 (1), 71-77. 14. Mizuguchi, T.; Higeta, Y.; Odagaki, T. Phys. Rev. E 2017, 95 (4), 42804. 15. Moore, E. B.; Molinero, V. Nature 2011, 479 (7374), 506-508. 16. Debenedetti, P. G.; Stillinger, F. H. Nature 2001, 410 (6825), 259-267. 17. Zhao, J.; Simon, S. L.; McKenna, G. B. Nat. Commun. 2013, 4, 1783. 18. Zhang, B.; Zhao, D. Q.; Pan, M. X.; Wang, W. H.; Greer, A. L. Phys. Rev. Lett. 2005, 94 (20), 205502.


28

Page 29 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

19. Liu, Y. H.; Fujita, T.; Aji, D. P. B.; Matsuura, M.; Chen, M. W. Nat. Commun. 2014, 5, 3238. 20. Yu, H.; Wang, W.; Samwer, K. Mater. Today 2013, 16 (5), 183-191. 21. Zhuravlev, E.; Schick, C. Thermochim. Acta 2010, 505 (1-2), 1-13. 22. Zhao, B.; Li, L.; Zhai, Q.; Gao, Y. Appl. Phys. Lett. 2013, 103 (13), 131913. 23. Zhao, B.; Li, L.; Lu, F.; Zhai, Q.; Yang, B.; Schick, C.; Gao, Y. Thermochim. Acta 2015, 603, 2-23. 24. Lee, D.; Sim, G.; Zhao, K.; Vlassak, J. J. Nano Lett. 2015, 15 (12), 8266-8270. 25. Minakov, A. A.; Schick, C. Rev. Sci. Instrum. 2007, 78 (7), 73902. 26. Lee, D.; Sim, G.-D.; Xiao, K.; Vlassak, J. J. J. Phys. Chem. C 2014, 118 (36), 2119221198. 27. Lee, D.; Vlassak, J. J.; Zhao, K. Nano Lett. 2015, 15 (10), 6553-6558. 28. Yang, B.; Gao, Y.; Zou, C.; Zhai, Q.; Abyzov, A. S.; Zhuravlev, E.; Schmelzer, J. W. P.; Schick, C. Appl. Phys. A 2011, 104 (1), 189-196. 29. Molina-Ruiz, M.; Lopeandía, A. F.; Pi, F.; Givord, D.; Bourgeois, O.; Rodríguez-Viejo, J. Phys. Rev. B 2011, 83 (14), 140407. 30. S. Pogatscher, P. J. U. Appl. Phys. Lett. 2014, 104, 251908. 31. Schroers, J.; Masuhr, A.; Johnson, W. L.; Busch, R. Phys. Rev. B 1999, 60 (17), 1185511858.


29

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 34

32. Hikima, T.; Adachi, Y.; Hanaya, M.; Oguni, M. Phys. Rev. B 1995, 52 (6), 3900-3908. 33. Zhang, B.; Pan, M. X.; Zhao, D. Q.; Wang, W. H. Appl. Phys. Lett. 2004, 85 (1), 61-63. 34. Matsunaga, T.; Akola, J.; Kohara, S.; Honma, T.; Kobayashi, K.; Ikenaga, E.; Jones, R. O.; Yamada, N.; Takata, M.; Kojima, R. Nat. Mater. 2011, 10 (2), 129-134. 35. Lan, S.; Ren, Y.; Wei, X. Y.; Wang, B.; Gilbert, E. P.; Shibayama, T.; Watanabe, S.; Ohnuma, M.; Wang, X. Nat. Commun. 2017, 8, 14679. 36. Salmerón Sánchez, M.; Mathot, V. B. F.; Vanden Poel, G.; Gómez Ribelles, J. L. Macromolecules 2007, 40 (22), 7989-7997. 37. Vyazovkin, S.; Dranca, I. J. Phys. Chem. B 2007, 111 (25), 7283-7287. 38. Paladi, F.; Oguni, M. Phys. Rev. B 2002, 65 (14), 144202. 39. Lin, X. H.; Johnson, W. L. J. Appl. Phys. 1995, 78 (11), 6514-6519. 40. Cacciuto, A.; Auer, S.; Frenkel, D. Nature 2004, 428 (6981), 404-406. 41. Zeng, J.; Zhu, C.; Tao, J.; Jin, M.; Zhang, H.; Li, Z.; Zhu, Y.; Xia, Y. Angew. Chem. Int. Ed. 2012, 51 (10), 2354-2358. 42. Kohda, M.; Haruyama, O.; Ohkubo, T.; Egami, T. Phys. Rev. B 2010, 81 (9), 92203. 43. Qin, Y. S.; Shi, Y.; Han, X. L.; Song, K. K.; Cao, C. D.; Li, X. L.; Wang, S. H.; He, J.; Wang, L.; Kaban, I.; Eckert, J. Mater. Today Commun. 2016, 8, 64-71. 44. Zhu, F.; Nguyen, H. K.; Song, S. X.; Aji, D. P. B.; Hirata, A.; Wang, H.; Nakajima, K.; Chen, M. W. Nat. Commun. 2016, 7, 11516.


30

Page 31 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

45. Angell, C. A.; Ngai, K. L.; McKenna, G. B.; McMillan, P. F.; Martin, S. W. J. Appl. Phys. 2000, 88 (6), 3113-3157. 46. Liu, X. F.; Wang, R. J.; Zhao, D. Q.; Pan, M. X.; Wang, W. H. Appl. Phys. Lett. 2007, 91 (4), 41901. 47. Van den Brande, N.; Van Assche, G.; Van Mele, B. Cryst. Growth Des. 2015, 15 (11), 5614-5623. 48. Xi, X.; Li, L. long; Zhang, B.; Wang, W.; Wu, Y. Phys. Rev. Lett. 2007, 99 (9), 95501. 49. Bingge Zhao; Quanliang Zhang; Ling Zhang; Chenyu Sun; Qijie Zhai; Yulai Gao. Bridging the amorphous structure and crystallization products in Ce68Al10Cu20Co2 bulk metallic glass by fast scanning calorimetry. 2017. 50. Lee, D.; Zhao, B.; Perim, E.; Zhang, H.; Gong, P.; Gao, Y.; Liu, Y.; Toher, C.; Curtarolo, S.; Schroers, J.; Vlassak, J. J. Acta Mater. 2016, 121, 68-77. 51. Kelton, K. F. Mater. Sci. Eng. A 1997, 226, 142-150. 52. Turnbull, D. J. Chem. Phys. 1950, 18 (2), 198-203. 53. Turnbull, D. J. Chem. Phys. 1952, 20 (3), 411-424. 54. Gebauer, D.; Cölfen, H. Nano Today 2011, 6 (6), 564-584. 55. Kim, B. J.; Tersoff, J.; Kodambaka, S.; Reuter, M. C.; Stach, E. A.; Ross, F. M. Science 2008, 322 (5904), 1070-1073.


31

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 34

56. Zhuravlev, E.; Schmelzer, J. W. P.; Wunderlich, B.; Schick, C. Polymer 2011, 52 (9), 1983-1997. 57. Evgeny Zhuravlev; Andreas Wurm; Petra Pötschke; Jürn W.P. Schmelzer; Christoph Schick. Eur. Polym. J. 2014, 52, 1-11. 58. Supaphol, P.; Spruiell, J. E. Polymer 2001, 42 (2), 699-712. 59. Rybnikář, F. J. Polym. Sci. 1960, 44 (144), 517-522. 60. Ikehara, T.; Nishi, T. Polymer 2000, 41 (21), 7855-7864. 61. Lim, B.; Jiang, M.; Yu, T.; Camargo, P. H. C.; Xia, Y. Nano Res. 2010, 3 (2), 69-80. 62. Wunderlich, B. Prog. Polym. Sci. 2003, 28 (3), 383-450. 63. Wang, J. Q.; Shen, Y.; Perepezko, J. H.; Ediger, M. D. Acta Mater. 2016, 104, 25-32. 64. Kobayashi, M.; Tanaka, H. Nat. Commun. 2016, 7, 13438. 65. Ichitsubo, T.; Matsubara, E.; Yamamoto, T.; Chen, H. S.; Nishiyama, N.; Saida, J.; Anazawa, K. Phys. Rev. Lett. 2005, 95 (24), 245501. 66. Sarac, B.; Bernasconi, A.; Wright, J.; Stoica, M.; Spieckermann, F.; Mühlbacher, M.; Keckes, J.; Bian, X.; Wang, G.; Eckert, J. Mater. Sci. Eng. A 2017, 707, 245-252. 67. Kündig, A. A.; Ohnuma, M.; Ohkubo, T.; Hono, K. Acta Mater. 2005, 53 (7), 2091-2099. 68. Guerdane, M.; Teichler, H.; Nestler, B. Phys. Rev. Lett. 2013, 110 (8), 86105. 69. Shibata, N.; Painter, G. S.; Becher, P. F.; Pennycook, S. J. Appl. Phys. Lett. 2006, 89 (5), 51908.


32

Page 33 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

70. Gao, X. Z.; Müser, M. H.; Kong, L. T.; Li, J. F. Model. Simul. Mater. Sci. Eng. 2014, 22 (6), 65007. 71. Kokotin, V.; Hermann, H.; Eckert, J. Intermetallics 2012, 30, 40-47. 72. Wang, Y.; Li, J.; Hamza, A. V.; Barbee, T. W. Proc. Natl. Acad. Sci. 2007, 104 (27), 11155-11160. 73. Chen, Z. P.; Gao, J. E.; Wu, Y.; Wang, H.; Liu, X. J.; Lu, Z. P. Sci. Rep. 2013, 3, 3353.


33

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 34

The graphic for the Table of Contents


34

Beating Homogeneous Nucleation and Tuning ... - ACS Publications

Recommend Documents