Academia.eduAcademia.edu

Low Temperature Hydrothermal Synthesis of Nanophase BaTiO3 and BaFe12O19 Powders

Profile image of Mehmet SarikayaMehmet Sarikaya

1996, MRS Proceedings

visibility

…

description

551 pages

link

1 file

AI-generated Abstract

This research focuses on the low temperature hydrothermal synthesis of nanophase barium titanate (BaTiO3) and barium hexaferrite (BaFe12O19) powders. It highlights the advancement in thermal conductivity achieved through the utilization of nanoparticles in fluids, offering potential applications in enhanced heat transfer systems. The study provides detailed methodologies for creating stable nanofluids and the characterization of resultant nanocrystalline powders, showcasing their implications for future thermal management technologies.

Figures (397)

a suspension containing large particles is more than doubled by decreasing the sphericity of the particles from a value of 1.0 to 0.3 (the sphericity is defined as the ratio of the surface area of a particle with a perfectly spherical shape to that of a non-spherical particle with the same volume). Since the surface area to volume ratio is 1000 times larger for particles with a 10 nm diameter than for particles with a 10mm diameter, a much more dramatic improvement in effective thermal conductivity is expected as a result of decreasing the particle size in a solution than can obtained by altering the particle shapes of large particles. a ) a a a as my
a suspension containing large particles is more than doubled by decreasing the sphericity of the particles from a value of 1.0 to 0.3 (the sphericity is defined as the ratio of the surface area of a particle with a perfectly spherical shape to that of a non-spherical particle with the same volume). Since the surface area to volume ratio is 1000 times larger for particles with a 10 nm diameter than for particles with a 10mm diameter, a much more dramatic improvement in effective thermal conductivity is expected as a result of decreasing the particle size in a solution than can obtained by altering the particle shapes of large particles. a ) a a a as my
Thermal conductivities were measured using a transient hot-wire apparatus shown schematically in Fig. 2. This system measures the electrical resistivity of the fluid systems and thermal conductivities are then calculated from the known relationship between electrical and thermal conductivities.
Thermal conductivities were measured using a transient hot-wire apparatus shown schematically in Fig. 2. This system measures the electrical resistivity of the fluid systems and thermal conductivities are then calculated from the known relationship between electrical and thermal conductivities.
Figure 3. Bright-field micrographs showing the typical agglomeration size of nanocrystalline (a) Cu produced by IGC and (b) CuO [5]. While the Cu in (a) exhibits a smaller grain size of approximately 18 nm compared to the typical grain size of 36 nm for CuO in (b), the CuO agglomerate sizes are smaller than those of Cu and thus CuO forms much more stable dispersions in deionized water, as seen in (c).
Figure 3. Bright-field micrographs showing the typical agglomeration size of nanocrystalline (a) Cu produced by IGC and (b) CuO [5]. While the Cu in (a) exhibits a smaller grain size of approximately 18 nm compared to the typical grain size of 36 nm for CuO in (b), the CuO agglomerate sizes are smaller than those of Cu and thus CuO forms much more stable dispersions in deionized water, as seen in (c).
Figure 4. Bright field TEM micrographs of nanocrystalline Al,O3 (a) produced by IGC at Argonne (average grain size < 3 nm) and (b) purchased from Nanophase Technologies [5]. The average grain size in (b) is 33 nm. Both (a) and (b) show powders prior to dispersion in deionized water, while (c) shows the powder in (a) after dispersion. Large grains are seen after dispersion. Beakers of Al,03 dispersed in HO are seen in (d). The beaker on the right contains a stable solution of powders from (b), while the powders in the beaker on the left (corresponding to (a) and (c)) settle rapidly and are thus unsuitable for heat transfer applications.
Figure 4. Bright field TEM micrographs of nanocrystalline Al,O3 (a) produced by IGC at Argonne (average grain size < 3 nm) and (b) purchased from Nanophase Technologies [5]. The average grain size in (b) is 33 nm. Both (a) and (b) show powders prior to dispersion in deionized water, while (c) shows the powder in (a) after dispersion. Large grains are seen after dispersion. Beakers of Al,03 dispersed in HO are seen in (d). The beaker on the right contains a stable solution of powders from (b), while the powders in the beaker on the left (corresponding to (a) and (c)) settle rapidly and are thus unsuitable for heat transfer applications.
Figure 7. Thermal conductivity of Cu dispersed in oil. Note that similar substantial enhancements in thermal conductivity are seen compared to the oxide-in-water systems shown in Figure 5. However, in the case of Cu-in-oil, this improved behavior is obtained with almost two orders of magnitude less dispersed nanocrystalline powder.
Figure 7. Thermal conductivity of Cu dispersed in oil. Note that similar substantial enhancements in thermal conductivity are seen compared to the oxide-in-water systems shown in Figure 5. However, in the case of Cu-in-oil, this improved behavior is obtained with almost two orders of magnitude less dispersed nanocrystalline powder.
Figure 1. Transmission electron microscopy showing the ribbon-like structure of V205,nH20 gels of the ribbons is observed when these gels are deposited onto a flat substrate. A series of 00/ peaks is observed by X-ray diffraction. The 1-D Patterson map suggests that V2Os layers are made of two V205 sheets facing each other at a distance of 2.8A [12]. The basal distance between ribbons depends on the amount of water in V20s5,nH20. It first increases by steps of about 2.8A corresponding to the intercalation of one water layer (d=8.7A for n= 0.5, d=11.5A for n=1.8 ). A continuous swelling is then observed beyond d=20A [13][14].
Figure 1. Transmission electron microscopy showing the ribbon-like structure of V205,nH20 gels of the ribbons is observed when these gels are deposited onto a flat substrate. A series of 00/ peaks is observed by X-ray diffraction. The 1-D Patterson map suggests that V2Os layers are made of two V205 sheets facing each other at a distance of 2.8A [12]. The basal distance between ribbons depends on the amount of water in V20s5,nH20. It first increases by steps of about 2.8A corresponding to the intercalation of one water layer (d=8.7A for n= 0.5, d=11.5A for n=1.8 ). A continuous swelling is then observed beyond d=20A [13][14].
Figure 2. 5l1v NMR of aqueous solutions formed via the dissolution of V2O5 in TMAOH
Figure 2. 5l1v NMR of aqueous solutions formed via the dissolution of V2O5 in TMAOH
Figure 3. Scanning electron microscopy of cluster shell polyvanadate crystals (a) (NMeq)6[V15036C1]4H20, (b) (NMe4)10[H3V 180421] 3H20 (size of the crystals ~ 300 jum)
Figure 3. Scanning electron microscopy of cluster shell polyvanadate crystals (a) (NMeq)6[V15036C1]4H20, (b) (NMe4)10[H3V 180421] 3H20 (size of the crystals ~ 300 jum)
Figure 4. Molecular structure of cluster shell polyvanadates (a) [VisO36Cl]® , (b) [VigOgaI]!3-
Figure 4. Molecular structure of cluster shell polyvanadates (a) [VisO36Cl]® , (b) [VigOgaI]!3-
Figure 5. Formation of V205,nH20 gels via the condensation of vanadic acid. (a) neutral precursor, (b) olation and oxolation lead to edge and comer sharing pyramids proceed via ion exchange or redox reactions leading to the formation of hybrid layered
Figure 5. Formation of V205,nH20 gels via the condensation of vanadic acid. (a) neutral precursor, (b) olation and oxolation lead to edge and comer sharing pyramids proceed via ion exchange or redox reactions leading to the formation of hybrid layered
Figure 1. Impedance spectra of partially sintered ceria (~ 15 nm grain size) without(w/o) contact effect and with (w/) contact resistance effect, at 550°C A: Two sides of the specimen were polished down to 1 [sm and coated with Au (0.15 um). B: One side was polished (down to 1 pm) and coated with Au. The other side was abraded with 400 grit abrasive paper. C: True sample resistance calculated from DC 4-point conductivity. Nanocrystalline ceramics are expected to have much smaller grain boundary capacitance than ~nF range, from the conventional brick layer model. The decreased capacitance can be affected by experimental artifacts, such as apparatus contributions, as reported before [4]. Also, poor contact between the specimen/electrode leads to an additional feature (i.e., an arc) as shown in Fig. 1. Poor electroding causes current-spreading at the specimen/electrode interface.
Figure 1. Impedance spectra of partially sintered ceria (~ 15 nm grain size) without(w/o) contact effect and with (w/) contact resistance effect, at 550°C A: Two sides of the specimen were polished down to 1 [sm and coated with Au (0.15 um). B: One side was polished (down to 1 pm) and coated with Au. The other side was abraded with 400 grit abrasive paper. C: True sample resistance calculated from DC 4-point conductivity. Nanocrystalline ceramics are expected to have much smaller grain boundary capacitance than ~nF range, from the conventional brick layer model. The decreased capacitance can be affected by experimental artifacts, such as apparatus contributions, as reported before [4]. Also, poor contact between the specimen/electrode leads to an additional feature (i.e., an arc) as shown in Fig. 1. Poor electroding causes current-spreading at the specimen/electrode interface.
As seen in Eq. (1), the contact resistance is a function of the conductivity (of the system under investigation) and total contact radius. The effect of contact radius can be tested by varying the mechanical contact pressure, since contact pressure changes the total contact radius. As shown in Fig. 2(a), the mechanical contact pressure controls the impedance arc at
As seen in Eq. (1), the contact resistance is a function of the conductivity (of the system under investigation) and total contact radius. The effect of contact radius can be tested by varying the mechanical contact pressure, since contact pressure changes the total contact radius. As shown in Fig. 2(a), the mechanical contact pressure controls the impedance arc at
Thermopower is dictated by the charge carrier concentration irrespective of the corresponding conduction mechanism, i.e., band conduction or small polaron conduction. Bulk cerium oxide is known to conduct through small polarons between Cet3 and Cet4 [8]. The temperature dependence of thermopower (Fig. 4(a)) was used to determine the reductior enthalpy. The calculated value of 1.84 eV is much smaller than the value of 4.67~4.98 eV ir bulk cerium oxide. In combination with the temperature dependence of the electronic conductivity (Fig. 4(b)), the enthalpy of electron mobility was determined to be 0.47 eV. which is consistent with small polaron conduction. Additionally, the calculated mobility satisfied the criterion of <0.1 cm2/Vsec [9]. Based upon the chemical analysis through Spark Source Mass Spectrometry (SSMS) and the measured thermopower, the nonstoichiometry was estimated for undoped nanocrystalline ceria. It is three orders of magnitude larger than for the bulk ceria. This effect can be introduced by space charge layer formation which can mean a different activation enthalpy for defect formation (reduction). Nanocrystalline cerium oxide possesses an extremely large fraction of grain boundaries. The effect of a space charge can be amplified and thereby alter the conductivity. Although the grain boundaries are believed to have higher conductivity and higher dielectric constant, the resultant time constant is not significantly larger than the grain interiors (or the bulk). The resultant impedance spectra can be a-single arc, as shown in this work.
Thermopower is dictated by the charge carrier concentration irrespective of the corresponding conduction mechanism, i.e., band conduction or small polaron conduction. Bulk cerium oxide is known to conduct through small polarons between Cet3 and Cet4 [8]. The temperature dependence of thermopower (Fig. 4(a)) was used to determine the reductior enthalpy. The calculated value of 1.84 eV is much smaller than the value of 4.67~4.98 eV ir bulk cerium oxide. In combination with the temperature dependence of the electronic conductivity (Fig. 4(b)), the enthalpy of electron mobility was determined to be 0.47 eV. which is consistent with small polaron conduction. Additionally, the calculated mobility satisfied the criterion of <0.1 cm2/Vsec [9]. Based upon the chemical analysis through Spark Source Mass Spectrometry (SSMS) and the measured thermopower, the nonstoichiometry was estimated for undoped nanocrystalline ceria. It is three orders of magnitude larger than for the bulk ceria. This effect can be introduced by space charge layer formation which can mean a different activation enthalpy for defect formation (reduction). Nanocrystalline cerium oxide possesses an extremely large fraction of grain boundaries. The effect of a space charge can be amplified and thereby alter the conductivity. Although the grain boundaries are believed to have higher conductivity and higher dielectric constant, the resultant time constant is not significantly larger than the grain interiors (or the bulk). The resultant impedance spectra can be a-single arc, as shown in this work.
Fig.3. Relative density of silica -doped TZP as a function of sintering temperature.
Fig.3. Relative density of silica -doped TZP as a function of sintering temperature.
frustrated transition. This central peak correlates adequately with the sharp anomaly in hypersonic damping in PMN which also is caused by fluctuations [4].
frustrated transition. This central peak correlates adequately with the sharp anomaly in hypersonic damping in PMN which also is caused by fluctuations [4].
Figure 1 (a)X-ray diffraction scans from nanocrystalline BaTiO, annealed at 700 °C in air for 2 hours. (b) X-ray diffraction pattern of standard polycrystalline BaTiO,. We employed a two-source evaporation process to simultaneously produce a homogenous mixture of partially oxidized Ba/Ti clusters. The processing technique and parameters have been reported somewhere else[5]. One important feature of the process used is that the Ti and BaTiO, source materials can be evaporated in an oxygen environment rather than the more common inter gas environment typically used with gas condensation.
Figure 1 (a)X-ray diffraction scans from nanocrystalline BaTiO, annealed at 700 °C in air for 2 hours. (b) X-ray diffraction pattern of standard polycrystalline BaTiO,. We employed a two-source evaporation process to simultaneously produce a homogenous mixture of partially oxidized Ba/Ti clusters. The processing technique and parameters have been reported somewhere else[5]. One important feature of the process used is that the Ti and BaTiO, source materials can be evaporated in an oxygen environment rather than the more common inter gas environment typically used with gas condensation.
After annealing the mixture of Ba/Ti oxidized powders at 700 °C in air for 1-2 hours, well crystallized nanocrystalline BaTiO, was obtained. Figure 1(a) is an x-ray scan for powders after annealing at 700°C for 2 hours, which indicates single phase BaTiO, is formed. The dominant phase is most likely a pseudo-cubic phase, although peak broadening due to small particle size makes it difficult to distinguish the phase from the tetragonal phase typically found for coarse grained materials. The XRD observations were also reproduced by electron diffraction. Bright- field and dark field TEM images of these nanoparticles are shown in Figure 2. The average particle size is less than 20nm. The nanocrystalline BaTiO, powders were next pelletized by cold isostatic pressing without use of a binder.
After annealing the mixture of Ba/Ti oxidized powders at 700 °C in air for 1-2 hours, well crystallized nanocrystalline BaTiO, was obtained. Figure 1(a) is an x-ray scan for powders after annealing at 700°C for 2 hours, which indicates single phase BaTiO, is formed. The dominant phase is most likely a pseudo-cubic phase, although peak broadening due to small particle size makes it difficult to distinguish the phase from the tetragonal phase typically found for coarse grained materials. The XRD observations were also reproduced by electron diffraction. Bright- field and dark field TEM images of these nanoparticles are shown in Figure 2. The average particle size is less than 20nm. The nanocrystalline BaTiO, powders were next pelletized by cold isostatic pressing without use of a binder.
Figure 4 The dependence of dielectric constants and losses on temperature for the sample made b' nanocrystalline BaTiO, from GC
Figure 4 The dependence of dielectric constants and losses on temperature for the sample made b' nanocrystalline BaTiO, from GC
Figure 3 Plot of the density as a function of sintering temperature of nanocrystalline BaTiO, pellets.
Figure 3 Plot of the density as a function of sintering temperature of nanocrystalline BaTiO, pellets.
Figure 5 P vs. E hysteresis behavior of BaTiO, ceramics sintered at 1200 °C/2h.
Figure 5 P vs. E hysteresis behavior of BaTiO, ceramics sintered at 1200 °C/2h.
Figure 6 (a) Thermogravimetric analysis and differential thermal analysis curves of the mixture of Ba/Ti oxidized clusters. Figure 6 (b) TGA of the mixture of Ba/Ti clusters at different heating rates.
Figure 6 (a) Thermogravimetric analysis and differential thermal analysis curves of the mixture of Ba/Ti oxidized clusters. Figure 6 (b) TGA of the mixture of Ba/Ti clusters at different heating rates.
The structure of the samples annealed below 300°C was amorphous (Fig. 2A) and Fig. 1 TGA curve of sol-gel zirconia prepared with different hydrolysis catalysts.
The structure of the samples annealed below 300°C was amorphous (Fig. 2A) and Fig. 1 TGA curve of sol-gel zirconia prepared with different hydrolysis catalysts.
Fig. 2 X-ray diffraction curves of sol-gel zirconia prepared with acetic acid as hydrolysis catalyst. Tick marks correspond to the tetragonal phase of zirconia.
Fig. 2 X-ray diffraction curves of sol-gel zirconia prepared with acetic acid as hydrolysis catalyst. Tick marks correspond to the tetragonal phase of zirconia.
Table 1. Non-Sulfated Zirconia. Phase Composition and Average Crystallite Size as a Function of Hydrolysis Catalyst and Temperature
Table 1. Non-Sulfated Zirconia. Phase Composition and Average Crystallite Size as a Function of Hydrolysis Catalyst and Temperature
Fig. 3 Rietveld refinement plot for sol-gel zirconia prepared with HCI and annealed at 600 °C. It has the tetragonal (upper tick marks) and monoclinic phases (lower tick marks). Fig. 3 Rietveld refinement plot for sol-gel zirconia
Fig. 3 Rietveld refinement plot for sol-gel zirconia prepared with HCI and annealed at 600 °C. It has the tetragonal (upper tick marks) and monoclinic phases (lower tick marks). Fig. 3 Rietveld refinement plot for sol-gel zirconia
Fig. 5 TGA curve of sulfated sol-gel zirconia prepared with different hydrolysis catalysts. Sulfated samples started loosing part of its weight when they were annealed above 500 °C. They had one transformation around 500 °C and a second one around 600 °C (Fig. 5). After annealing the sample at 800 °C, the total weight loss depended on the hydrolysis catalyst used in the preparation; it was 31.7% for HCl, 29.3% for C2H,OQ2, and 41.5% for NH,OH. In the transformation at 500 °C, the sulfated samples prepared with HCl and NH,OH as hydrolysis catalyst only lost 2.4 and 14.7% in weight respectively; the rest of the weight loss occurred above 600 °C. In contrast to this, the samples prepared with acetic acid lost most of their weight (24.1 of 29.3%) in the first transformation. If the weight loss, observed around 500 °C and above 600 °C, was associated to the evaporation of SO, ions from the sample, then, the above results will suggest that these SO, ions had a strong interaction with the sol-gel zirconia prepared with hydrochloric
Fig. 5 TGA curve of sulfated sol-gel zirconia prepared with different hydrolysis catalysts. Sulfated samples started loosing part of its weight when they were annealed above 500 °C. They had one transformation around 500 °C and a second one around 600 °C (Fig. 5). After annealing the sample at 800 °C, the total weight loss depended on the hydrolysis catalyst used in the preparation; it was 31.7% for HCl, 29.3% for C2H,OQ2, and 41.5% for NH,OH. In the transformation at 500 °C, the sulfated samples prepared with HCl and NH,OH as hydrolysis catalyst only lost 2.4 and 14.7% in weight respectively; the rest of the weight loss occurred above 600 °C. In contrast to this, the samples prepared with acetic acid lost most of their weight (24.1 of 29.3%) in the first transformation. If the weight loss, observed around 500 °C and above 600 °C, was associated to the evaporation of SO, ions from the sample, then, the above results will suggest that these SO, ions had a strong interaction with the sol-gel zirconia prepared with hydrochloric
Table 2. Sulfated Zirconia. Phase Composition and Average Crystallite Size as a Function of Hydrolysis Catalyst and Temperature acid and ammonium hydroxide, and a weak interaction with the sol-gel zirconia prepared with acetic acid.
Table 2. Sulfated Zirconia. Phase Composition and Average Crystallite Size as a Function of Hydrolysis Catalyst and Temperature acid and ammonium hydroxide, and a weak interaction with the sol-gel zirconia prepared with acetic acid.
ure 1 Schematics of zeolite and sodalite cage frameworks. Zeolites are constructed from TO, tetrahedra and each apical oxygen atom is shared with an adjacent tetrahedron [1-3]. Silicon atom forms bonds with four neighboring atoms in a
ure 1 Schematics of zeolite and sodalite cage frameworks. Zeolites are constructed from TO, tetrahedra and each apical oxygen atom is shared with an adjacent tetrahedron [1-3]. Silicon atom forms bonds with four neighboring atoms in a
S1X kinds of zeolites (Zeolite A, X, Y, Mordenite, ZSM-5 and Ferrierite supplied by TOSOH), crystal-SiO, (a- quartz), amorphous-SiO, (quartz glass and ultrafine particle), and a-Al,O, were ablated as the targets by an Nd:YAG laser (the second harmonic wave 532nm and the third harmonic wave 355nm). The structures of zeolites and the Si/Al ratios are listed in TABLE I. The zeolites and SiO, pellet were 20mm in diameter and ~Smm in thickness. After sintered in an electric furnace at 523K for 10 hours, these pellets were set on the electric plate of the acceleration region and ablation was performed. Other materials were ground and the pellets were prepared, BLEI. The Si/AI ratios of zeolites which were used in this experiment ablated as the targets by an Nd:YAG
S1X kinds of zeolites (Zeolite A, X, Y, Mordenite, ZSM-5 and Ferrierite supplied by TOSOH), crystal-SiO, (a- quartz), amorphous-SiO, (quartz glass and ultrafine particle), and a-Al,O, were ablated as the targets by an Nd:YAG laser (the second harmonic wave 532nm and the third harmonic wave 355nm). The structures of zeolites and the Si/Al ratios are listed in TABLE I. The zeolites and SiO, pellet were 20mm in diameter and ~Smm in thickness. After sintered in an electric furnace at 523K for 10 hours, these pellets were set on the electric plate of the acceleration region and ablation was performed. Other materials were ground and the pellets were prepared, BLEI. The Si/AI ratios of zeolites which were used in this experiment ablated as the targets by an Nd:YAG
were ablated, these peaks were weaker, but these two peaks appeared clearly from the Ferrierite. The characteristic peaks from zeolite were m/z=148 and 163. These two peaks appeared from all ablated zeolites. The mass spectra due to the zeolites did not depend on the Si/AI ratios (see Laser : Nd:YAG 532nm : Laser power : 255mJ pulse"! em-2
were ablated, these peaks were weaker, but these two peaks appeared clearly from the Ferrierite. The characteristic peaks from zeolite were m/z=148 and 163. These two peaks appeared from all ablated zeolites. The mass spectra due to the zeolites did not depend on the Si/AI ratios (see Laser : Nd:YAG 532nm : Laser power : 255mJ pulse"! em-2
Zeolite A was ablated by Nd:YAG laser and five peaks aré obtained up to m/z = 200. The peaks at m/z = 28, 44, 70, 148 and 163 were assigned to T*, TO*, T20*, T3047 and T305", respectively.
Zeolite A was ablated by Nd:YAG laser and five peaks aré obtained up to m/z = 200. The peaks at m/z = 28, 44, 70, 148 and 163 were assigned to T*, TO*, T20*, T3047 and T305", respectively.
Zeolite X and Y. They have the same topology.) In this experiment, three kinds of Zeolite A ( The cations in the structure were different, K*, Na* and Ca’* ) were used, but the cations in the structures have no influence on the spectra.
Zeolite X and Y. They have the same topology.) In this experiment, three kinds of Zeolite A ( The cations in the structure were different, K*, Na* and Ca’* ) were used, but the cations in the structures have no influence on the spectra.
Table I: Samples and Heat Treatments
Table I: Samples and Heat Treatments
Figure 2: Representative example of an impedance spectrum for a-CeO, ,
Figure 2: Representative example of an impedance spectrum for a-CeO, ,
Annealing causes the relative size of the boundary arc to increase as shown in Figure 2. The coarsest sample, c-CeO,, has a still larger boundary arc (Figure 3) which is typical of polycrystalline ionic conductors in which grain boundary impedance has been attributed to impurity segregation [6, 7, 8].
Annealing causes the relative size of the boundary arc to increase as shown in Figure 2. The coarsest sample, c-CeO,, has a still larger boundary arc (Figure 3) which is typical of polycrystalline ionic conductors in which grain boundary impedance has been attributed to impurity segregation [6, 7, 8].
Figure 4(a) shows the conductivity versus PO, results. Both the n-CeO,, and the a-CeO, , exhibit a PO, dependence between -1/6 and -1/4. The coarse sample, c-CeO,, however, exhibits a very shallow slope of approximately -1/20. The nearly PO,-independent behavior suggests extrinsic, ionic behavior. The slight PO, dependence may result from reduction of the smaller grains which are only a small fraction of the whole and therefore have a minor effect on the total conductivity. The c-CeO, sample behaves nearly extrinsically under the same conditions where the nanocrystalline materials show behavior consistent with reduction. Figure 4 (b) shows the temperature dependence of the conductivity for these samples along with their activation eneroles The coarecened camnile c-Cef) chowe an activation eneroy aff 322
Figure 4(a) shows the conductivity versus PO, results. Both the n-CeO,, and the a-CeO, , exhibit a PO, dependence between -1/6 and -1/4. The coarse sample, c-CeO,, however, exhibits a very shallow slope of approximately -1/20. The nearly PO,-independent behavior suggests extrinsic, ionic behavior. The slight PO, dependence may result from reduction of the smaller grains which are only a small fraction of the whole and therefore have a minor effect on the total conductivity. The c-CeO, sample behaves nearly extrinsically under the same conditions where the nanocrystalline materials show behavior consistent with reduction. Figure 4 (b) shows the temperature dependence of the conductivity for these samples along with their activation eneroles The coarecened camnile c-Cef) chowe an activation eneroy aff 322
Fig. 3. TEM BF images from (a) hydrothermal BaFe;20j 9 heat treated at 850°C and (b) hydrothermal BaFe;2019 at 95°C displaying the distribution of crystalline nanophase particles. The EEL spectra in (c) are from the hexagonal face-on suspended particles A and B in Figures (a) and (b) respectively.
Fig. 3. TEM BF images from (a) hydrothermal BaFe;20j 9 heat treated at 850°C and (b) hydrothermal BaFe;2019 at 95°C displaying the distribution of crystalline nanophase particles. The EEL spectra in (c) are from the hexagonal face-on suspended particles A and B in Figures (a) and (b) respectively.
Figure 1. Gas-phase number size distributions of iron oxide particles produced in air (a), N2 (b) or H2/N> (c) at different temperatures.
Figure 1. Gas-phase number size distributions of iron oxide particles produced in air (a), N2 (b) or H2/N> (c) at different temperatures.
Figure 1: XRD powder patterns of calcined powders obtained at different carboxylic group to cation ratio, for pH =7.2.
Figure 1: XRD powder patterns of calcined powders obtained at different carboxylic group to cation ratio, for pH =7.2.
Figure 2: XRD powder patterns of calcined powders obtained at various pH, with carboxylic group to cation ratio kept constant at 3.5. rroup to cation ratio held at 3.5:1, shows a average crystallite size of less than 10nm. One has to also note here that for obtaining phase pure powders it was very important that calcining of the dried precipitates not be done at temperatures higher than 400 °C, because of a tendency of the phase to decompose into hematite and another cubic spinel phase (CuFe20,). Tha racultae af SOTITD eatunratinn maanatizatinn at 19h far threasa nntuaore ctahiiienad «+ nn nat
Figure 2: XRD powder patterns of calcined powders obtained at various pH, with carboxylic group to cation ratio kept constant at 3.5. rroup to cation ratio held at 3.5:1, shows a average crystallite size of less than 10nm. One has to also note here that for obtaining phase pure powders it was very important that calcining of the dried precipitates not be done at temperatures higher than 400 °C, because of a tendency of the phase to decompose into hematite and another cubic spinel phase (CuFe20,). Tha racultae af SOTITD eatunratinn maanatizatinn at 19h far threasa nntuaore ctahiiienad «+ nn nat
Figure 6: High temperature x-ray plot showing reaction occuring in nanosize divalent copper ferrite with temperature on slow heating.
Figure 6: High temperature x-ray plot showing reaction occuring in nanosize divalent copper ferrite with temperature on slow heating.
Fig. 1 Images and selected area electron diffraction patterns of polyhedral nanophase TiO, (scale applies to both images). The anatase particle (a, b) is oriented along [010]. The rutile particle (c, d) is oriented along [111] and contains a twin on (011).
Fig. 1 Images and selected area electron diffraction patterns of polyhedral nanophase TiO, (scale applies to both images). The anatase particle (a, b) is oriented along [010]. The rutile particle (c, d) is oriented along [111] and contains a twin on (011).
Unusual particles were noted in two of the sample sets (ow and medium temperatures, lower height in flame). These particles are generally 100 nm or larger and are much larger than the majority of particles in the sample sets. They have central features that are rounded or faceted as shown in the examples of Figures 2a,b. Tilting of the particles showed that the central features are in the interior of the particles and energy dispersive x-ray analysis confirmed that the particles are TiO, SAED patterns obtained thus far are consistent with an identification of rutile. Rarely, these particles show twinning on {011} as shown in Figures 2b-d.
Unusual particles were noted in two of the sample sets (ow and medium temperatures, lower height in flame). These particles are generally 100 nm or larger and are much larger than the majority of particles in the sample sets. They have central features that are rounded or faceted as shown in the examples of Figures 2a,b. Tilting of the particles showed that the central features are in the interior of the particles and energy dispersive x-ray analysis confirmed that the particles are TiO, SAED patterns obtained thus far are consistent with an identification of rutile. Rarely, these particles show twinning on {011} as shown in Figures 2b-d.
to longer wavelength. The above-mentioned change in the behavior of line profiles for 0.1% Eu”: Y20; is negligible, and the excitation lines remain sharp and symmetric. The excitation peaks corresponding to sites B and C are resolved in this spectrum, unlike the spectrum of th: Fig.2. ’Fo->°"Dp excitation spectra of nanocrystalline (a) EuzO3 (b) 5.0% Eu’: Y20; and (c) 0.1% Eu: Y20; monitoring “D)—>’F, fluorescence at 624.0 nm using wide-band 0.25 m
to longer wavelength. The above-mentioned change in the behavior of line profiles for 0.1% Eu”: Y20; is negligible, and the excitation lines remain sharp and symmetric. The excitation peaks corresponding to sites B and C are resolved in this spectrum, unlike the spectrum of th: Fig.2. ’Fo->°"Dp excitation spectra of nanocrystalline (a) EuzO3 (b) 5.0% Eu’: Y20; and (c) 0.1% Eu: Y20; monitoring “D)—>’F, fluorescence at 624.0 nm using wide-band 0.25 m
Fig. 1 X-ray diffraction of monoclinic 0.1% Eu**:¥203 nanocrystals (average size of 23 nm). nanocrystals. The excitation spectra in Fig. 2 were obtained by monitoring “Do > "F, fluorescence at 624 nm with the wide-band 0.25 m monochromator. We recorded the excitation spectra with short and long boxcar gate widths and delays to uncover any discrimination among the various sites and phases that might be present. At a boxcar delay of 20 us the excitation spectrum for 5% Eu”: Y.03 shows broadened and asymmetric lines with higher intensity towards the Eu,O; transitions. At a boxcar delay of 200 ms, the lines appear symmetric and the line positions shifted
Fig. 1 X-ray diffraction of monoclinic 0.1% Eu**:¥203 nanocrystals (average size of 23 nm). nanocrystals. The excitation spectra in Fig. 2 were obtained by monitoring “Do > "F, fluorescence at 624 nm with the wide-band 0.25 m monochromator. We recorded the excitation spectra with short and long boxcar gate widths and delays to uncover any discrimination among the various sites and phases that might be present. At a boxcar delay of 20 us the excitation spectrum for 5% Eu”: Y.03 shows broadened and asymmetric lines with higher intensity towards the Eu,O; transitions. At a boxcar delay of 200 ms, the lines appear symmetric and the line positions shifted
Contrary to pure Eu systems, the probability of depopulation through ion-ion interaction decreases as the concentration of europium ions decreases in doped systems. At sufficiently low concentration the decay mechanism is mainly radiative. This results in longer and almost exponential decay curves. In the 0.1% Eu: Y2Os nanocrystals, although the energy difference between “Dp levels for site B and C is less than the thermal energy at 10 K interaction between sites B and C appears very weak. In the cubic system also no observable interactions between Eu” ions have been reported for concentrations below 0.1% of Eu [14]. However, we notice significant energy transfer from site A to B and C in nanocrystalline monoclinic Eu*: Y203. Phonon-assisted energy transfer could be a possible explanation of this observed behavior [15]. Because of small energy mismatch between the “Dp levels of Eu at sites B and C, the one- phonon assisted energy transfer between these two sites will be inhibited as compared to transfer from site A to B and C. The fluorescence from site A appears to be inefficient as compared to site B and C. The decay characteristic of site A is rather complex and fluorescence from this site appears inefficient as compared to sites B and C. Table 1 Fluorescence decay times in monoclinic 0.1% Eu: Y20; at 10 K.
Contrary to pure Eu systems, the probability of depopulation through ion-ion interaction decreases as the concentration of europium ions decreases in doped systems. At sufficiently low concentration the decay mechanism is mainly radiative. This results in longer and almost exponential decay curves. In the 0.1% Eu: Y2Os nanocrystals, although the energy difference between “Dp levels for site B and C is less than the thermal energy at 10 K interaction between sites B and C appears very weak. In the cubic system also no observable interactions between Eu” ions have been reported for concentrations below 0.1% of Eu [14]. However, we notice significant energy transfer from site A to B and C in nanocrystalline monoclinic Eu*: Y203. Phonon-assisted energy transfer could be a possible explanation of this observed behavior [15]. Because of small energy mismatch between the “Dp levels of Eu at sites B and C, the one- phonon assisted energy transfer between these two sites will be inhibited as compared to transfer from site A to B and C. The fluorescence from site A appears to be inefficient as compared to site B and C. The decay characteristic of site A is rather complex and fluorescence from this site appears inefficient as compared to sites B and C. Table 1 Fluorescence decay times in monoclinic 0.1% Eu: Y20; at 10 K.
Figure 1: The average grain size and microstrains as determined from X-ray line broadening as function of milling time for iron powder From wide angle X-ray spectra, the information about lattice defects (grain boundaries, dislocations etc.) is obtained via their disturbing influence on the coherent superposition of radiation diffracted at the atomic lattice sites which causes the broadening of Bragg peaks. In small angle scattering experiments the lattice defects themselves give rise to a scattering contrast because of the (scattering length) density fluctuations associated with them. aia a sa aa wii a am mes te
Figure 1: The average grain size and microstrains as determined from X-ray line broadening as function of milling time for iron powder From wide angle X-ray spectra, the information about lattice defects (grain boundaries, dislocations etc.) is obtained via their disturbing influence on the coherent superposition of radiation diffracted at the atomic lattice sites which causes the broadening of Bragg peaks. In small angle scattering experiments the lattice defects themselves give rise to a scattering contrast because of the (scattering length) density fluctuations associated with them. aia a sa aa wii a am mes te
After 30 hs milling time, a drop of the scattering intensity is observed over the whole q-r ympared to the 0.5 hs sample. Especially the drop at high q values is surprising since fror y measurements it is known that the average grain size is further reduced to a volume ave ‘about 16 nm which is good agreement with the average grain size derived from the RD out 15 nm after 30 hs milling. Furthermore, the RDF also shows a decrease in magni ithout any significant shift of their maxima (Fig. 3). bviously the observed change of the SANS intensity cannot be solely explained by a shi ystallite size distribution to smaller distances in real space during the milling process. Inste believed that the scattering contrast due to dislocations plays an important role for terpretation of the measured data. Since the volume dilatation in the vicinity of dislocatio nall their nuclear scattering contrast is small and disappears for pure screw dislocations | heir magnetic scattering contrast due to orientation fluctuations of the magnetic moments ¢ oms may exceed by factors up to 10-100 their contribution to nuclear scattering [15]. T uctuations are caused by magnetoelastic coupling between the magnetic moments and slocation strain field.
After 30 hs milling time, a drop of the scattering intensity is observed over the whole q-r ympared to the 0.5 hs sample. Especially the drop at high q values is surprising since fror y measurements it is known that the average grain size is further reduced to a volume ave ‘about 16 nm which is good agreement with the average grain size derived from the RD out 15 nm after 30 hs milling. Furthermore, the RDF also shows a decrease in magni ithout any significant shift of their maxima (Fig. 3). bviously the observed change of the SANS intensity cannot be solely explained by a shi ystallite size distribution to smaller distances in real space during the milling process. Inste believed that the scattering contrast due to dislocations plays an important role for terpretation of the measured data. Since the volume dilatation in the vicinity of dislocatio nall their nuclear scattering contrast is small and disappears for pure screw dislocations | heir magnetic scattering contrast due to orientation fluctuations of the magnetic moments ¢ oms may exceed by factors up to 10-100 their contribution to nuclear scattering [15]. T uctuations are caused by magnetoelastic coupling between the magnetic moments and slocation strain field.
gure 3: Radial distribution functions (volume weighted) calculated from spectra in Fig. 2 (milling times 0.5 hs, 30 hs)
gure 3: Radial distribution functions (volume weighted) calculated from spectra in Fig. 2 (milling times 0.5 hs, 30 hs)
Specific heat increase at room temperature in comparison to the unmilled state as function of the stored enthalpy AH (given as percentage of AH) for bec and hcp nanocrystalline metals Figure 4: Large differences generally also arise in the specific heat c, at constant pressure. The specific heat of the heavily deformed powder particles was measured in the range from 130 K to 300 K i.e, at low enough temperatures to prevent the recovery processes from taking place. For al samples, a considerable increase in c, has been found experimentally after 24 hours milling reaching values up to 15% for Ru. These data are also included in Table I given as a percentage of heat capacity increase in comparison to the unmilled state at 300 K. For pure metals, a linea: correlation between the heat capacity change Ac, and the stored enthalpy AH given as : percentage of the heat of fusion (AH/AH,) after extended mechanical attrition is observed (Fig 4). Such a relationship is also predicted by the free volume model for grain boundaries.
Specific heat increase at room temperature in comparison to the unmilled state as function of the stored enthalpy AH (given as percentage of AH) for bec and hcp nanocrystalline metals Figure 4: Large differences generally also arise in the specific heat c, at constant pressure. The specific heat of the heavily deformed powder particles was measured in the range from 130 K to 300 K i.e, at low enough temperatures to prevent the recovery processes from taking place. For al samples, a considerable increase in c, has been found experimentally after 24 hours milling reaching values up to 15% for Ru. These data are also included in Table I given as a percentage of heat capacity increase in comparison to the unmilled state at 300 K. For pure metals, a linea: correlation between the heat capacity change Ac, and the stored enthalpy AH given as : percentage of the heat of fusion (AH/AH,) after extended mechanical attrition is observed (Fig 4). Such a relationship is also predicted by the free volume model for grain boundaries.
By simple geometric considerations [35,36] the specific grain boundary excess enthalpy is estimated to be about 2.1 J/m’. This would correspond to a value for non-equilibrium unrelaxed grain boundaries, whereas after relaxation, the grain boundary energy is reduced to 1.5 J/m’. Values resulting from computer simulations suggest excess enthalpies between 1.2 and 1.8 J/m’ it has been found that after a fast increase at early times the average grain size d changes from 16 nm to about 30 - 40 nm. The average grain size remains constant for t > 2400 sec and reaches values of 100 - 200 nm at temperatures about 600 °C. As such, two regimes with and without grain growth can be distinguished. However, since the influence of lattice point defects and lattice dislocations is negligible, the enthalpy release can be clearly assigned to the existence of grain boundaries. The reduction of the microstrains are probably caused by grain boundary relaxation and annihilation of secondary grain boundary dislocations. Based on elastic theory it is estimated that this contribution to the overall energy is less than about 5%. On this basis, the grain boundary energy can be estimated.
By simple geometric considerations [35,36] the specific grain boundary excess enthalpy is estimated to be about 2.1 J/m’. This would correspond to a value for non-equilibrium unrelaxed grain boundaries, whereas after relaxation, the grain boundary energy is reduced to 1.5 J/m’. Values resulting from computer simulations suggest excess enthalpies between 1.2 and 1.8 J/m’ it has been found that after a fast increase at early times the average grain size d changes from 16 nm to about 30 - 40 nm. The average grain size remains constant for t > 2400 sec and reaches values of 100 - 200 nm at temperatures about 600 °C. As such, two regimes with and without grain growth can be distinguished. However, since the influence of lattice point defects and lattice dislocations is negligible, the enthalpy release can be clearly assigned to the existence of grain boundaries. The reduction of the microstrains are probably caused by grain boundary relaxation and annihilation of secondary grain boundary dislocations. Based on elastic theory it is estimated that this contribution to the overall energy is less than about 5%. On this basis, the grain boundary energy can be estimated.
ure 8: Hardness measurement using a nanoindentation device on polycrystalline (upper curves) and nanocrystalline (lower curves) attrited iron powder samples As a further consequence of the grain size reduction a drastic change in the mechanical properties has been observed. Local mechanical properties can be measured by nano-indentation methods. Here the load as well as the indentation depths are monitored continuously during the loading and unloading process (Fig. 8). Typical results for nanocrystalline Fe powder samples exhibit an increase in hardness by a factor 7 (9.3 GPa for nx-Fe with d about 16 nm versus 1.3 GPa for annealed px-Fe). In general, the hardness follows a trend similar to the Hall-Petch relationship though the deformation mechanism in the nanocrystalline regime remains unclear. Furthermore, the Youngs modulus can be measured by this method as well and shows a decrease by 10 - 20 % in comparison with the polycrystal. Therefore, it is suggested that the mechanical properties of nanophase materials prepared by mechanical attrition after extended periods of milling are not being controlled by the plasticity of the crystal due to dislocation movement anymore but rather by the cohesion of the nanocrystalline material across its grain boundaries. From the considerable increase of hardness and the principal changes of the deformation
ure 8: Hardness measurement using a nanoindentation device on polycrystalline (upper curves) and nanocrystalline (lower curves) attrited iron powder samples As a further consequence of the grain size reduction a drastic change in the mechanical properties has been observed. Local mechanical properties can be measured by nano-indentation methods. Here the load as well as the indentation depths are monitored continuously during the loading and unloading process (Fig. 8). Typical results for nanocrystalline Fe powder samples exhibit an increase in hardness by a factor 7 (9.3 GPa for nx-Fe with d about 16 nm versus 1.3 GPa for annealed px-Fe). In general, the hardness follows a trend similar to the Hall-Petch relationship though the deformation mechanism in the nanocrystalline regime remains unclear. Furthermore, the Youngs modulus can be measured by this method as well and shows a decrease by 10 - 20 % in comparison with the polycrystal. Therefore, it is suggested that the mechanical properties of nanophase materials prepared by mechanical attrition after extended periods of milling are not being controlled by the plasticity of the crystal due to dislocation movement anymore but rather by the cohesion of the nanocrystalline material across its grain boundaries. From the considerable increase of hardness and the principal changes of the deformation
The Ti-Al alloy deposits produced by the evaporator were between 50 [um and 400 jum in thickness after deposition runs of between 45 mins and 5 hrs, and were deposited at a rate of 1-3 ym per minute.
The Ti-Al alloy deposits produced by the evaporator were between 50 [um and 400 jum in thickness after deposition runs of between 45 mins and 5 hrs, and were deposited at a rate of 1-3 ym per minute.
imaging, and is directly related to compositional fluctuations. TEM observations on a Ti-40% Al alloy, showed a grain size of 20-50 nm in the as-deposited condition. Fig. 2. As-deposited Ti-40 to 50% Al alloy. (a) and (b) show the deposit produced at low rotation rate (500 r.p.m). Note the defect content within the material and layered nature of its appearance in the SEM (BSE image). Selected area diffraction patterns from the o, grains from planar sections parallel to the growth direction showed a nearly (0001) basal orientation diffraction pattern indicative of the c-axis orientation of the hexagonal ©, phase parallel to the growth axis (Fig. 3). Voids, 5-10 nm size, were also seen in the TEM. The grain boundary ‘pinning’ from these and the 2-phase contributes to retaining fine grain size in these alloys.
imaging, and is directly related to compositional fluctuations. TEM observations on a Ti-40% Al alloy, showed a grain size of 20-50 nm in the as-deposited condition. Fig. 2. As-deposited Ti-40 to 50% Al alloy. (a) and (b) show the deposit produced at low rotation rate (500 r.p.m). Note the defect content within the material and layered nature of its appearance in the SEM (BSE image). Selected area diffraction patterns from the o, grains from planar sections parallel to the growth direction showed a nearly (0001) basal orientation diffraction pattern indicative of the c-axis orientation of the hexagonal ©, phase parallel to the growth axis (Fig. 3). Voids, 5-10 nm size, were also seen in the TEM. The grain boundary ‘pinning’ from these and the 2-phase contributes to retaining fine grain size in these alloys.
Effect of Heat-treatment and Consolidation
Effect of Heat-treatment and Consolidation
Fig. 4. Ti 40 to 50% AI (500 r.p.m), heat treated for 2 hrs at 700°C. Note that phase separation o: fine eqiuaxed gamma particles has occurred in (a) and (b). Note void healing and disruption of the normal microstructure in the regions which contained defective material. as-deposited 2-phase o,+ y alloys, now appeared to be healed after heat treatment, with affected regions apparent by disruption of the ordered microstructure with Al rich deposits at the sites of the
Fig. 4. Ti 40 to 50% AI (500 r.p.m), heat treated for 2 hrs at 700°C. Note that phase separation o: fine eqiuaxed gamma particles has occurred in (a) and (b). Note void healing and disruption of the normal microstructure in the regions which contained defective material. as-deposited 2-phase o,+ y alloys, now appeared to be healed after heat treatment, with affected regions apparent by disruption of the ordered microstructure with Al rich deposits at the sites of the
Fig. 5. Fracture surface observations of consolidated Ti-40 to 50%AI. (a) shows a general view of the fracture surface (b) shows the predominantly intergranular nature of the failure, with cleavage of occasional thin layers of brittle o,- Ti,Al. that a surface flaw had initiated failure. Isolated thin bands of material which had failed by formation of cleavage steps were occasionally observed, due to compositional variations within the material which resulted in formation of thin layers of brittle Ti,A]. Compression testing of stacl consolidated deposits revealed that initial plastic yielding of the material occurred at approximately 900 MPa, which is higher than for conventionally processed near gamma material [2].
Fig. 5. Fracture surface observations of consolidated Ti-40 to 50%AI. (a) shows a general view of the fracture surface (b) shows the predominantly intergranular nature of the failure, with cleavage of occasional thin layers of brittle o,- Ti,Al. that a surface flaw had initiated failure. Isolated thin bands of material which had failed by formation of cleavage steps were occasionally observed, due to compositional variations within the material which resulted in formation of thin layers of brittle Ti,A]. Compression testing of stacl consolidated deposits revealed that initial plastic yielding of the material occurred at approximately 900 MPa, which is higher than for conventionally processed near gamma material [2].
High temperature tensile tests were conducted on the two phase alloy at a temperature of 850°C and stain rate 5x10° s'. The material showed a flow strength of about 20 MPa, and failure elongation of about 90% engineering strain. This was a lower limit failure strain since rupture invariably
High temperature tensile tests were conducted on the two phase alloy at a temperature of 850°C and stain rate 5x10° s'. The material showed a flow strength of about 20 MPa, and failure elongation of about 90% engineering strain. This was a lower limit failure strain since rupture invariably
Table I. Reaction conditions and characterization results for aluminum powders prepared from AIH,e(N(CH,),),.
Table I. Reaction conditions and characterization results for aluminum powders prepared from AIH,e(N(CH,),),.
Fig. 1. Scanning electron micrographs of aluminum powders. H-3 is a commercial aluminum powder, while the other three powders were prepared from AlH,e(NMe,), at 110°C using 0.25% Ti(OCH(CH,).), (B), 2.2% Ti(OCH(CH,),), (F), and 0.40% TiCl, (M). All pictures originally at 20,000X magnification. The particle size is most sensitive to catalyst concentration among the variables examined (see entries E and F in Table I). The particle size roughly doubles for an order of magnitude decrease it catalyst concentration. Thus, the number of particles, which varies as the inverse third power of particle size, appears to be roughly proportional to catalyst concentration over the range 0.023- 2.2% Ti(O-i-Pr),. This indicates that the catalyst affects the nucleation of aluminum particles.
Fig. 1. Scanning electron micrographs of aluminum powders. H-3 is a commercial aluminum powder, while the other three powders were prepared from AlH,e(NMe,), at 110°C using 0.25% Ti(OCH(CH,).), (B), 2.2% Ti(OCH(CH,),), (F), and 0.40% TiCl, (M). All pictures originally at 20,000X magnification. The particle size is most sensitive to catalyst concentration among the variables examined (see entries E and F in Table I). The particle size roughly doubles for an order of magnitude decrease it catalyst concentration. Thus, the number of particles, which varies as the inverse third power of particle size, appears to be roughly proportional to catalyst concentration over the range 0.023- 2.2% Ti(O-i-Pr),. This indicates that the catalyst affects the nucleation of aluminum particles.
Fig. 3. Dependence of active aluminum content on particle size. The curve was calculated based on uniform spherical particles with a 3.68 nm oxide layer. TGA data was used to determine the weight percent of active aluminum for each powder.
Fig. 3. Dependence of active aluminum content on particle size. The curve was calculated based on uniform spherical particles with a 3.68 nm oxide layer. TGA data was used to determine the weight percent of active aluminum for each powder.
The molecular silver nanocrystals under investigation here are from the same class of materials that has enjoyed rapidly growing interest in the past several years. They are charge-neutral entities comprised entirely of an elemental metal (Au, Ag, Pt) core and a dense mantle of alkylthiol(ate) surfactant groups. Elemental Ag is evaporated at temperatures ranging from 1200 to 1500 K into a flowing, preheated atmosphere of ultrahigh purity helium. The flowstream is cooled over a short distance (and flow-time) to ca. 400 K, stimulating growth of nanocrystals in the desired size- range, here ca. 4 - 6 nm. Growth is abruptly terminated by expansion through a conical funnel accompanied by dilution in a great excess of cool helium. Depending on conditions, most of the material produced is soluble in non-polar solvents, from which excess surfactant can be removed. The soluble material is found to consist entirely of nanocrystal Ag molecules (NCAM s). A Hitachi HF-2000 cold field emission gun TEM (200 kV) has been used to perform imaging and diffraction experiments. Each TEM specimen was prepared by depositing a highly concentrated drop of toluene solution of NCAMs on an ultrathin amorphous carbon film substrate, and allowing it to dry slowly in air prior to TEM observations. Nanocrystal superlattices, such as those shown in figure 1, appear on the carbon film, in the form of highly-oriented thin crystalline films, whose lateral dimensions can be as large as 1 micrometer. Our first objective has been to establish the packing structure of the nanocrystal superlattice Figure 2a shows a TEM image of an ultrathin (monolayer) film of 5.0 nm diameter NCAM: (6.4+0.2 nm center-to-center spacing), in which 2-D close-packing is deduced from the clear six fold symmetry. The electron diffraction pattern obtained from this region reveals the periodic structure of the nanocrystal lattices, and also provide an absolute calibration of the superlattice constant, because the fcc-Ag {111} ring is recorded simultaneously in the same photograph (figur 2b). Figures 2c and 2d show a TEM image and the corresponding diffraction patter corresponding to the [110] orientation, thus confirming that the superlattice has the fcc structure Such [110],-oriented films are the most frequently observed structure in our specimens, and are
The molecular silver nanocrystals under investigation here are from the same class of materials that has enjoyed rapidly growing interest in the past several years. They are charge-neutral entities comprised entirely of an elemental metal (Au, Ag, Pt) core and a dense mantle of alkylthiol(ate) surfactant groups. Elemental Ag is evaporated at temperatures ranging from 1200 to 1500 K into a flowing, preheated atmosphere of ultrahigh purity helium. The flowstream is cooled over a short distance (and flow-time) to ca. 400 K, stimulating growth of nanocrystals in the desired size- range, here ca. 4 - 6 nm. Growth is abruptly terminated by expansion through a conical funnel accompanied by dilution in a great excess of cool helium. Depending on conditions, most of the material produced is soluble in non-polar solvents, from which excess surfactant can be removed. The soluble material is found to consist entirely of nanocrystal Ag molecules (NCAM s). A Hitachi HF-2000 cold field emission gun TEM (200 kV) has been used to perform imaging and diffraction experiments. Each TEM specimen was prepared by depositing a highly concentrated drop of toluene solution of NCAMs on an ultrathin amorphous carbon film substrate, and allowing it to dry slowly in air prior to TEM observations. Nanocrystal superlattices, such as those shown in figure 1, appear on the carbon film, in the form of highly-oriented thin crystalline films, whose lateral dimensions can be as large as 1 micrometer. Our first objective has been to establish the packing structure of the nanocrystal superlattice Figure 2a shows a TEM image of an ultrathin (monolayer) film of 5.0 nm diameter NCAM: (6.4+0.2 nm center-to-center spacing), in which 2-D close-packing is deduced from the clear six fold symmetry. The electron diffraction pattern obtained from this region reveals the periodic structure of the nanocrystal lattices, and also provide an absolute calibration of the superlattice constant, because the fcc-Ag {111} ring is recorded simultaneously in the same photograph (figur 2b). Figures 2c and 2d show a TEM image and the corresponding diffraction patter corresponding to the [110] orientation, thus confirming that the superlattice has the fcc structure Such [110],-oriented films are the most frequently observed structure in our specimens, and are
very advantageous, as described previously [11], in that the projection from individual columns of nanocrystals can be viewed without obstruction by neighboring columns, allowing one to clearly view a projected shape of the nanocrystals, as we now describe. [The subscript s denotes the nanocrystal superlattice, to distinguish it from the reflections of atomic lattice of elemental Ag.] Our second objective has been to establish a relationship between this 3-D superlattice structure and the orientation of the NCAMs within it. Figures 3a andc are two TEM images recorded from the same specimen region under differing defocus conditions. Three striking features are observed. First, the superlattice contains "coherent" {200}, twin (T) and {111}, stacking fault (SF) structures. Close examination shows that the (projected) nanocrystal shape is modified across the twin boundary, accompanying the transformation in superlattice orientation. Second, profiles of the faceted structure of the NCAM cores can be seen in the image recorded at the condition close to in-focus (Figures 3a and b), and the co-aligned arrangement of these projected shapes is apparent. The majority of the nanocrystal profiles observed are consistent with the truncated octahedral (TO) morphology, which exposes {100} and {111} facets, as described previously [13]. Viewed along [110], two {100} facets and four {111} facets are projected edge-on (Figure 3b). This establishes the orientational relationship between the NCAM core's lattice and the superlattice as [110] II [110], and [002] lI [110],. Third, the image recorded at the out-focus condition exposes a web-like pattern of directional "contacts" among neighboring NCAMs, as indicated by dotted lines in Figure 3d, where the Ag cores are in dark contrast and the bright spots are channels enclosed by regions of enhanced density of surfactant chains. By comparison with Figure 3b, one can see that these dense regions interconnect between facets of Ag cores, rather than edges or vertices. It is therefore natural to suppose that the overlap of surfactant groups emanating from facets on neighboring nanocrystals serves as the prime contact which holds, or "bonds", and orients the particles in their 3-D superlattice. We further propose that materials of special type be called “highly oriented molecular (Ag) nanocrystal arrays", HOM(A)NA s. A precise orientational relationship between nanocrystal core lattices and the superlattice has Eee: PR RENN A fr Ee. Me A. er sn. Cae a a, ae” aes ee ee ae eS oe
very advantageous, as described previously [11], in that the projection from individual columns of nanocrystals can be viewed without obstruction by neighboring columns, allowing one to clearly view a projected shape of the nanocrystals, as we now describe. [The subscript s denotes the nanocrystal superlattice, to distinguish it from the reflections of atomic lattice of elemental Ag.] Our second objective has been to establish a relationship between this 3-D superlattice structure and the orientation of the NCAMs within it. Figures 3a andc are two TEM images recorded from the same specimen region under differing defocus conditions. Three striking features are observed. First, the superlattice contains "coherent" {200}, twin (T) and {111}, stacking fault (SF) structures. Close examination shows that the (projected) nanocrystal shape is modified across the twin boundary, accompanying the transformation in superlattice orientation. Second, profiles of the faceted structure of the NCAM cores can be seen in the image recorded at the condition close to in-focus (Figures 3a and b), and the co-aligned arrangement of these projected shapes is apparent. The majority of the nanocrystal profiles observed are consistent with the truncated octahedral (TO) morphology, which exposes {100} and {111} facets, as described previously [13]. Viewed along [110], two {100} facets and four {111} facets are projected edge-on (Figure 3b). This establishes the orientational relationship between the NCAM core's lattice and the superlattice as [110] II [110], and [002] lI [110],. Third, the image recorded at the out-focus condition exposes a web-like pattern of directional "contacts" among neighboring NCAMs, as indicated by dotted lines in Figure 3d, where the Ag cores are in dark contrast and the bright spots are channels enclosed by regions of enhanced density of surfactant chains. By comparison with Figure 3b, one can see that these dense regions interconnect between facets of Ag cores, rather than edges or vertices. It is therefore natural to suppose that the overlap of surfactant groups emanating from facets on neighboring nanocrystals serves as the prime contact which holds, or "bonds", and orients the particles in their 3-D superlattice. We further propose that materials of special type be called “highly oriented molecular (Ag) nanocrystal arrays", HOM(A)NA s. A precise orientational relationship between nanocrystal core lattices and the superlattice has Eee: PR RENN A fr Ee. Me A. er sn. Cae a a, ae” aes ee ee ae eS oe
Figure 3 [110], TEM images recorded from the a single region of a film under (a) in-focu: and (c) out-focus conditions, showing the faceted shape and the directional intermolecula bonds of the particles. The white lines indicate twin (T) planes, stacking faults (SF) and the rotation of the {111}, plane across the twin plane. An angle, a, is indicated which measure: the projected orientation between the {111} facet of a nanocrystal (dotted line) and the {111}. plane of the superlattice. A structural model for the formation of {100}, twins in the fcc superlattice is inset in (c) according to the structure model to be given in Figure 5b. (b) anc (d) are enlarged TEM images selected from (a) and (c), respectively, showing the relationshiy of the {111} and {100} facets of the truncated octahedron particles to the projected unit cel (white lines) of the superlattice. The dotted lines in (d) indicate the directional intermoleculai contacts formed by the groups of surfactant chains from facets on neighboring nanocrystals.
Figure 3 [110], TEM images recorded from the a single region of a film under (a) in-focu: and (c) out-focus conditions, showing the faceted shape and the directional intermolecula bonds of the particles. The white lines indicate twin (T) planes, stacking faults (SF) and the rotation of the {111}, plane across the twin plane. An angle, a, is indicated which measure: the projected orientation between the {111} facet of a nanocrystal (dotted line) and the {111}. plane of the superlattice. A structural model for the formation of {100}, twins in the fcc superlattice is inset in (c) according to the structure model to be given in Figure 5b. (b) anc (d) are enlarged TEM images selected from (a) and (c), respectively, showing the relationshiy of the {111} and {100} facets of the truncated octahedron particles to the projected unit cel (white lines) of the superlattice. The dotted lines in (d) indicate the directional intermoleculai contacts formed by the groups of surfactant chains from facets on neighboring nanocrystals.
Figure 5. (b) A schematic model of the fec packing of the TO Ag cores and following the orientational relations of [110] Il [110], and [002] II [110],. (a) The [110], projection of the mode shown in (b). The shaded polyhedra are out-of-plane by a distance (¥2/4)a and the angle ais indicated corresponding to that shown in Figure 3a. The intermolecular interactions are represente by coils that interconnect the {111}-facets with {111}-facets on neighboring nanocrysals and {100} with {100} facets.
Figure 5. (b) A schematic model of the fec packing of the TO Ag cores and following the orientational relations of [110] Il [110], and [002] II [110],. (a) The [110], projection of the mode shown in (b). The shaded polyhedra are out-of-plane by a distance (¥2/4)a and the angle ais indicated corresponding to that shown in Figure 3a. The intermolecular interactions are represente by coils that interconnect the {111}-facets with {111}-facets on neighboring nanocrysals and {100} with {100} facets.
Figure 3 shows the dependence of the cluster radius of gold on the time of heat treatment at 1000°C. Figure 4 presents the natural logarithm of all radii as a function of the reciprocal of the heat treatment temperature, all for equal heat treatment times. From the slopes, we find that the activation energies for metal cluster formation in silica are 80, 83 and 290 meV for gold, silver and copper, respectively.
Figure 3 shows the dependence of the cluster radius of gold on the time of heat treatment at 1000°C. Figure 4 presents the natural logarithm of all radii as a function of the reciprocal of the heat treatment temperature, all for equal heat treatment times. From the slopes, we find that the activation energies for metal cluster formation in silica are 80, 83 and 290 meV for gold, silver and copper, respectively.
Figure 4 The natural logarithm of metal cluster radii versus reciprocal heat treatment temperature. The lines determine the activation energies given in the text. Figure 3 Dependence of the metal cluster radius on time of heat treatment at 1100°C
Figure 4 The natural logarithm of metal cluster radii versus reciprocal heat treatment temperature. The lines determine the activation energies given in the text. Figure 3 Dependence of the metal cluster radius on time of heat treatment at 1100°C
Fig. 1 Outline of the modified VERL process used for the preparation of metal nano- suspensions. The vacuum chamber has a diameter of 700mm.The size of the rotating stainless stee] drum is @200mm x 400mm.
Fig. 1 Outline of the modified VERL process used for the preparation of metal nano- suspensions. The vacuum chamber has a diameter of 700mm.The size of the rotating stainless stee] drum is @200mm x 400mm.
Sputtering of Ag has been performed at a constant input power of 1kW and at Ar gas pressures ranging from | to 30 Pa. The mean particle sizes have been determined from fits of the respective size histograms to log-normal size distributions. Fig.2 gives an idea of the quality of these fits. The results are plotted as a function of Ar gas pressure in fig. 3. Fig.2 TEM bright field image (left) and respective particle size histogram (right) of Ag- nanoparticles in mineral oil prepared at a sputtering pressure of 5 Pa and a sputtering input power of 1 kW. The size histogram was fitted to a log-normal distribution function (equ.(1)) using a mean particle size of <d> =9.5 nm and a standard deviation of o =1.28.
Sputtering of Ag has been performed at a constant input power of 1kW and at Ar gas pressures ranging from | to 30 Pa. The mean particle sizes have been determined from fits of the respective size histograms to log-normal size distributions. Fig.2 gives an idea of the quality of these fits. The results are plotted as a function of Ar gas pressure in fig. 3. Fig.2 TEM bright field image (left) and respective particle size histogram (right) of Ag- nanoparticles in mineral oil prepared at a sputtering pressure of 5 Pa and a sputtering input power of 1 kW. The size histogram was fitted to a log-normal distribution function (equ.(1)) using a mean particle size of <d> =9.5 nm and a standard deviation of o =1.28.
Fig. 3: Mean size of Ag particles in the prepared suspensions as a function of sputtering gas pressure. The data have been derived from fits of the respective size histograms (TEM) to log-normal distribution functions. The solid line fits the data with a slope of 0.40. In contrast to thermal evaporation in the case of sputtering a second concurrent process may affect the resulting particle sizes: Keeping both sputtering current and voltage constant the sputtering rate strongly decreases with increasing gas pressure due to a reduction of the average kinetic energy of the impinging Ar atoms [3]. Here this effect shows up in a reduction of metal content of the liquid suspension with increasing gas pressure (fig. 4).
Fig. 3: Mean size of Ag particles in the prepared suspensions as a function of sputtering gas pressure. The data have been derived from fits of the respective size histograms (TEM) to log-normal distribution functions. The solid line fits the data with a slope of 0.40. In contrast to thermal evaporation in the case of sputtering a second concurrent process may affect the resulting particle sizes: Keeping both sputtering current and voltage constant the sputtering rate strongly decreases with increasing gas pressure due to a reduction of the average kinetic energy of the impinging Ar atoms [3]. Here this effect shows up in a reduction of metal content of the liquid suspension with increasing gas pressure (fig. 4).
Fig. 3 Transient reflectivity changes in the three samples as a function of the probe time delay. Dashed lines: fitting curves. of the size dependence of the ullrarast Wansient ronecuvity jr}. These measurements were performed in a pump-probe configuration, using a Ti:sapphire laser with chirped pulse amplification, which provides pulses of 150-fs duration at 780 nm with energy up to 750 «J at 1-kHz repetition rate. In the experiment, the pump and probe wavelengths were 390 and 780 nm, respectively. The pump beam was obtained by frequency doubling a fraction of the laser beam in a LiB3Os; crystal. The pump pulse duration was 180 fs and the energy used in the measurements was in the range 6-60 nJ. Degenerate transient reflectivity measurements were performed on a subpicosecond time scale by using ultrashort pulses (10 fs) at 780 nm, with energy up to 240 pJ. Pump and probe beam polarizations were made orthogonal to avoid thermal grating buildup.
Fig. 3 Transient reflectivity changes in the three samples as a function of the probe time delay. Dashed lines: fitting curves. of the size dependence of the ullrarast Wansient ronecuvity jr}. These measurements were performed in a pump-probe configuration, using a Ti:sapphire laser with chirped pulse amplification, which provides pulses of 150-fs duration at 780 nm with energy up to 750 «J at 1-kHz repetition rate. In the experiment, the pump and probe wavelengths were 390 and 780 nm, respectively. The pump beam was obtained by frequency doubling a fraction of the laser beam in a LiB3Os; crystal. The pump pulse duration was 180 fs and the energy used in the measurements was in the range 6-60 nJ. Degenerate transient reflectivity measurements were performed on a subpicosecond time scale by using ultrashort pulses (10 fs) at 780 nm, with energy up to 240 pJ. Pump and probe beam polarizations were made orthogonal to avoid thermal grating buildup.
The transient reflectivity changes, AR, in the samples are shown in fig. 3 as a function of the probe delay. The time decay can be separated in three distinct time scales, each one characterized by a time constant t resulting from a fit of the curves using three exponential functions. The values of these time constants 1), t2, 13, given in Table 1, are largely different, being related to three different relaxation phenomena. Table 1 Time constants obtained from the fitting of the transient reflectivity measurements These time constants can be associated to the processes of hot-electron thermalization by electron-electron (e-e), electron-phonon (e-p) and electron-surface (e-s) collisions [4],[5]. The electron thermalization usually occurs on a subpicosecond timescale, with the formation of a Fermi distribution with T. > To (where Tp is the initial sample temperature). The hot electron system thermalizes with the lattice by e-e, but mainly by e-p and e-s collisions on a time scale of a few picoseconds [6]. The excess energy is then dispersed toward the matrix, in tens of picoseconds [7],[8], and the whole system decays to the original temperature in a time scale of hundreds of ps.
The transient reflectivity changes, AR, in the samples are shown in fig. 3 as a function of the probe delay. The time decay can be separated in three distinct time scales, each one characterized by a time constant t resulting from a fit of the curves using three exponential functions. The values of these time constants 1), t2, 13, given in Table 1, are largely different, being related to three different relaxation phenomena. Table 1 Time constants obtained from the fitting of the transient reflectivity measurements These time constants can be associated to the processes of hot-electron thermalization by electron-electron (e-e), electron-phonon (e-p) and electron-surface (e-s) collisions [4],[5]. The electron thermalization usually occurs on a subpicosecond timescale, with the formation of a Fermi distribution with T. > To (where Tp is the initial sample temperature). The hot electron system thermalizes with the lattice by e-e, but mainly by e-p and e-s collisions on a time scale of a few picoseconds [6]. The excess energy is then dispersed toward the matrix, in tens of picoseconds [7],[8], and the whole system decays to the original temperature in a time scale of hundreds of ps.
Figure 3. shows the L, edge for Pt on porous silicon compared with that of the bulk metal. An intense white line is clearly evident where it was absent in the bulk metal. This spectroscopic change is present in all the metal on porous silicon samples examined. Figure 4. shows the SEM of a gold cluster on porous silicon. The cluster is an aggregate of yet smaller clusters. At several hundred nanometres in diameter, Au clusters are much larger (approx. one order of magnitude), than the clusters of other metals. This is consistent with the X-ray absorption spectra where the difference in intensity between bulk Au and Au on porous silicon is very emall.
Figure 3. shows the L, edge for Pt on porous silicon compared with that of the bulk metal. An intense white line is clearly evident where it was absent in the bulk metal. This spectroscopic change is present in all the metal on porous silicon samples examined. Figure 4. shows the SEM of a gold cluster on porous silicon. The cluster is an aggregate of yet smaller clusters. At several hundred nanometres in diameter, Au clusters are much larger (approx. one order of magnitude), than the clusters of other metals. This is consistent with the X-ray absorption spectra where the difference in intensity between bulk Au and Au on porous silicon is very emall.
Fig. 1. Schematic diagram of the surface melting process.
Fig. 1. Schematic diagram of the surface melting process.
Surface Melting and Associated Diffusivity 3. RESULTS AND DISCUSSION
Surface Melting and Associated Diffusivity 3. RESULTS AND DISCUSSION
Fig. 2. Predicted surface melting and melting temperatures of alumina nanoparticles. Also shown (dashed) is the bulk melting temperature. replaced by the surface melting temperature T,,,;. If we further postulate that Q, is explicitly size independent, then we have the surface self-diffusion coefficient for small particles: a. estimated from the method proposed in the references. b. taken as the lattice spacing of o-alumina.
Fig. 2. Predicted surface melting and melting temperatures of alumina nanoparticles. Also shown (dashed) is the bulk melting temperature. replaced by the surface melting temperature T,,,;. If we further postulate that Q, is explicitly size independent, then we have the surface self-diffusion coefficient for small particles: a. estimated from the method proposed in the references. b. taken as the lattice spacing of o-alumina.
Fig. 4. Coagulation and sintering characteristic times used to determine Al,03 spherule sizes. Beyond the crossover particle size, ts becomes much longer than tc and particle growth is expected to stop at /near the indicated crossover size. Fig. 3. Predicted size-dependent surface self- diffusivity for AlpO3 particles, showing the enhancement for small particles. Also shown is the macroscopic diffusivity. The corresponding surface self-diffusivity is plotted in Fig. 3, from which we see that the enhancement of this property is especially impressive for small particles. Of course, for a sintering process via surface diffusion we are only concerned with the surface structure, not the state inside each nanoparticle.
Fig. 4. Coagulation and sintering characteristic times used to determine Al,03 spherule sizes. Beyond the crossover particle size, ts becomes much longer than tc and particle growth is expected to stop at /near the indicated crossover size. Fig. 3. Predicted size-dependent surface self- diffusivity for AlpO3 particles, showing the enhancement for small particles. Also shown is the macroscopic diffusivity. The corresponding surface self-diffusivity is plotted in Fig. 3, from which we see that the enhancement of this property is especially impressive for small particles. Of course, for a sintering process via surface diffusion we are only concerned with the surface structure, not the state inside each nanoparticle.
Figure 1. X-ray diffraction patterns of (a) a low oxygen concentration Mo,N film and (b) an MoO, contaminated Mo,N film. The peak broadening is due to small grain size. MESGY Ch Bw LEAGNE CASED CEA The nitridation of the Ti substrates reduced the MoO, present in the converted Mo,N films. The nitrided titanium layer made up only 5-10% of the bulk, but it was concentrated at the surface of the substrate. Due to their excellent oxidation resistance, the Ti nitrided substrates did not oxidize to the extent of the Ti substrates during the preparation of MoO, films. It was discovered that once titanium oxides (TiO, and TiO) were formed on the substrate, it was thermodynamically impossible at the reaction temperatures and pressures employed to convert these oxides to Ti,N or TiN. As the Gibbs free energy of formation of MoO, possesses a more negative value than TiO at temperatures below 1020 K, the presence of TiO was found to inhibit the conversion of MoO, to
Figure 1. X-ray diffraction patterns of (a) a low oxygen concentration Mo,N film and (b) an MoO, contaminated Mo,N film. The peak broadening is due to small grain size. MESGY Ch Bw LEAGNE CASED CEA The nitridation of the Ti substrates reduced the MoO, present in the converted Mo,N films. The nitrided titanium layer made up only 5-10% of the bulk, but it was concentrated at the surface of the substrate. Due to their excellent oxidation resistance, the Ti nitrided substrates did not oxidize to the extent of the Ti substrates during the preparation of MoO, films. It was discovered that once titanium oxides (TiO, and TiO) were formed on the substrate, it was thermodynamically impossible at the reaction temperatures and pressures employed to convert these oxides to Ti,N or TiN. As the Gibbs free energy of formation of MoO, possesses a more negative value than TiO at temperatures below 1020 K, the presence of TiO was found to inhibit the conversion of MoO, to
Figure 2. Scanning electron micrographs (SEM) of (a) a bulk oxide free Mo,N film and (b) an MoO, contaminated Mo,N film. The particle size variation in (b) is believed to be due to the formation of MoO, in place of Mo,N. Rags ie: LA Bes ee ae Ie en eo ea Cn een ne Se To monitor contamination in the Mo,N films, AES and EDX was conducted to determine lhe presence of oxygen and chlorine, respectively. The AES depth profile shown in Figure 4a of an Mo,N film on a nitrided Ti substrate indicates that oxygen levels were near background soncentrations adjacent to film/substrate interface. Conversely, in Figure 4b, oxygen is indicated near the interface region in an Mo,N film on Ti substrates prepared by using the same temperature programmed reaction. The EDX data (not shown) of these two films did not indicate the presence of chlorine. Therefore at the MoO, deposition temperatures employed, complete pyrolysis of MoCl. was achieved.
Figure 2. Scanning electron micrographs (SEM) of (a) a bulk oxide free Mo,N film and (b) an MoO, contaminated Mo,N film. The particle size variation in (b) is believed to be due to the formation of MoO, in place of Mo,N. Rags ie: LA Bes ee ae Ie en eo ea Cn een ne Se To monitor contamination in the Mo,N films, AES and EDX was conducted to determine lhe presence of oxygen and chlorine, respectively. The AES depth profile shown in Figure 4a of an Mo,N film on a nitrided Ti substrate indicates that oxygen levels were near background soncentrations adjacent to film/substrate interface. Conversely, in Figure 4b, oxygen is indicated near the interface region in an Mo,N film on Ti substrates prepared by using the same temperature programmed reaction. The EDX data (not shown) of these two films did not indicate the presence of chlorine. Therefore at the MoO, deposition temperatures employed, complete pyrolysis of MoCl. was achieved.
Figure 2 (a) Weak-beam dark-field image (g=220) of the film; the particles at A and B show dark contrast on a bright background. (b) Weak-beam dark-field image of the same region of the film with illumination tilted relative to (a). The particles at A and B now show a bright contrast on a dark background, indicating a local thickness increase and confirming their epitaxial orientation. At this point the particles are almost invisible in bright-field.
Figure 2 (a) Weak-beam dark-field image (g=220) of the film; the particles at A and B show dark contrast on a bright background. (b) Weak-beam dark-field image of the same region of the film with illumination tilted relative to (a). The particles at A and B now show a bright contrast on a dark background, indicating a local thickness increase and confirming their epitaxial orientation. At this point the particles are almost invisible in bright-field.
Figure 4 Cross-sectional illustration of the reorientation process occurring upon annealing an as- deposited particle. (a) Particle prior to annealing, showing the presence of a grain boundary close to the particle: bulk interface, (b) fluctuation of the position of the grain boundary upon heating and (c) epitaxial reorientation of the particle. Shaded lines represent copper (100) planes.
Figure 4 Cross-sectional illustration of the reorientation process occurring upon annealing an as- deposited particle. (a) Particle prior to annealing, showing the presence of a grain boundary close to the particle: bulk interface, (b) fluctuation of the position of the grain boundary upon heating and (c) epitaxial reorientation of the particle. Shaded lines represent copper (100) planes.
The compute-intensive part of MD simulations is the calculation of interparticle interactions. Highly efficient algorithms have been designed to compute these interactions on parallel machines [11]. The long-range Coulomb interaction is calculated with a divide-and-conquer scheme, called the fast multipole method (FMM), which reduces the computational complexity from O(N2) to O(N). For short-ranged interactions, we have employed a multiple time-step (MTS) approach in which a significant reduction in computation is achieved by exploiting different time scales for different force components.
The compute-intensive part of MD simulations is the calculation of interparticle interactions. Highly efficient algorithms have been designed to compute these interactions on parallel machines [11]. The long-range Coulomb interaction is calculated with a divide-and-conquer scheme, called the fast multipole method (FMM), which reduces the computational complexity from O(N2) to O(N). For short-ranged interactions, we have employed a multiple time-step (MTS) approach in which a significant reduction in computation is achieved by exploiting different time scales for different force components.
Simulation of nanophase materials is characterized by irregular atomic distribution. Uniform spatial decomposition on parallel computers would result in unequal partition of workloads among processors and low efficiency. Recently, we have added a dynamic-load- balancing capability to an existing parallel MD code which was based on uniform mesh decomposition [13]. The new load-balancing scheme uses adaptive curvilinear coordinates to represent partition boundaries. Workloads are partitioned with a uniform 3-dimensional mesh in the curvilinear coordinate system (Fig. 2). Simulated annealing is used to determine the optimal coordinate system which minimizes load-imbalance and communication costs. Periodic boundary conditions are naturally incorporated, and the new scheme requires no change in the data structures of the MD program. The computational overhead to perform load balancing is only 3.7% of the total MD execution time for a 633,696-atom SiO, system on 32 nodes of the SP machine. Inclusion of load balancing reduces the execution time by a factor of 4.2. FIG. 2: Partition of 2,964 atoms into 2 x 2 x 1 processors: (left) curved partition boundaries in the Euclidean space; (right) uniform mesh partition in the curvilinear space.
Simulation of nanophase materials is characterized by irregular atomic distribution. Uniform spatial decomposition on parallel computers would result in unequal partition of workloads among processors and low efficiency. Recently, we have added a dynamic-load- balancing capability to an existing parallel MD code which was based on uniform mesh decomposition [13]. The new load-balancing scheme uses adaptive curvilinear coordinates to represent partition boundaries. Workloads are partitioned with a uniform 3-dimensional mesh in the curvilinear coordinate system (Fig. 2). Simulated annealing is used to determine the optimal coordinate system which minimizes load-imbalance and communication costs. Periodic boundary conditions are naturally incorporated, and the new scheme requires no change in the data structures of the MD program. The computational overhead to perform load balancing is only 3.7% of the total MD execution time for a 633,696-atom SiO, system on 32 nodes of the SP machine. Inclusion of load balancing reduces the execution time by a factor of 4.2. FIG. 2: Partition of 2,964 atoms into 2 x 2 x 1 processors: (left) curved partition boundaries in the Euclidean space; (right) uniform mesh partition in the curvilinear space.
At this strain we observed the growth of a few crack branches in addition to sligh progression of the notch in the y direction. These local branches and nanoclusters decelerate the crack front. When the motion of the crack front subsided, we increased the strain along x by displacing the same boundary-layer atoms by 1% and then we relaxed the system for 10 ps. The crack front then advanced significantly in the y direction. We also observed that interfacia regions allowed the crack front to meander and cause local crack branches to sprout in the system. At a strain of 14%, a secondary crack with several local branches merged with th primary crack. With further increase in the strain the secondary front advanced toward the initia notch and finally when the strain reaches 30% the two fronts joined and the system wa completely fractured. Figure 3 shows a snapshot of the nanophase system just before it fractures It should be noted that the critical strain (30%) at which the nanophase system fractures i enormous compared to what the crystalline Si3N4 system can sustain (at an applied strain of only 3% the crystal undergoes cleavage fracture).
At this strain we observed the growth of a few crack branches in addition to sligh progression of the notch in the y direction. These local branches and nanoclusters decelerate the crack front. When the motion of the crack front subsided, we increased the strain along x by displacing the same boundary-layer atoms by 1% and then we relaxed the system for 10 ps. The crack front then advanced significantly in the y direction. We also observed that interfacia regions allowed the crack front to meander and cause local crack branches to sprout in the system. At a strain of 14%, a secondary crack with several local branches merged with th primary crack. With further increase in the strain the secondary front advanced toward the initia notch and finally when the strain reaches 30% the two fronts joined and the system wa completely fractured. Figure 3 shows a snapshot of the nanophase system just before it fractures It should be noted that the critical strain (30%) at which the nanophase system fractures i enormous compared to what the crystalline Si3N4 system can sustain (at an applied strain of only 3% the crystal undergoes cleavage fracture).
To investigate the nature of self-affine fracture surfaces in the nanophase Si3N 4, we calculate the height-height correlation functions both in and out of the fracture plane y-z (see Fig. 4). Figure 5a shows the best fit to the out-of-plane height-height correlation function gxx(z) (= <[x(z+z,) - x(Z,)17>"”) for the fracture profile x(z) requires two roughness exponents: C, = 0.84+0.12 above a certain length scale (64A) and C, = 0.58+0.14 otherwise. The inset shows the MD results for the other out-of-plane height-height correlation function gxx(y). In this case the best fit to the results gives a roughness exponent G, = 0.7540.08. The MD results for C, and are very close to experimental values [14]. We have also determined the in-plane roughness exponent, €. Figure Sb shows that the best fit to the corresponding height-height correlation function gives C = 0.57+0.08, which is in good agreement with experiments [14]. unction gives C = 0.57+0.08, which is in good agreement with experiments [14].
To investigate the nature of self-affine fracture surfaces in the nanophase Si3N 4, we calculate the height-height correlation functions both in and out of the fracture plane y-z (see Fig. 4). Figure 5a shows the best fit to the out-of-plane height-height correlation function gxx(z) (= <[x(z+z,) - x(Z,)17>"”) for the fracture profile x(z) requires two roughness exponents: C, = 0.84+0.12 above a certain length scale (64A) and C, = 0.58+0.14 otherwise. The inset shows the MD results for the other out-of-plane height-height correlation function gxx(y). In this case the best fit to the results gives a roughness exponent G, = 0.7540.08. The MD results for C, and are very close to experimental values [14]. We have also determined the in-plane roughness exponent, €. Figure Sb shows that the best fit to the corresponding height-height correlation function gives C = 0.57+0.08, which is in good agreement with experiments [14]. unction gives C = 0.57+0.08, which is in good agreement with experiments [14].
Figure 2 shows the stress-strain curve for Ni_15 and Ni_50. At stresses lower then 80 MPa the deformation can be said to be elastic and the curve fits to a straight line within a standard deviation of 1%. On the plot only the linear fit for ni_15 is shown, because of the very little difference. Unloading the sample leads to the original sample size. From linear fitting of the Strain versus stress in longitudinal and transverse direction, the elastic constants C); and C2 can be calculated. For Ni_15 the values are 170 and 90 Gpa and for Ni_50, 150 and 60 GPa respectively. This means a reduction of about 30 % in Young modulus for a sample with mean grain size of 5.4 nm and a reduction of 40 % for a sample with a mean grain size of 3.2 nm relative to the value of a single crystal calculated with this potential (Cj, :250 Gpa, Cy. : 155 Gpa). The reduction in E-modulus is less then the measured values in consolidated samples and is in the range of the value (50%) reported in other computer simulations on Fe with mean grain
Figure 2 shows the stress-strain curve for Ni_15 and Ni_50. At stresses lower then 80 MPa the deformation can be said to be elastic and the curve fits to a straight line within a standard deviation of 1%. On the plot only the linear fit for ni_15 is shown, because of the very little difference. Unloading the sample leads to the original sample size. From linear fitting of the Strain versus stress in longitudinal and transverse direction, the elastic constants C); and C2 can be calculated. For Ni_15 the values are 170 and 90 Gpa and for Ni_50, 150 and 60 GPa respectively. This means a reduction of about 30 % in Young modulus for a sample with mean grain size of 5.4 nm and a reduction of 40 % for a sample with a mean grain size of 3.2 nm relative to the value of a single crystal calculated with this potential (Cj, :250 Gpa, Cy. : 155 Gpa). The reduction in E-modulus is less then the measured values in consolidated samples and is in the range of the value (50%) reported in other computer simulations on Fe with mean grain
Figure 1a shows histograms of potential energy for the two extreme sizes, Ni_15 and Ni_50, it the undeformed state at 0 K. Bin width is 40 meV, which is approximately the valu corresponding to potential thermal energy at room temperature. In this scale, it is clearly seen th relevance of the interfaces in terms of the number of atoms that have potential energy above th perfect crystal value, within a range of thermal energies: these numbers are 50% for Ni_15, anc 70% for Ni_50, with 5.4 and 3.2 nm respectively; if we count the atoms with energy larger tha the latent heat of melting, these numbers reduce to 12% and 18% respectively. This figure no only tells us about the significant perturbation to the crystalline structure that these grail boundaries introduce, but also shows that the average potential energy those defected atom store is comparable to the latent heat of melting of this Ni model potential, 180 meV/atom. Thi: gives strong support to the interpretation of boundaries in terms of amorphous (or liquid-like structures. Also shown in this figure is the grain size distributions for Ni_SO and Ni_15.
Figure 1a shows histograms of potential energy for the two extreme sizes, Ni_15 and Ni_50, it the undeformed state at 0 K. Bin width is 40 meV, which is approximately the valu corresponding to potential thermal energy at room temperature. In this scale, it is clearly seen th relevance of the interfaces in terms of the number of atoms that have potential energy above th perfect crystal value, within a range of thermal energies: these numbers are 50% for Ni_15, anc 70% for Ni_50, with 5.4 and 3.2 nm respectively; if we count the atoms with energy larger tha the latent heat of melting, these numbers reduce to 12% and 18% respectively. This figure no only tells us about the significant perturbation to the crystalline structure that these grail boundaries introduce, but also shows that the average potential energy those defected atom store is comparable to the latent heat of melting of this Ni model potential, 180 meV/atom. Thi: gives strong support to the interpretation of boundaries in terms of amorphous (or liquid-like structures. Also shown in this figure is the grain size distributions for Ni_SO and Ni_15.
NR An evaluation of the energy distribution after 10% deformation (fig.2b) followed by a short unloading, reveals that deformation takes place increasing slightly the total number of defected atoms; 13% more atoms join the grain boundary region, but less then 1% of them end in an amorphous-like position. The strain field induced by the deformation extends from the boundary Figure 3 shows strain versus time for all samples (Ni_15.. Ni_50O). It becomes apparent that they show a linear behavior, with increasing strain rate for smaller grain size. It is worth noticing that the density of all samples is reduced to about 98% of the initial value during deformation, and that relaxation after ultimate loading does recover density to 99% of that value, so one can say that deformation takes place at constant volume (i. e. no pore formation).
NR An evaluation of the energy distribution after 10% deformation (fig.2b) followed by a short unloading, reveals that deformation takes place increasing slightly the total number of defected atoms; 13% more atoms join the grain boundary region, but less then 1% of them end in an amorphous-like position. The strain field induced by the deformation extends from the boundary Figure 3 shows strain versus time for all samples (Ni_15.. Ni_50O). It becomes apparent that they show a linear behavior, with increasing strain rate for smaller grain size. It is worth noticing that the density of all samples is reduced to about 98% of the initial value during deformation, and that relaxation after ultimate loading does recover density to 99% of that value, so one can say that deformation takes place at constant volume (i. e. no pore formation).
size of Inm [12]. The reduction is however larger then the predicted values from the three- compound model described in reference[6]. At higher loads the curve start to deviate from linear behaviour. The inset in fig.2 shows the deviation from linear behaviour expressed in %. The exact onset of plastic deformation is difficult to determine because of the very slowly or even maybe never reached maximum deformation. In order to reduce error, values given are all strain values loading during 30 psec, when strain rate was very slow. As can be observed, the strains reached in Ni_5O are higher then those reached in Ni_15, which means that there is more plastic deformation at the lower grain size. This indicates a lowering in yield stress or a reduced hardness when grain size is reduced from 5.4 to 3.2 nm. At 5.4 nm the yield stress is between 0.1 and 0.2 Gpa, at 3.2 nm the yield stress is below 0.1 Gpa.
size of Inm [12]. The reduction is however larger then the predicted values from the three- compound model described in reference[6]. At higher loads the curve start to deviate from linear behaviour. The inset in fig.2 shows the deviation from linear behaviour expressed in %. The exact onset of plastic deformation is difficult to determine because of the very slowly or even maybe never reached maximum deformation. In order to reduce error, values given are all strain values loading during 30 psec, when strain rate was very slow. As can be observed, the strains reached in Ni_5O are higher then those reached in Ni_15, which means that there is more plastic deformation at the lower grain size. This indicates a lowering in yield stress or a reduced hardness when grain size is reduced from 5.4 to 3.2 nm. At 5.4 nm the yield stress is between 0.1 and 0.2 Gpa, at 3.2 nm the yield stress is below 0.1 Gpa.
Fig. 1 Flow chart of computer programm simulating microstructure development
Fig. 1 Flow chart of computer programm simulating microstructure development
An analysis of Equation 4 indicates that increase in overpotential results in a decrease in activation energy. Hence, the nucleation probability increases. During simulation, nucleation takes place at the deposition stage. A block at this stage has the possibility of becoming a nucleus or ad-block. Increasing the nucleation probability results in a higher chance for a block to become a nucleus with a new orientation rather than remain an ad-block. Therefore, the resultant grain growth process is depressed. Thus, for relatively high values of the overpotential, the grain growth along the deposition direction can be effectively restricted and smaller equi-axis grains in the microstructure should be expected. The input parameters used in the simulation are specified in Table II. The microstructure obtained from simulation for five values of overpotential, 400 500, 600, 700 and 800 mV, are shown in Fig. Sa-e. The grain size distribution as a function of grain radius, which was derived from Fig.5e, is shown in Fig. 5f. It is seen, that the grain size decreases rapidly as the overpotential increases. During electrodeposition at lower overpotentials, columnar grains are formed (Fig. 5a). Conversely, at higher overpotentials, smaller and equi-axis grains are formed. Both the overpotential and current density were found to contribute to the microstructure development.
An analysis of Equation 4 indicates that increase in overpotential results in a decrease in activation energy. Hence, the nucleation probability increases. During simulation, nucleation takes place at the deposition stage. A block at this stage has the possibility of becoming a nucleus or ad-block. Increasing the nucleation probability results in a higher chance for a block to become a nucleus with a new orientation rather than remain an ad-block. Therefore, the resultant grain growth process is depressed. Thus, for relatively high values of the overpotential, the grain growth along the deposition direction can be effectively restricted and smaller equi-axis grains in the microstructure should be expected. The input parameters used in the simulation are specified in Table II. The microstructure obtained from simulation for five values of overpotential, 400 500, 600, 700 and 800 mV, are shown in Fig. Sa-e. The grain size distribution as a function of grain radius, which was derived from Fig.5e, is shown in Fig. 5f. It is seen, that the grain size decreases rapidly as the overpotential increases. During electrodeposition at lower overpotentials, columnar grains are formed (Fig. 5a). Conversely, at higher overpotentials, smaller and equi-axis grains are formed. Both the overpotential and current density were found to contribute to the microstructure development.
Fig. 5 Simulated evolution of microstructure for different overpotentials. (a) n=400mv; (b) n=500mv; (c) n=600mv; (d) n=700mv; (e) y=800mv; and (f) the grain size distribution for microstructure shown in Fig. 5 (e)
Fig. 5 Simulated evolution of microstructure for different overpotentials. (a) n=400mv; (b) n=500mv; (c) n=600mv; (d) n=700mv; (e) y=800mv; and (f) the grain size distribution for microstructure shown in Fig. 5 (e)
Fig. 1 Nanophase Si3Nq with 1.08 million atoms (a) before consolidation (1.0 g/cc) and (b) after consolidation (2.94g/cc) .
Fig. 1 Nanophase Si3Nq with 1.08 million atoms (a) before consolidation (1.0 g/cc) and (b) after consolidation (2.94g/cc) .
Fig. 2 (a) Si-N pair-distribution functions (PDF); (b) N-Si-N and (c) Si-N-Si bond-angle distributions (BAD). (Dashed and solid curves correspond to the interior regions of nanoclusters and the intercluster regions of the nanophase system, respectively.) The inset in (a) shows the first peaks in the PDF. Fig. 3. Spatial distribution of the average Si coordination, projected onto the x-y plane, in the nanophase system at 2.94 g/cc. Mw ee ee ee ee SRS We have calculated various correlation functions to determine the structure of the consolidated nanophase Si,N, at mass density 2.94 g/cc. In Fig. 2 (a), we depict partial (Si-N) pair-distribution functions (PDF) for atoms inside the nanoclusters (dashed curve) and in intercluster regions (solid curve) at 5 K. The sharp peaks in the PDF for atoms inside the clusters indicate that the interior regions of these clusters remain crystalline. On the other hand, the PDF for intercluster regions has much broader peaks (except the first one) than those inside the nanoclusters. The inset in Fig. 2(a) shows that the height of the first peak in the PDF for intercluster regions is only one fourth of the height of the first peak for interior regions. It should also be noted that the position of the first peak for intercluser regions is shifted to a lower value relative to that for the interior region of nanoclusters. This shift implies a decrease of the nearest- neighbor coordination for Si atoms in intercluster regions. Lower Si coordination in intercluster regions is also evident in Fig. 3. The figure shows that the average Si coordination in intercluster regions is approximately 3.5 which means that there are 50% three-fold coordinated Si atoms while the remaining Si atoms are four-fold coordinated. (In crystalline Si3N4 the coordination of Si is four [12].) The figure also shows that there are a few pores (black regions) inside the system. We have also examined N-Si-N and Si-N-Si bond-angle distributions (BAD) inside the nanoclusters and in intercluster regions. We find that the BADs for intercluster regions are much broader than those corresponding to the interior region of nanoclusters. This also indicates that the local structure in the intercluster regions is highly distorted [13].
Fig. 2 (a) Si-N pair-distribution functions (PDF); (b) N-Si-N and (c) Si-N-Si bond-angle distributions (BAD). (Dashed and solid curves correspond to the interior regions of nanoclusters and the intercluster regions of the nanophase system, respectively.) The inset in (a) shows the first peaks in the PDF. Fig. 3. Spatial distribution of the average Si coordination, projected onto the x-y plane, in the nanophase system at 2.94 g/cc. Mw ee ee ee ee SRS We have calculated various correlation functions to determine the structure of the consolidated nanophase Si,N, at mass density 2.94 g/cc. In Fig. 2 (a), we depict partial (Si-N) pair-distribution functions (PDF) for atoms inside the nanoclusters (dashed curve) and in intercluster regions (solid curve) at 5 K. The sharp peaks in the PDF for atoms inside the clusters indicate that the interior regions of these clusters remain crystalline. On the other hand, the PDF for intercluster regions has much broader peaks (except the first one) than those inside the nanoclusters. The inset in Fig. 2(a) shows that the height of the first peak in the PDF for intercluster regions is only one fourth of the height of the first peak for interior regions. It should also be noted that the position of the first peak for intercluser regions is shifted to a lower value relative to that for the interior region of nanoclusters. This shift implies a decrease of the nearest- neighbor coordination for Si atoms in intercluster regions. Lower Si coordination in intercluster regions is also evident in Fig. 3. The figure shows that the average Si coordination in intercluster regions is approximately 3.5 which means that there are 50% three-fold coordinated Si atoms while the remaining Si atoms are four-fold coordinated. (In crystalline Si3N4 the coordination of Si is four [12].) The figure also shows that there are a few pores (black regions) inside the system. We have also examined N-Si-N and Si-N-Si bond-angle distributions (BAD) inside the nanoclusters and in intercluster regions. We find that the BADs for intercluster regions are much broader than those corresponding to the interior region of nanoclusters. This also indicates that the local structure in the intercluster regions is highly distorted [13].
Dhaest VUAUELE Iw BY LEE CUA VUE VE NA av Pee bat oY BARN WEE WEILL ANE LER PO fe CO a phase c, is calculated from the effective volume of the crystalline part of nanoclusters; and the concentration of amorphous intercluster regions c3 is determined from the condition, c+ c2+¢3= 1. The bulk and shear moduli of individual phases are obtained from MD calculations for the o- crystal and the amorphous Si;N, system [8]. (K; = G, = 0; Ky = 289 GPa and G, = 145 GPa for the o-crystal; K; = 181 GPa and G; = 109 GPa for the amorphous system.) Using those values we solved Eqs. (1) and (2) for K and G. In Fig. 5 we show these results by solid circles (K), triangles (K) and squares (G). Evidently the three-phase model for the elastic moduli explains successfully the MD results for the dependence of elastic moduli on both porosity and cluster size. The elastic moduli of amorphous $i3N4q are smaller than the crystal and a nanophase system with 45 A clusters has larger concentration of amorphous regions than the system with 60 A clusters. The system with smaller grains is therefore less stiff than the system with larger grains.
Dhaest VUAUELE Iw BY LEE CUA VUE VE NA av Pee bat oY BARN WEE WEILL ANE LER PO fe CO a phase c, is calculated from the effective volume of the crystalline part of nanoclusters; and the concentration of amorphous intercluster regions c3 is determined from the condition, c+ c2+¢3= 1. The bulk and shear moduli of individual phases are obtained from MD calculations for the o- crystal and the amorphous Si;N, system [8]. (K; = G, = 0; Ky = 289 GPa and G, = 145 GPa for the o-crystal; K; = 181 GPa and G; = 109 GPa for the amorphous system.) Using those values we solved Eqs. (1) and (2) for K and G. In Fig. 5 we show these results by solid circles (K), triangles (K) and squares (G). Evidently the three-phase model for the elastic moduli explains successfully the MD results for the dependence of elastic moduli on both porosity and cluster size. The elastic moduli of amorphous $i3N4q are smaller than the crystal and a nanophase system with 45 A clusters has larger concentration of amorphous regions than the system with 60 A clusters. The system with smaller grains is therefore less stiff than the system with larger grains.
Fig. 5 Snapshots of fractured Si3N4: (a) The consolidated nanophase Si3N4 under an applied strain of 5%; (b) the nanophase system under an applied strain of 14% strain; and (c) the crystalline SizN4 under an applied strain of 4%.
Fig. 5 Snapshots of fractured Si3N4: (a) The consolidated nanophase Si3N4 under an applied strain of 5%; (b) the nanophase system under an applied strain of 14% strain; and (c) the crystalline SizN4 under an applied strain of 4%.
Central to the development of the nanocomposite systems examined in this study are the thermal properties of the ferromagnetic metal nitrides utilized as precursors. Their relatively limited stability allows selective decomposition at moderate temperatures, ~ 600 °C, while the ceramic matrix is unaffected. The thermal decomposition of pure CoN and FeN was examined by Suzuki et al [7,8] who conducted annealing treatments in a stepwise manner; on the other hand, the thermal decomposition of pure Ni;N was examined[6] using a continuous incremental ramp. Pure Co and Fe foils were sputtered in the present study to confirm the stoichiometry of the products formed under the experimental conditions described above. The nitrides produced in both cases were the ones with a 1:1 stoichiometry. Both elements form nitrides with lower nitrogen content, Co,N and Co;N in the case of cobalt, and in the case of iron, Fe,N asa additional phases. The thermal decomposition profile of CoN shows two nitrogen evolution events at 340 and 400°C. The integrated areas for the two events are approximately equal, suggesting a stepwise decomposition to CoN in the first step and formation of metallic cobalt in the second step. The profile is given in Fig. 1 along with the result of a run that was quenched halfway to establish weight loss and to examine the crystalline phase present at that stage. Fig. 1. Thermal decomposition profile, showing mass analyzer response to ion m/e = 28. Trace A decomposition of pure CoN quenched. Trace B pure CoN.
Central to the development of the nanocomposite systems examined in this study are the thermal properties of the ferromagnetic metal nitrides utilized as precursors. Their relatively limited stability allows selective decomposition at moderate temperatures, ~ 600 °C, while the ceramic matrix is unaffected. The thermal decomposition of pure CoN and FeN was examined by Suzuki et al [7,8] who conducted annealing treatments in a stepwise manner; on the other hand, the thermal decomposition of pure Ni;N was examined[6] using a continuous incremental ramp. Pure Co and Fe foils were sputtered in the present study to confirm the stoichiometry of the products formed under the experimental conditions described above. The nitrides produced in both cases were the ones with a 1:1 stoichiometry. Both elements form nitrides with lower nitrogen content, Co,N and Co;N in the case of cobalt, and in the case of iron, Fe,N asa additional phases. The thermal decomposition profile of CoN shows two nitrogen evolution events at 340 and 400°C. The integrated areas for the two events are approximately equal, suggesting a stepwise decomposition to CoN in the first step and formation of metallic cobalt in the second step. The profile is given in Fig. 1 along with the result of a run that was quenched halfway to establish weight loss and to examine the crystalline phase present at that stage. Fig. 1. Thermal decomposition profile, showing mass analyzer response to ion m/e = 28. Trace A decomposition of pure CoN quenched. Trace B pure CoN.
Fig.2 Diffraction pattems of (A) pure crystalline CoN, (B) CoN/BN nanocomposite, and (C) pyrolysis product, pattern corresponds to hep cobalt. The decomposition profiles of all the nanocomposite films studied, NisN/AIN, CoN/SisN,, CoN/BN, and FeN/SisN, exhibit similar patterns to the pure metal nitride compounds. The x-ray diffraction of the nanocomposite films identified the metal nitride as the only crystalline phase. There was significant line broadening which was used to calculate an average particle size that turned out to be < 10 nm in every case. A very significant characteristic of these systems is that the particle size of metal produced by the thermal treatment does not show any crystal grain stowth, probably due to the ceramic matrix which prevents aggregation. This is illustrated in Fig. 2 for the cobalt nitride system. The presence of the ceramic matrix in each composite was revealed by their corresponding typical infrared spectra and also confirmed by direct chemical analysis.
Fig.2 Diffraction pattems of (A) pure crystalline CoN, (B) CoN/BN nanocomposite, and (C) pyrolysis product, pattern corresponds to hep cobalt. The decomposition profiles of all the nanocomposite films studied, NisN/AIN, CoN/SisN,, CoN/BN, and FeN/SisN, exhibit similar patterns to the pure metal nitride compounds. The x-ray diffraction of the nanocomposite films identified the metal nitride as the only crystalline phase. There was significant line broadening which was used to calculate an average particle size that turned out to be < 10 nm in every case. A very significant characteristic of these systems is that the particle size of metal produced by the thermal treatment does not show any crystal grain stowth, probably due to the ceramic matrix which prevents aggregation. This is illustrated in Fig. 2 for the cobalt nitride system. The presence of the ceramic matrix in each composite was revealed by their corresponding typical infrared spectra and also confirmed by direct chemical analysis.
There are extensive experimental observations of size effects on coercivity and its temperature dependence that have been explained[9] in relation to the size of single-domain particles. Coercivity reaches a maximum for single domain particles but decreases both for lar; particles that are multidomain and for smaller particles which fall within the realm of superparamagnetic behavior. Superparamagnetism is dependent on temperature and is evident above a critical temperature, the blocking temperature, Tp, above which thermal motions cont the alignment of the assembly with the applied field and the system is magnetically reversible(coercive field, Hc=0). The assembly is stable at temperatures lower than Ty and the system shows hysteresis. Experimentally we derive Ts from plots of Hc vs T’”. Ty values are 392 K for Ni/AIN, 206 K for Co/BN, 50 K for Co/SisN, and 100 K for Fe/SisN,. The coercive field at room temperature for nickel is 35 Oe; on the other hand, the other systems should not show any coercivity at room temperature under ideal conditions since the Tg values are much wPeReL Aka BD ated = Saturation magnetization of the ferromagnetic nanocomposites derived from the thermal treatment were to about 74% of that for pure metal for Ni/AIN. The corresponding figures are 94 % for Co/BN and 50 % for of Fe/Si3N,. The lower saturation magnetization is common for nanostructured metals and has been explained as arising from the oxidation of the clusters and disorder at interfaces. The iron case is somewhat more complicated since the pyrolysis has to be conducted at temperatures above 700 °C to insure that all of the nitride is converted to the metal. Under those conditions, however reaction between the metal and the ceramic matrix is possible and silicides were formed which accounts in part for the lower saturation magnetization for iron. In common with nanostructured ferromagnetic metals, the nanocomposites examined in this ctdv chawed hvcterecie unoan revereal nf the Feld in a oven temneratnre rance TTwnical
There are extensive experimental observations of size effects on coercivity and its temperature dependence that have been explained[9] in relation to the size of single-domain particles. Coercivity reaches a maximum for single domain particles but decreases both for lar; particles that are multidomain and for smaller particles which fall within the realm of superparamagnetic behavior. Superparamagnetism is dependent on temperature and is evident above a critical temperature, the blocking temperature, Tp, above which thermal motions cont the alignment of the assembly with the applied field and the system is magnetically reversible(coercive field, Hc=0). The assembly is stable at temperatures lower than Ty and the system shows hysteresis. Experimentally we derive Ts from plots of Hc vs T’”. Ty values are 392 K for Ni/AIN, 206 K for Co/BN, 50 K for Co/SisN, and 100 K for Fe/SisN,. The coercive field at room temperature for nickel is 35 Oe; on the other hand, the other systems should not show any coercivity at room temperature under ideal conditions since the Tg values are much wPeReL Aka BD ated = Saturation magnetization of the ferromagnetic nanocomposites derived from the thermal treatment were to about 74% of that for pure metal for Ni/AIN. The corresponding figures are 94 % for Co/BN and 50 % for of Fe/Si3N,. The lower saturation magnetization is common for nanostructured metals and has been explained as arising from the oxidation of the clusters and disorder at interfaces. The iron case is somewhat more complicated since the pyrolysis has to be conducted at temperatures above 700 °C to insure that all of the nitride is converted to the metal. Under those conditions, however reaction between the metal and the ceramic matrix is possible and silicides were formed which accounts in part for the lower saturation magnetization for iron. In common with nanostructured ferromagnetic metals, the nanocomposites examined in this ctdv chawed hvcterecie unoan revereal nf the Feld in a oven temneratnre rance TTwnical
Fig. 4 Atomic force (left side) and magnetic force (right side) micrographs of Ni/AIN composite. Images are about 400 nm on the side. Examination of our films by Atomic Force Microscopy, AFM, and Magnetic Force Microscopy, MEM, reveals a botryoidal morphology, consisting of aggregates of smaller clusters, about 15 nm in diameter. The AFM and MFM images are similar, other than subtle differences in contrast, since the MFM responds to a field gradient generated by the presence of the ferromagnetic particles. This is illustrated in Fig. 4. The AFM and MFM produce an image of the outer ceramic matrix and thus complement the XRD and TEM observations which are sensitive to the denser crystalline metal component in the nanocomposite.
Fig. 4 Atomic force (left side) and magnetic force (right side) micrographs of Ni/AIN composite. Images are about 400 nm on the side. Examination of our films by Atomic Force Microscopy, AFM, and Magnetic Force Microscopy, MEM, reveals a botryoidal morphology, consisting of aggregates of smaller clusters, about 15 nm in diameter. The AFM and MFM images are similar, other than subtle differences in contrast, since the MFM responds to a field gradient generated by the presence of the ferromagnetic particles. This is illustrated in Fig. 4. The AFM and MFM produce an image of the outer ceramic matrix and thus complement the XRD and TEM observations which are sensitive to the denser crystalline metal component in the nanocomposite.
Fig. la. Overview of RF plasma torch reactor showing plasma torch head, expansion and reaction vessel, cyclone separator, and porous metal filters. 1b. Detailed schematic of plasm: torch head used to produce nanoparticles.
Fig. la. Overview of RF plasma torch reactor showing plasma torch head, expansion and reaction vessel, cyclone separator, and porous metal filters. 1b. Detailed schematic of plasm: torch head used to produce nanoparticles.
Fig. 2. Conventional TEM of bare (uncoated) FeCo nanoparticles collected from the RF plasm: torch reactor wall. No carbon source was used during this synthesis. Fig. 3. Carbon coated FeCo nanoparticles produced in an RF plasma torch using metal powder feedstocks and acetylene as the carbon source. The processing conditions used to generate this product are described in the text.
Fig. 2. Conventional TEM of bare (uncoated) FeCo nanoparticles collected from the RF plasm: torch reactor wall. No carbon source was used during this synthesis. Fig. 3. Carbon coated FeCo nanoparticles produced in an RF plasma torch using metal powder feedstocks and acetylene as the carbon source. The processing conditions used to generate this product are described in the text.
Fig. 4. X-ray powder diffraction of bare and carbon coated FeCo nanoparticles collected from the reactor wall. The lower trace (bare FeCo, no carbon) shows signs of the oxidation ("ox" indicates oxide peaks) expected from ultrafine particles lacking a protective carbon coating.
Fig. 4. X-ray powder diffraction of bare and carbon coated FeCo nanoparticles collected from the reactor wall. The lower trace (bare FeCo, no carbon) shows signs of the oxidation ("ox" indicates oxide peaks) expected from ultrafine particles lacking a protective carbon coating.
Magnetic results showed a decrease in the coercivity due to the annealing. Representativ hysteresis curves taken at a high (300° K) and low (5° K) temperature both before and after the annealing are shown in Figure 1. The annealing caused little detectable structural change. TEM measurements revealed that the particles appeared spherical both before and after being annealed, and that the particle size was close (withing statistical measurement error) to being the same. Similarly, x-ray diffraction did not reveal any measurable lattice parameter change. The lack of structural changes shows that the graphite sheets effectively protected the nanocrystals at 550° C. Similarly, the lack of a change in the lattice parameter showed that if any carbon was trapped in the iy (which would have shown a lattice parameter change if dissolved carbon was present)”, it was at a low concentration, resulting in no detectable lattice parameter change after annealing. Similarly, x-ray diffraction revealed that the particles were composed of FCC Cobalt both before and after the anneal. FCC is the high temperature phase for cobalt, normally only being present above 420°C. As has been advance previously, this, was probably due the small particle nature of the material, as opposed to any “quenching” effects*. The lattice parameter and particle size data are summarized in Table I.
Magnetic results showed a decrease in the coercivity due to the annealing. Representativ hysteresis curves taken at a high (300° K) and low (5° K) temperature both before and after the annealing are shown in Figure 1. The annealing caused little detectable structural change. TEM measurements revealed that the particles appeared spherical both before and after being annealed, and that the particle size was close (withing statistical measurement error) to being the same. Similarly, x-ray diffraction did not reveal any measurable lattice parameter change. The lack of structural changes shows that the graphite sheets effectively protected the nanocrystals at 550° C. Similarly, the lack of a change in the lattice parameter showed that if any carbon was trapped in the iy (which would have shown a lattice parameter change if dissolved carbon was present)”, it was at a low concentration, resulting in no detectable lattice parameter change after annealing. Similarly, x-ray diffraction revealed that the particles were composed of FCC Cobalt both before and after the anneal. FCC is the high temperature phase for cobalt, normally only being present above 420°C. As has been advance previously, this, was probably due the small particle nature of the material, as opposed to any “quenching” effects*. The lattice parameter and particle size data are summarized in Table I.
Figure 1. Hysteresis curves for graphite encapsulated cobalt nanocrystals at high and low temperatures both before and after annealing at 550° C for 6 hours. The coercivity is seen t decrease at both temperatures. For superparamagnetic particles, the coercivity is expected to obey the formula”:
Figure 1. Hysteresis curves for graphite encapsulated cobalt nanocrystals at high and low temperatures both before and after annealing at 550° C for 6 hours. The coercivity is seen t decrease at both temperatures. For superparamagnetic particles, the coercivity is expected to obey the formula”:
Figure 3. Diffusion of 99% of the carbon from a 10 nm diameter Co particle. At any temperature above 600 ° C, this takes less than one tenth of a second’.
Figure 3. Diffusion of 99% of the carbon from a 10 nm diameter Co particle. At any temperature above 600 ° C, this takes less than one tenth of a second’.
Figure 1: Magnetization M on linear axis versus applied field H on logarithmic axis for n-Ni (@) and Ni reference sample (©) at 298K. Lines are guides to the eye.
Figure 1: Magnetization M on linear axis versus applied field H on logarithmic axis for n-Ni (@) and Ni reference sample (©) at 298K. Lines are guides to the eye.
278K, SANS data was recorded at the 30m instrument on beamline 7 at the NIST Cold Neutron Research Facil- ity, with a wavelength of 0.6nm and with the magnetic field applied horizontally and perpendicular to the (horizontal) neutron beam. Figure 2 displays azimuthal averaged differential scattering cross sections versus the modulus k of the scattering vector, recorded at 300K with different applied magnetic fields. The intensity is high for small fields; as the field is increased to the maximum experimental value of 1440kA/m (18.1kOe), the intensity decreases by more than two orders of magnitude. The intensity decrease is not a simple scaling, but the shape of the scattering curves changes: the wavevector at which the (negative) curvature on a log-log plot of do/dQ versus k is at maximum increases systematically with increasing field. This indicates the exis- tence of a characteristic length scale of the magnetic microstructure which decreases with increasing H,. Figure 3 compares the experimental data to scattering cross sections computed from Equation (8), with do,,,/dQ and S,, (also displayed) determined from fits to the experimental data. Demagnetizing fields for use in Equation (3) were computed from the experimental magnetization data, with estimates for N, for spheroids with the same axis ratio as the samples. It is seen that the fit is in excellent agreement with experiment over ais CALS aA BE ALAR RAGASAA Bate) ALA lm (496) eR 1AAD Aw 1108 TLOEN 6 Pisxdatiane cee ciithin
278K, SANS data was recorded at the 30m instrument on beamline 7 at the NIST Cold Neutron Research Facil- ity, with a wavelength of 0.6nm and with the magnetic field applied horizontally and perpendicular to the (horizontal) neutron beam. Figure 2 displays azimuthal averaged differential scattering cross sections versus the modulus k of the scattering vector, recorded at 300K with different applied magnetic fields. The intensity is high for small fields; as the field is increased to the maximum experimental value of 1440kA/m (18.1kOe), the intensity decreases by more than two orders of magnitude. The intensity decrease is not a simple scaling, but the shape of the scattering curves changes: the wavevector at which the (negative) curvature on a log-log plot of do/dQ versus k is at maximum increases systematically with increasing field. This indicates the exis- tence of a characteristic length scale of the magnetic microstructure which decreases with increasing H,. Figure 3 compares the experimental data to scattering cross sections computed from Equation (8), with do,,,/dQ and S,, (also displayed) determined from fits to the experimental data. Demagnetizing fields for use in Equation (3) were computed from the experimental magnetization data, with estimates for N, for spheroids with the same axis ratio as the samples. It is seen that the fit is in excellent agreement with experiment over ais CALS aA BE ALAR RAGASAA Bate) ALA lm (496) eR 1AAD Aw 1108 TLOEN 6 Pisxdatiane cee ciithin
In order to demonstrate the effect unambiguously it is important to use far-infrared (FIR) light. In this frequency range the electromagnetic field penetration depth 5 is much smaller than the wavelength of light 4 and the metal behaves as an almost perfect reflector. Here, we examine the FIR transmission of thick (90 um) samples of 10-~m diameter In wire arrays. The wire arrays were prepared by filling the channels of a glass microchannel plate (MCP) with indium. The dense, 76% metal content, microstructure is highly reproducible. The composite is conducting as the In metal fills the full length of the channels. Our previous work has demonstrated the feasibility of injecting porous dielectrics of characteristic pore diameter as small as 5 nm with some selected low melting point materials (Se, Te, In) and alloys (Bi,Te,) by pressure-forcing their melts into the interconnected pores.® The MCP samples’ were cleaned by subsequently washing in acetone, methanol, and distilled H,O, and then dried at 150 °C for several hours; they were injected with the In melt at 400 °C and 60,000 psi. The composite samples were mechanically polished in the shape of a plate with the wires running perpendicular to the plate surface. Fig. 1 shows an electron micrograph of the wire array composite.
In order to demonstrate the effect unambiguously it is important to use far-infrared (FIR) light. In this frequency range the electromagnetic field penetration depth 5 is much smaller than the wavelength of light 4 and the metal behaves as an almost perfect reflector. Here, we examine the FIR transmission of thick (90 um) samples of 10-~m diameter In wire arrays. The wire arrays were prepared by filling the channels of a glass microchannel plate (MCP) with indium. The dense, 76% metal content, microstructure is highly reproducible. The composite is conducting as the In metal fills the full length of the channels. Our previous work has demonstrated the feasibility of injecting porous dielectrics of characteristic pore diameter as small as 5 nm with some selected low melting point materials (Se, Te, In) and alloys (Bi,Te,) by pressure-forcing their melts into the interconnected pores.® The MCP samples’ were cleaned by subsequently washing in acetone, methanol, and distilled H,O, and then dried at 150 °C for several hours; they were injected with the In melt at 400 °C and 60,000 psi. The composite samples were mechanically polished in the shape of a plate with the wires running perpendicular to the plate surface. Fig. 1 shows an electron micrograph of the wire array composite.
Infrared absorption spectra were measured with a Bomem FTIR DA& spectrophotometer.’° A Hg lamp, Mylar film beam splitters, and DTGS detector were used. The instrument’s beam divergence at the sample corresponds to a ratio of focal length to bear diameter of 4. An area of 3 mm? of sample was probed by the beam. Spectra were taken by averaging 8,000 scans. Figure 2 shows the transmission I/I, of the empty MCP in the k < 100 cm” FIR region of transparency of the glass. The reference I, is the intensity transmittec through an aperture of the same size as the sample under the same experimental conditions. The resolution is 1 cm”. The noise is typical of measurements in this frequency range, i increases at around 20 cm” as the lower limit of operation of the instrumental resources is approached. The empty MCP transmission spectrum shows a slow decrease with increasing frequency. This trend continues in the range between 100 and 200 cm" (not shown). Alsc shown in Fig. 2 is the transmission of the In-MCP composite (8000 scans). The transmission of the wire array also shows an overall decrease with increasing frequency and a much stronger frequency dependence. Figure 2. (a) Unpolarized FIR light impinges parallel to the channels. (b) Transmission of the microchannel plate (MCP) and of the 10-micron diameter In wire array composite (MCP-In). The dotted line is the result of a calculation based on Eq. 1 as described in the text. The thickness of the MCP is 100 ym and that of the composite is 91 pm.
Infrared absorption spectra were measured with a Bomem FTIR DA& spectrophotometer.’° A Hg lamp, Mylar film beam splitters, and DTGS detector were used. The instrument’s beam divergence at the sample corresponds to a ratio of focal length to bear diameter of 4. An area of 3 mm? of sample was probed by the beam. Spectra were taken by averaging 8,000 scans. Figure 2 shows the transmission I/I, of the empty MCP in the k < 100 cm” FIR region of transparency of the glass. The reference I, is the intensity transmittec through an aperture of the same size as the sample under the same experimental conditions. The resolution is 1 cm”. The noise is typical of measurements in this frequency range, i increases at around 20 cm” as the lower limit of operation of the instrumental resources is approached. The empty MCP transmission spectrum shows a slow decrease with increasing frequency. This trend continues in the range between 100 and 200 cm" (not shown). Alsc shown in Fig. 2 is the transmission of the In-MCP composite (8000 scans). The transmission of the wire array also shows an overall decrease with increasing frequency and a much stronger frequency dependence. Figure 2. (a) Unpolarized FIR light impinges parallel to the channels. (b) Transmission of the microchannel plate (MCP) and of the 10-micron diameter In wire array composite (MCP-In). The dotted line is the result of a calculation based on Eq. 1 as described in the text. The thickness of the MCP is 100 ym and that of the composite is 91 pm.
We have calculated the FIR absorption of the metal wire array composite using Eq. 1 for the hexagonal lattice. In this calculation we use the index of refraction and absorption coefficient of the MCP glass as given above. The optical constants of bulk In metal (e,,) in the far-infrared are taken from the measurements of Golovashkin et al.!® We adopt their modification of the Drude model for the weak anomalous skin effect. Fig. 3 shows the In absorbance as obtained from their data. The composite absorbance as calculated from Eq. 1 is also shown in Fig. 3, its frequency dependence contrasts with the experimental results. Clearly, this simple approximation does not adequately represent the absorption of the composite, i.e., the calculated absorption is too small for k<50 cm" and it has the wrong frequency dependence. The optical properties of small spherical metal particles have been studied experimentally and theoretically for over twenty years because of their anomalous FIR absorption.” Sample preparation has been an important issue. For coated particles, the absorption is dominated by the enhancement of the electric dipole absorption due to the oxide coating. In that case, the absorption has a k” frequency dependence (like oxide glasses). It is believed that for bare particles clustering allows more closed paths for eddy currents to flow and enhances the absorption due to the magnetic dipole contribution. Although the matter of
We have calculated the FIR absorption of the metal wire array composite using Eq. 1 for the hexagonal lattice. In this calculation we use the index of refraction and absorption coefficient of the MCP glass as given above. The optical constants of bulk In metal (e,,) in the far-infrared are taken from the measurements of Golovashkin et al.!® We adopt their modification of the Drude model for the weak anomalous skin effect. Fig. 3 shows the In absorbance as obtained from their data. The composite absorbance as calculated from Eq. 1 is also shown in Fig. 3, its frequency dependence contrasts with the experimental results. Clearly, this simple approximation does not adequately represent the absorption of the composite, i.e., the calculated absorption is too small for k<50 cm" and it has the wrong frequency dependence. The optical properties of small spherical metal particles have been studied experimentally and theoretically for over twenty years because of their anomalous FIR absorption.” Sample preparation has been an important issue. For coated particles, the absorption is dominated by the enhancement of the electric dipole absorption due to the oxide coating. In that case, the absorption has a k” frequency dependence (like oxide glasses). It is believed that for bare particles clustering allows more closed paths for eddy currents to flow and enhances the absorption due to the magnetic dipole contribution. Although the matter of
FIGURE 1: Equilibrium phase diagram of the Zn-Bi system showing the wide miscibility gap characteristic of this system [12]. In order to investigate the application of ultrasound to the formation of metal/metal matrix composites, the Zn-Bi case was selected as a model system, since both metals ar relatively easy to handle based on their chemical reactivity and low melting points. Additionally the wide miscibility gap in the Zn-Bi phase diagram (See Figure 1) made this system particularly attractive candidate for the ultrasonic formation of metallic emulsions in varyin, concentrations.
FIGURE 1: Equilibrium phase diagram of the Zn-Bi system showing the wide miscibility gap characteristic of this system [12]. In order to investigate the application of ultrasound to the formation of metal/metal matrix composites, the Zn-Bi case was selected as a model system, since both metals ar relatively easy to handle based on their chemical reactivity and low melting points. Additionally the wide miscibility gap in the Zn-Bi phase diagram (See Figure 1) made this system particularly attractive candidate for the ultrasonic formation of metallic emulsions in varyin, concentrations.
FIGURE 2: Schematic representation of the ultrasonic system. vibrations. A schematic representation of the ultrasonic processing system 1s shown in Figure 2. The Zn-Bi composition selected for the present experiments consisted of 10 wt.% Bi. Once the two metals were weighed to achieve the appropriate proportions, they were then melted in a SiO2 tube using a propane torch and were heated to approximately 650°C. To minimize oxide-formation, argon gas containing 4% Hz was continuously sprayed over the surface. When the desired melt temperature was reached, the torch was turned off and the sonication process was initiated by immersing the vibrating microtip in the liquid. While the Ti- alloy used for the horn and tip represents the best horn material from the mechanical and acoustic point of view — combining outstanding acoustic properties with lightness, strength, abrasion resistance and chemical inertness — it tends to form an alloy with the Zn-phase. This interaction results in degradation of the microtip when it is immersed in the melt. Accordingly, in order to minimize the reaction of the Ti-tip with the sample, the sonication process had to be limited to short durations — typically 10-30 seconds. In the present experiments, the melt is first sonicated for 10-15 seconds while it begins cooling down at a typical cooling rate of 10°C per second. The solidification process is subsequently accelerated by spraying water on the outside of the Si02 tube, while maintaining the sonication conditions. until the solidification is complete. This results
FIGURE 2: Schematic representation of the ultrasonic system. vibrations. A schematic representation of the ultrasonic processing system 1s shown in Figure 2. The Zn-Bi composition selected for the present experiments consisted of 10 wt.% Bi. Once the two metals were weighed to achieve the appropriate proportions, they were then melted in a SiO2 tube using a propane torch and were heated to approximately 650°C. To minimize oxide-formation, argon gas containing 4% Hz was continuously sprayed over the surface. When the desired melt temperature was reached, the torch was turned off and the sonication process was initiated by immersing the vibrating microtip in the liquid. While the Ti- alloy used for the horn and tip represents the best horn material from the mechanical and acoustic point of view — combining outstanding acoustic properties with lightness, strength, abrasion resistance and chemical inertness — it tends to form an alloy with the Zn-phase. This interaction results in degradation of the microtip when it is immersed in the melt. Accordingly, in order to minimize the reaction of the Ti-tip with the sample, the sonication process had to be limited to short durations — typically 10-30 seconds. In the present experiments, the melt is first sonicated for 10-15 seconds while it begins cooling down at a typical cooling rate of 10°C per second. The solidification process is subsequently accelerated by spraying water on the outside of the Si02 tube, while maintaining the sonication conditions. until the solidification is complete. This results
FIGURE 3 (a) and (b): SEM image of a Zn-Bi composite obtained by sonication of the two molten immiscible liquid metals.
FIGURE 3 (a) and (b): SEM image of a Zn-Bi composite obtained by sonication of the two molten immiscible liquid metals.
SAYERS ALY MOL do LULLY UCTIS® ao UPahiled layed Alloy layers of Al(O) can also be deposited with PLD by sequentially ablating Al and Al,( targets [6]. Composition is determined by the ratio of pulses per target, and the amount | material deposited per cycle is kept at ~1 monolayer to achieve a near-homogeneous depositio The energies of atoms in the laser-ablated plasma plume are ~10 eV, which is insufficient penetrate into the layer. However, the atoms reach the specimen with greater energies sever hundred times k,T (0.025 eV) and may thermally break surface bonds or directly displace surfa atoms, which are less strongly bound. The dark-field image in Fig. 3b shows a microstructure. 5 - 25 nm fcc Al grains with ~1 nm oxides, similar to that with ECR but on a finer scale. TI strengths of PLD alloys with 20 and 29 at.% O are 1.7 and 2.5 GPa, respectively [6], which a higher than for corresponding ECR alloys but below implanted alloys. Thus this relatively lo energy deposition method also produces a nanocomposite microstructure with high yield stress. Tes BY'D anA DT TY lasseew ol laser iret awarne eon reremanrtias wl thameaninitotad ‘allan laces
SAYERS ALY MOL do LULLY UCTIS® ao UPahiled layed Alloy layers of Al(O) can also be deposited with PLD by sequentially ablating Al and Al,( targets [6]. Composition is determined by the ratio of pulses per target, and the amount | material deposited per cycle is kept at ~1 monolayer to achieve a near-homogeneous depositio The energies of atoms in the laser-ablated plasma plume are ~10 eV, which is insufficient penetrate into the layer. However, the atoms reach the specimen with greater energies sever hundred times k,T (0.025 eV) and may thermally break surface bonds or directly displace surfa atoms, which are less strongly bound. The dark-field image in Fig. 3b shows a microstructure. 5 - 25 nm fcc Al grains with ~1 nm oxides, similar to that with ECR but on a finer scale. TI strengths of PLD alloys with 20 and 29 at.% O are 1.7 and 2.5 GPa, respectively [6], which a higher than for corresponding ECR alloys but below implanted alloys. Thus this relatively lo energy deposition method also produces a nanocomposite microstructure with high yield stress. Tes BY'D anA DT TY lasseew ol laser iret awarne eon reremanrtias wl thameaninitotad ‘allan laces
Figure 6. Nanoindentation of AlgsN5 and fit to low-load portion using indicated alloy parameters. WV REAL CUO LUPE LLU OLE WV PEWVE VLA) ae BANEE UAE CLI AD OMI 1 L1G. SU. Nanoindentation was done on a specimen implanted with 5 at.% N to a depth of 0.5 um, and the response curve is shown in Fig. 6. For loads < 0.4 mN (depths < 70 nm) the alloy layer appears quite strong; finite-element simulation using a yield stress of 2.15 GPa is seen to fit this portion of the curve well. However at higher loads, abrupt penetrations to greater depth are seen, termed “‘pop-ins” [7]. The origin of these features is not yet clear for N-implanted Al; they were not observed in O-implanted Al. Considering the initial part of the curve only, the Al,,N, alloy layer appears stronger than the corresponding Al,,O, alloy with yield stress = 1.4 GPa.
Figure 6. Nanoindentation of AlgsN5 and fit to low-load portion using indicated alloy parameters. WV REAL CUO LUPE LLU OLE WV PEWVE VLA) ae BANEE UAE CLI AD OMI 1 L1G. SU. Nanoindentation was done on a specimen implanted with 5 at.% N to a depth of 0.5 um, and the response curve is shown in Fig. 6. For loads < 0.4 mN (depths < 70 nm) the alloy layer appears quite strong; finite-element simulation using a yield stress of 2.15 GPa is seen to fit this portion of the curve well. However at higher loads, abrupt penetrations to greater depth are seen, termed “‘pop-ins” [7]. The origin of these features is not yet clear for N-implanted Al; they were not observed in O-implanted Al. Considering the initial part of the curve only, the Al,,N, alloy layer appears stronger than the corresponding Al,,O, alloy with yield stress = 1.4 GPa.
structure of AIN differs from that of cubic y-Al,O, and its bonds are more covalent. Annealed Al was implanted with 10 at.% N at room temperature and examined with TEM. Electron diffraction patterns show extra spots matching reflections of AIN, which align in the pattern of intense fcc Al reflections as in Fig. 5a. The precipitates align with the c-axis along a <]11> matrix direction and with their {1-100} planes parallel to {2-20} Al planes. Dark-field imaging with these reflections shows precipitates ~2 nm in diameter as seen in Fig. 5b.
structure of AIN differs from that of cubic y-Al,O, and its bonds are more covalent. Annealed Al was implanted with 10 at.% N at room temperature and examined with TEM. Electron diffraction patterns show extra spots matching reflections of AIN, which align in the pattern of intense fcc Al reflections as in Fig. 5a. The precipitates align with the c-axis along a <]11> matrix direction and with their {1-100} planes parallel to {2-20} Al planes. Dark-field imaging with these reflections shows precipitates ~2 nm in diameter as seen in Fig. 5b.
Fig.4 The relationship between weight percentages of UFP and bending strength.
Fig.4 The relationship between weight percentages of UFP and bending strength.
Figure 3- X-ray diffraction pattern from a series of Cu.g9gCo.29 samples heat treated for 1 hour in flowing hydrogen gas at the specified temperatures. separate hydroxide and oxide crystals. Thus, upon reduction separate Fe and Cu crystallites are formed. Further, the decrease in surface energy of forming coherent fcc interfaces appears insufficient to overcome the tendency of Fe to crystallize into a bcc structure. The situation appears to be different in the case of Cu/Co. For this system, the data is most easily interpreted by supposing that either the hydroxide or oxide precursors contain a mixture of Cu and Co. Even though the X-ray spectrum consists of peaks consistent with a single structure, after the
Figure 3- X-ray diffraction pattern from a series of Cu.g9gCo.29 samples heat treated for 1 hour in flowing hydrogen gas at the specified temperatures. separate hydroxide and oxide crystals. Thus, upon reduction separate Fe and Cu crystallites are formed. Further, the decrease in surface energy of forming coherent fcc interfaces appears insufficient to overcome the tendency of Fe to crystallize into a bcc structure. The situation appears to be different in the case of Cu/Co. For this system, the data is most easily interpreted by supposing that either the hydroxide or oxide precursors contain a mixture of Cu and Co. Even though the X-ray spectrum consists of peaks consistent with a single structure, after the
indicate that a homogeneous supersaturated solid solution of Cu,g9Co,20 should not have any ferromagnetic component. The saturation magnetization? > of as- prepared, ball-milled samples of Cu.g9Co.29 also is much smaller than the values shown in Fig. 6. Thus, despite the x-ray data, it is clear that our Cu.g9Co.29 sample heated for one hour at 215 °C is not a homogeneous solid solution. On the other hand, since the saturation magnetization per gram of Cois smaller than that of bulk fcc Co (161 emu/gm), some of the Co is in some form other than pure Co. Some of the Co is either in a supersaturated solid solution or is in the form of an oxide.
indicate that a homogeneous supersaturated solid solution of Cu,g9Co,20 should not have any ferromagnetic component. The saturation magnetization? > of as- prepared, ball-milled samples of Cu.g9Co.29 also is much smaller than the values shown in Fig. 6. Thus, despite the x-ray data, it is clear that our Cu.g9Co.29 sample heated for one hour at 215 °C is not a homogeneous solid solution. On the other hand, since the saturation magnetization per gram of Cois smaller than that of bulk fcc Co (161 emu/gm), some of the Co is in some form other than pure Co. Some of the Co is either in a supersaturated solid solution or is in the form of an oxide.
Figure 7 - Fourier transformed (FT) EXAFS data versus radial distance from the absorbing atom. The peaks superscripted 'o' and 'm' correspond to neighbors in an oxide and close-packed metal phases, respectively. The subscript denotes the atomic shell number relative to the absorbing atom. Cu foil data are included in panels (b) and (d) for comparison purposes. (Data analyzied using a k-range of 2.5-12.5A-! and k2-weighting.)
Figure 7 - Fourier transformed (FT) EXAFS data versus radial distance from the absorbing atom. The peaks superscripted 'o' and 'm' correspond to neighbors in an oxide and close-packed metal phases, respectively. The subscript denotes the atomic shell number relative to the absorbing atom. Cu foil data are included in panels (b) and (d) for comparison purposes. (Data analyzied using a k-range of 2.5-12.5A-! and k2-weighting.)
Bilge J. ACHSNC SUCSS-SUiall CULYES OF NANOCrystaiine alloys A-f. alurntecaye et ~D.\ aluminum (69 GPa) [27]. By contrast, the stresses realized in these tests are significantly higher than the generally referenced value of the yield strength of bulk aluminum (cy=20 MPa). The stress-strain curve for nanocrystalline aluminium (A) exhibits nonlinear permanent deformation and is suggestive of dislocation motion and plasticity. By comparison, the Al-Zr alloys remained very linear and fractured with very little or no plasticity. This lack of plasticity, even at very high stresses, had been taken as a strong indication that dislocation motion is effectively inhibited or blocked in these nanocrystalline alloys. This brittle behaviour is expected when the size of the grains of a previously ductile metal is reduced to the nanometer scale [4, 23]. It is however, interesting to note that the fracture strength decreases with increasing Zr content, as seen on Fig. 5, which emphasizes the role of the flaws in this type of alloy [28]. Material from the same disks as used for samples A and B had been tested previously at Northwestern. The slightly higher values reported here, especially in the case of the more ductile sample (A) have been attributed to a miscalibration of the compliance of the glue and the strain gauge used in the previous study (but the maximum strains have been found to be very close). The fracture stresses of sample B are in decent agreement between the two studies (250 Mpa compared to 300 Mpa here).
Bilge J. ACHSNC SUCSS-SUiall CULYES OF NANOCrystaiine alloys A-f. alurntecaye et ~D.\ aluminum (69 GPa) [27]. By contrast, the stresses realized in these tests are significantly higher than the generally referenced value of the yield strength of bulk aluminum (cy=20 MPa). The stress-strain curve for nanocrystalline aluminium (A) exhibits nonlinear permanent deformation and is suggestive of dislocation motion and plasticity. By comparison, the Al-Zr alloys remained very linear and fractured with very little or no plasticity. This lack of plasticity, even at very high stresses, had been taken as a strong indication that dislocation motion is effectively inhibited or blocked in these nanocrystalline alloys. This brittle behaviour is expected when the size of the grains of a previously ductile metal is reduced to the nanometer scale [4, 23]. It is however, interesting to note that the fracture strength decreases with increasing Zr content, as seen on Fig. 5, which emphasizes the role of the flaws in this type of alloy [28]. Material from the same disks as used for samples A and B had been tested previously at Northwestern. The slightly higher values reported here, especially in the case of the more ductile sample (A) have been attributed to a miscalibration of the compliance of the glue and the strain gauge used in the previous study (but the maximum strains have been found to be very close). The fracture stresses of sample B are in decent agreement between the two studies (250 Mpa compared to 300 Mpa here).
Except for a low density of round pores the compaction and HIiPping processes resulted in a complete dense material. However, some of the interfaces between ribbons were not bonded perfectly and appeared dark as marked by arrows in Figure 1a. Occasionally regions with a much coarser microstructure were found. Figure 1b presents an example of such microstructure. These regions are believed to be pieces of the master alloy which did not melt completely and were ejected with the melt during melt spinning. As shown in Figure 1b these remains are very brittle and cracked during the final swaging process.
Except for a low density of round pores the compaction and HIiPping processes resulted in a complete dense material. However, some of the interfaces between ribbons were not bonded perfectly and appeared dark as marked by arrows in Figure 1a. Occasionally regions with a much coarser microstructure were found. Figure 1b presents an example of such microstructure. These regions are believed to be pieces of the master alloy which did not melt completely and were ejected with the melt during melt spinning. As shown in Figure 1b these remains are very brittle and cracked during the final swaging process.
Figure 3. SEM fractograph of a HiPped and swaged tensile sample. Figure 2. TEM micrograph of the as- HIPped material. The tensile specimens made from the wire sample showed an average ultimate tensile strength of 514 Mpa (459 to 542 MPa) and an average maximum uniform strain and total elongation of 2.5% and 6.7%, respectively. Figure 3 presents an SEM fractograph of the fracture surface. The fracture occurred by microvoid coalescence mechanism. The voids were submicron in size and were possibly initiated at the niobium particles. Delamination was detected at two locations on the fracture surface as can be seen in Figure 3: at the ribbon boundaries (marked as B) and at the regions with the coarse microstructure (marked as C). The delamination process is believed to contribute to the scatter observed in the tensile results. Occasionally, the pores that formed upon the HIPping were also detected on the fracture surface (marked as A).
Figure 3. SEM fractograph of a HiPped and swaged tensile sample. Figure 2. TEM micrograph of the as- HIPped material. The tensile specimens made from the wire sample showed an average ultimate tensile strength of 514 Mpa (459 to 542 MPa) and an average maximum uniform strain and total elongation of 2.5% and 6.7%, respectively. Figure 3 presents an SEM fractograph of the fracture surface. The fracture occurred by microvoid coalescence mechanism. The voids were submicron in size and were possibly initiated at the niobium particles. Delamination was detected at two locations on the fracture surface as can be seen in Figure 3: at the ribbon boundaries (marked as B) and at the regions with the coarse microstructure (marked as C). The delamination process is believed to contribute to the scatter observed in the tensile results. Occasionally, the pores that formed upon the HIPping were also detected on the fracture surface (marked as A).
RESULTS power diffractometer Siemens D-500 instrument at room temperature. The ND experi- ments were taken in the ISIS Rutherford Appleton Laboratory on a ”Liquids and Amor- phous Differactometer” by ”in stu” measurements of the diffraction patterns during the heat treatment. The Méssbauer spectra were measured at room temperature using a °’Co in Rh source with calibration of a-Fe foil, and fitted with a histogram magnetic hyperfine- field distribution by the Brand [7] and Hesse-Rubartsch method[8]. The sub-spectra of the nanocrystalline Fe(Si) phase were fitted by superimposing sexetets of Lorentzian li- nes corresponding to the different Fe neighbourhoods in Fe(Si) phase, in these case the Gaussian-shaped hyperfine distribution was related with the remaining amorphous phase.
RESULTS power diffractometer Siemens D-500 instrument at room temperature. The ND experi- ments were taken in the ISIS Rutherford Appleton Laboratory on a ”Liquids and Amor- phous Differactometer” by ”in stu” measurements of the diffraction patterns during the heat treatment. The Méssbauer spectra were measured at room temperature using a °’Co in Rh source with calibration of a-Fe foil, and fitted with a histogram magnetic hyperfine- field distribution by the Brand [7] and Hesse-Rubartsch method[8]. The sub-spectra of the nanocrystalline Fe(Si) phase were fitted by superimposing sexetets of Lorentzian li- nes corresponding to the different Fe neighbourhoods in Fe(Si) phase, in these case the Gaussian-shaped hyperfine distribution was related with the remaining amorphous phase.
Figure 2: The JMAE exponent,n, of Fe(Si) crystallization process as a function of transformed fraction, annealing the samp- les r=5 and 2=8 at 490, 500 and 510°C. beginning of crystallization is different for the different samples. This means that the controlling mechanisms during the crystallization process are different. For the sample z=5 it is an interface controlled grain growth with a increasing nucleation rate at the beginning of crystallization{9], then the nucleation rate saturates and diffusion limited growth dominates. For the sample z=8, the beginning of crystallization is diffusion con- trolled growth with increasing nucleation rate[9], and then diffusion limited growth till the end of the process.
Figure 2: The JMAE exponent,n, of Fe(Si) crystallization process as a function of transformed fraction, annealing the samp- les r=5 and 2=8 at 490, 500 and 510°C. beginning of crystallization is different for the different samples. This means that the controlling mechanisms during the crystallization process are different. For the sample z=5 it is an interface controlled grain growth with a increasing nucleation rate at the beginning of crystallization{9], then the nucleation rate saturates and diffusion limited growth dominates. For the sample z=8, the beginning of crystallization is diffusion con- trolled growth with increasing nucleation rate[9], and then diffusion limited growth till the end of the process.
Figure 3: The lattice parameter of Fe(Si) phase obtained by XRD in the samples z=5, 7, 8 and 12 as a function of silicon content.
Figure 3: The lattice parameter of Fe(Si) phase obtained by XRD in the samples z=5, 7, 8 and 12 as a function of silicon content.
Figure 6: The relative intensity of near- neighbour sites of Fe(Si) phase for sampes z=5(a) and x=12 (b).
Figure 6: The relative intensity of near- neighbour sites of Fe(Si) phase for sampes z=5(a) and x=12 (b).
UMBCLOSCOPYy \ 1 ioivi). BDBT was performed to characterize the mechanical behavior of both nanocrystalline anc coarse grained NiAl samples. All samples were mechanically polished prior to testing to produce « 0.0544m surface finish, A compression cage fixture was designed and built for thes measurements. A schematic representation of the BDBT apparatus is shown in Fig. 1. The 9 mn diameter specimens were freely supported on a 7 mm diameter ring and load was applied to the center of the disk through a flat punch having a 1.15 mm diameter. The BDBT load-displacemen data were analyzed following the procedures outlined in [5].
UMBCLOSCOPYy \ 1 ioivi). BDBT was performed to characterize the mechanical behavior of both nanocrystalline anc coarse grained NiAl samples. All samples were mechanically polished prior to testing to produce « 0.0544m surface finish, A compression cage fixture was designed and built for thes measurements. A schematic representation of the BDBT apparatus is shown in Fig. 1. The 9 mn diameter specimens were freely supported on a 7 mm diameter ring and load was applied to the center of the disk through a flat punch having a 1.15 mm diameter. The BDBT load-displacemen data were analyzed following the procedures outlined in [5].
tdi") eli iia Tica aka I a a i a a with an average size of ~6 nm in this particular case. Before disk bend testing nanocrystalline NiAl specimens, the performance of the BDB’ apparatus was evaluated by testing standard coarse grained stainless steel and aluminum samples These materials have very well characterized mechanical properties. Load-displacement curve: obtained were similar to load-displacement curves obtained by other investigators on similar]; ductile materials when subjected to disk bend testing [7]. Also, yield stress and modulus value: obtained were within +5% of the literature values for the same materials tested by more conventional tensile test methods.
tdi") eli iia Tica aka I a a i a a with an average size of ~6 nm in this particular case. Before disk bend testing nanocrystalline NiAl specimens, the performance of the BDB’ apparatus was evaluated by testing standard coarse grained stainless steel and aluminum samples These materials have very well characterized mechanical properties. Load-displacement curve: obtained were similar to load-displacement curves obtained by other investigators on similar]; ductile materials when subjected to disk bend testing [7]. Also, yield stress and modulus value: obtained were within +5% of the literature values for the same materials tested by more conventional tensile test methods.
A representative room temperature load-displacement curve for n-NiAl-1 and coarse grained NiAl sample strained at 1.4x10-6 sec! is shown in Fig. 3. After correction for the expected initial non linear region [7], the yield stresses of nanocrystalline samples were calculated from the total loads at yielding (taking into account the sample thicknesses according to the procedures described in [5]). Yield stress values are given in Table I. Coarse-grained NiAI samples fractured prior to yielding and the stresses at fracture were typically approximately 75 MPa, well below the yield stresses of nanocrystalline samples. It has been shown by Li et al. [7] that for disk bend tests plastic yielding occurs at the onset of the deviation from linearity in the load-displacement curve and that the yield strength obtained by disk-bend testing is comparable to the tensile yield strength. A significant and most important observation in the present study is that a deviation from linearity is observed in the load- displacement curve for nanocrystalline NiAl, indicating that NiAl in nanocrystalline form exhibits yield behavior and thus measurable room temperature ductility. A deviation from linearity was not observed in eight CG-NiAl samples tested under similar conditions. Strains-to-failure of 0.1 - 0.2% and 0.016 - 0.06% were observed at room temperature in the present study for n-NiAl and CG-NiAI, respectively. The values of total strain obtained for n-NiAl samples may be hampered by the residual porosity since stress concentrations could develop at the voids resulting premature fracture. Therefore, the measured strain-to-failure values in the present case are most likely lower limits. Further improvements in ductility may be obtained if denser specimens can be obtained in the future. The failure mode of n-NiAl was determined to be ductile from postmortem examination of broken pieces using SEM, as seen in Fig. 4. The apparent improvement in the room temperature ductility in n-NiAl can be understood
A representative room temperature load-displacement curve for n-NiAl-1 and coarse grained NiAl sample strained at 1.4x10-6 sec! is shown in Fig. 3. After correction for the expected initial non linear region [7], the yield stresses of nanocrystalline samples were calculated from the total loads at yielding (taking into account the sample thicknesses according to the procedures described in [5]). Yield stress values are given in Table I. Coarse-grained NiAI samples fractured prior to yielding and the stresses at fracture were typically approximately 75 MPa, well below the yield stresses of nanocrystalline samples. It has been shown by Li et al. [7] that for disk bend tests plastic yielding occurs at the onset of the deviation from linearity in the load-displacement curve and that the yield strength obtained by disk-bend testing is comparable to the tensile yield strength. A significant and most important observation in the present study is that a deviation from linearity is observed in the load- displacement curve for nanocrystalline NiAl, indicating that NiAl in nanocrystalline form exhibits yield behavior and thus measurable room temperature ductility. A deviation from linearity was not observed in eight CG-NiAl samples tested under similar conditions. Strains-to-failure of 0.1 - 0.2% and 0.016 - 0.06% were observed at room temperature in the present study for n-NiAl and CG-NiAI, respectively. The values of total strain obtained for n-NiAl samples may be hampered by the residual porosity since stress concentrations could develop at the voids resulting premature fracture. Therefore, the measured strain-to-failure values in the present case are most likely lower limits. Further improvements in ductility may be obtained if denser specimens can be obtained in the future. The failure mode of n-NiAl was determined to be ductile from postmortem examination of broken pieces using SEM, as seen in Fig. 4. The apparent improvement in the room temperature ductility in n-NiAl can be understood
Figure 2. XRD patterns of Fe-10wt.%Al powders milled for 0, 40, 60 and 540 min.
Figure 2. XRD patterns of Fe-10wt.%Al powders milled for 0, 40, 60 and 540 min.
Figure 5. XRD patterns of Fe-10wt.%A] powder milled for 10 min and annealed prior to (sample A) and after (sample B) the first exothermic peak and after the second exothermic peak (sample C) in DSC diagram. In order to study the kinetics of the formation of solid solution during ball milling, the ball milled Fe-4wt.%Al, which has single exothermic peak in the thermal diagram, was employed for XRD and DSC analysis. It is reasonable to assume here that the formation of solid solution consists of both a mechanically induced component (during bal! milling) and a thermally induced component (during DSC _ heating), with the latter being reflected in the DSC diagrams (As seen in Figures 3 and 4). If the thermal component of enthalpy of solid solution formation is known, the mechanical component of enthalpy of formation of solid solution may be calculated provided that the overall standard enthalpy of formation (AH{ota1) is available. In the present experiment, the standard enthalpy of formation of solid solution Fe,Al;.x (x atomic content) at 293 K can be approximated by [9]
Figure 5. XRD patterns of Fe-10wt.%A] powder milled for 10 min and annealed prior to (sample A) and after (sample B) the first exothermic peak and after the second exothermic peak (sample C) in DSC diagram. In order to study the kinetics of the formation of solid solution during ball milling, the ball milled Fe-4wt.%Al, which has single exothermic peak in the thermal diagram, was employed for XRD and DSC analysis. It is reasonable to assume here that the formation of solid solution consists of both a mechanically induced component (during bal! milling) and a thermally induced component (during DSC _ heating), with the latter being reflected in the DSC diagrams (As seen in Figures 3 and 4). If the thermal component of enthalpy of solid solution formation is known, the mechanical component of enthalpy of formation of solid solution may be calculated provided that the overall standard enthalpy of formation (AH{ota1) is available. In the present experiment, the standard enthalpy of formation of solid solution Fe,Al;.x (x atomic content) at 293 K can be approximated by [9]
Figure 6. Formation of Fe-4wt.%Al solid solution as a function of milling time.
Figure 6. Formation of Fe-4wt.%Al solid solution as a function of milling time.
‘igure 1. a) Left: Transmission electron micrograph of WC/Co namocomposite. Rignt: Electron diffraction of binder area C b) EDS spectrum of area C of a): the composition is 43 wt% W, 57 wt% Co
‘igure 1. a) Left: Transmission electron micrograph of WC/Co namocomposite. Rignt: Electron diffraction of binder area C b) EDS spectrum of area C of a): the composition is 43 wt% W, 57 wt% Co
conventional composites but the increase of hardness in the nano-structured composites does no’ decrease their bulk fracture toughness (Figure 3). Figure 2: Hardness of the WC/Co composites as a function of the dislocation mean free path in the Co binder phase. The symbols identify the samples, the numbers the cobalt content. Figure 3: Vickers indentation hardness of WC/Co composites as a function of their hardness. Note that the toughness decreases with increasing hardness for conventional, but not for nanostructured cermets.
conventional composites but the increase of hardness in the nano-structured composites does no’ decrease their bulk fracture toughness (Figure 3). Figure 2: Hardness of the WC/Co composites as a function of the dislocation mean free path in the Co binder phase. The symbols identify the samples, the numbers the cobalt content. Figure 3: Vickers indentation hardness of WC/Co composites as a function of their hardness. Note that the toughness decreases with increasing hardness for conventional, but not for nanostructured cermets.
K. Jia and T, E. Fischer, Wear of Materials 1997, San Diego Cal. April 21
K. Jia and T, E. Fischer, Wear of Materials 1997, San Diego Cal. April 21
Fig.1 (a) A cross section SAD pattern from the 7.5/2.5 multilayer. The spots have been indexed on the basis of the <2110> and <1010> zone axes for hcp Ti and <110> and <121> zone axes for fec Al. (b) A high resolution image from the same multilayer showing the hcp stacking in the Ti layers and fec stacking in the Al layer.
Fig.1 (a) A cross section SAD pattern from the 7.5/2.5 multilayer. The spots have been indexed on the basis of the <2110> and <1010> zone axes for hcp Ti and <110> and <121> zone axes for fec Al. (b) A high resolution image from the same multilayer showing the hcp stacking in the Ti layers and fec stacking in the Al layer.
Fig.2. A high resolution image showing the foc stacking sequence in both Ti and Al layers of the 5/5 multilayer.
Fig.2. A high resolution image showing the foc stacking sequence in both Ti and Al layers of the 5/5 multilayer.
Fig.4. A schematic plot of AG/A versus the bilayer (1). Fig.5. A schematic plot representing the effect of The f/f line represents only Ti transforming from hcp to changing the Ti:Al ratio on the boundaries of different fcc. Similarly, the h/h line represents Al transforming. phase stability regions. Three different Ti:Al ratios, 1:3, The h/f line represents the bulk structures for both Ti 1:1 and 3:1 have been considered. CMW value of 10 nm and Al with no transitions occuring in either. is expected to lie in the hashed region.
Fig.4. A schematic plot of AG/A versus the bilayer (1). Fig.5. A schematic plot representing the effect of The f/f line represents only Ti transforming from hcp to changing the Ti:Al ratio on the boundaries of different fcc. Similarly, the h/h line represents Al transforming. phase stability regions. Three different Ti:Al ratios, 1:3, The h/f line represents the bulk structures for both Ti 1:1 and 3:1 have been considered. CMW value of 10 nm and Al with no transitions occuring in either. is expected to lie in the hashed region.
The formation of y-TiAl in Ti/Al thin film diffusion couples has not been reported previously. It should be noted that research on compound formation in Ti/AI bilayers has been carried out for more than 20 years [13]; but there has never been any report of formation of intermetallic com- pounds other than Al3Ti in such bilayers. A possible explanation for the absence of intermetallic compounds other than Al3Ti was suggested by Gosele and Tu [14], although their model refers to planar morphologies rather than the more equiaxed transformation products observed in the present study. They introduced the concept of missing compounds in thin film diffusion couples as com- pared with bulk diffusion couples. A critical thickness is defined where a transition occurs from interfacial reaction to diffusion control of the rate of reaction. A second reaction product layer will
The formation of y-TiAl in Ti/Al thin film diffusion couples has not been reported previously. It should be noted that research on compound formation in Ti/AI bilayers has been carried out for more than 20 years [13]; but there has never been any report of formation of intermetallic com- pounds other than Al3Ti in such bilayers. A possible explanation for the absence of intermetallic compounds other than Al3Ti was suggested by Gosele and Tu [14], although their model refers to planar morphologies rather than the more equiaxed transformation products observed in the present study. They introduced the concept of missing compounds in thin film diffusion couples as com- pared with bulk diffusion couples. A critical thickness is defined where a transition occurs from interfacial reaction to diffusion control of the rate of reaction. A second reaction product layer will
The final orientation studied was also an arrester type but with the crack plane perpendicular to [101] and the crack front in the [101] direction. As expected with the presence of a <110>{111} slip system two Shockley partials similar to the ones observed for pure copper and nickel, in Figure 1, were emitted. In Figure 4a, the crack was observed to close at a stress of 0.64 as the dislocations partially relieve the lattice strain. At a stress intensity of 1.28, Fig. 4b, the crack remains in its original position and two more Shockley partials were emitted. As in the pure metals, all the dislocations where of the 1/6<121> type on {111} planes. As the crack blunts, the emitted dislocations appear to stay at the interfaces. These dislocations will control the fracture behavior in this ductile orientation. Figure 2: Crack growth in the delamination orientation, (010)[001], at stresses of a). 0.32 and b). 0.48 MPa*m’”.
The final orientation studied was also an arrester type but with the crack plane perpendicular to [101] and the crack front in the [101] direction. As expected with the presence of a <110>{111} slip system two Shockley partials similar to the ones observed for pure copper and nickel, in Figure 1, were emitted. In Figure 4a, the crack was observed to close at a stress of 0.64 as the dislocations partially relieve the lattice strain. At a stress intensity of 1.28, Fig. 4b, the crack remains in its original position and two more Shockley partials were emitted. As in the pure metals, all the dislocations where of the 1/6<121> type on {111} planes. As the crack blunts, the emitted dislocations appear to stay at the interfaces. These dislocations will control the fracture behavior in this ductile orientation. Figure 2: Crack growth in the delamination orientation, (010)[001], at stresses of a). 0.32 and b). 0.48 MPa*m’”.
Figure 4: Dislocation emission in the (101)[101] orientation and stresses of a). 0.64 and b). 1.28 MPa*m’”,
Figure 4: Dislocation emission in the (101)[101] orientation and stresses of a). 0.64 and b). 1.28 MPa*m’”,
ee ae a, i, Laie a a a ae these host compounds in their pure form at reduced temperatures (700- -900 °C). The principal aim of this paper is to demonstrate the feasibility of a simple polymerizable complex (PC) route, known as the Pechini method [19], for the synthesis of photocatalysts. BaTi,O, and its composites combined with RuO, have been chosen as the focus of the present work, and the PC technique was successfully applied to the synthesis of pure BaTi,O, at 700-900 °C [20]. The PC method is based on polymerization between citric acid (CA) and ethylene glyco! (EG) in the presence of soluble metal salts with required amounts to form a polyester-type resin, inside which metal-CA complex species remain soluble to sustain the scale of mixing of different metals almost molecularly homogeneous [21-27]. An important aspect of this technique in the synthesis of BaTi,O, is that the individual metal CA complexes can be immobilized in a rigid polyester network while preserving the initial stoichiometric ratio of Ba and Ti upon polymerizatior (see a schematic picture shown in Fig.1). Immobilization of the metal complexes in a rigid polymer
ee ae a, i, Laie a a a ae these host compounds in their pure form at reduced temperatures (700- -900 °C). The principal aim of this paper is to demonstrate the feasibility of a simple polymerizable complex (PC) route, known as the Pechini method [19], for the synthesis of photocatalysts. BaTi,O, and its composites combined with RuO, have been chosen as the focus of the present work, and the PC technique was successfully applied to the synthesis of pure BaTi,O, at 700-900 °C [20]. The PC method is based on polymerization between citric acid (CA) and ethylene glyco! (EG) in the presence of soluble metal salts with required amounts to form a polyester-type resin, inside which metal-CA complex species remain soluble to sustain the scale of mixing of different metals almost molecularly homogeneous [21-27]. An important aspect of this technique in the synthesis of BaTi,O, is that the individual metal CA complexes can be immobilized in a rigid polyester network while preserving the initial stoichiometric ratio of Ba and Ti upon polymerizatior (see a schematic picture shown in Fig.1). Immobilization of the metal complexes in a rigid polymer
Fig. 2 Flow chart for the polymerizable complex (PC) procedure used to prepare BaTi,O,.
Fig. 2 Flow chart for the polymerizable complex (PC) procedure used to prepare BaTi,O,.
Fig.3 X-ray diffraction patterns of the Ba/Ti=1/4 composition powder precursor (derived from an ammonia containing solution (pH=3)) calcined in static air for 2 h at 600 °C (a), 650 °C (b), 700 °C (c), 800 °C (d), 900 °C (e), 1000 °C (f), and 1200 °C (g). One of the common problems in the synthesis of BaTi,O, by wet chemical routes is the difficulty in controlling the rapid hydrolysis of Ti-alkoxides (chemicals most frequently used as sources of Ti) compared with barium species. For instance, in the previously reported sol-gel synthesis of BaTi,O, [18], gels derived from Ba and Ti alkoxide precursors have producec strongly multiphase samples even after the heat-treatment at 1200 °C. The preferential hydrolysis of Ti-alkoxides can form Ti-rich clusters, which destroys the cation composition of the original solution. This result tells us that Ti-alkoxides should be modified by certain organic compounds such as acetic acid to control the degree of hydrolysis and subsequent polycondensation reactions [21, 28-30]. A similar event has been observed in the present PC synthesis of BaTi,O,, when the pH of the original solution was adjusted to 3 by addition of ammonia. An interesting comparisor can then be made as to phase evolution of BaTi,O, between the original Ba/Ti=1/4 CA/EG solution (pH~0.1) and the one modified by ammonia (pH~3). Figure 3 shows XRD patterns of the powder precursor (derived from the ammonia containing solution) calcined in air at different temperatures
Fig.3 X-ray diffraction patterns of the Ba/Ti=1/4 composition powder precursor (derived from an ammonia containing solution (pH=3)) calcined in static air for 2 h at 600 °C (a), 650 °C (b), 700 °C (c), 800 °C (d), 900 °C (e), 1000 °C (f), and 1200 °C (g). One of the common problems in the synthesis of BaTi,O, by wet chemical routes is the difficulty in controlling the rapid hydrolysis of Ti-alkoxides (chemicals most frequently used as sources of Ti) compared with barium species. For instance, in the previously reported sol-gel synthesis of BaTi,O, [18], gels derived from Ba and Ti alkoxide precursors have producec strongly multiphase samples even after the heat-treatment at 1200 °C. The preferential hydrolysis of Ti-alkoxides can form Ti-rich clusters, which destroys the cation composition of the original solution. This result tells us that Ti-alkoxides should be modified by certain organic compounds such as acetic acid to control the degree of hydrolysis and subsequent polycondensation reactions [21, 28-30]. A similar event has been observed in the present PC synthesis of BaTi,O,, when the pH of the original solution was adjusted to 3 by addition of ammonia. An interesting comparisor can then be made as to phase evolution of BaTi,O, between the original Ba/Ti=1/4 CA/EG solution (pH~0.1) and the one modified by ammonia (pH~3). Figure 3 shows XRD patterns of the powder precursor (derived from the ammonia containing solution) calcined in air at different temperatures
Fig.5 X-ray diffraction patterns of the Ba/Ti=1/4 composition powder precursor (derived from the original solution nonmodified by ammonia) calcined in static air for 2 h at 650 °C (a), 700 °C (b) and 800 °C (c).
Fig.5 X-ray diffraction patterns of the Ba/Ti=1/4 composition powder precursor (derived from the original solution nonmodified by ammonia) calcined in static air for 2 h at 650 °C (a), 700 °C (b) and 800 °C (c).
Fig.8 A high-resolution transmission electron microscopic image of PC-BaTi,O,/RuO, (1 wt% Ru relative to BaTi,O,) prepared at 700 °C. A lot of dark spots correspond to RuO, particles. A pure BaTi,O, has been successfully synthesized by heat-treating the PC powder precursors ir air at reduced temperatures (700-900 °C) for 2 h. The success in lowering the processing temperature for BaTi,O, down to 700 °C implies that the molecular-level mixing of Ba and T seems to be likely throughout the PC process. It has been shown that the PC route is < promising methodology for preparing BaTi,O, with relatively large surface areas that can be subsequently used as highly active nanocomposite BaTi,O,/RuO, photocatalysts for wate decomposition.
Fig.8 A high-resolution transmission electron microscopic image of PC-BaTi,O,/RuO, (1 wt% Ru relative to BaTi,O,) prepared at 700 °C. A lot of dark spots correspond to RuO, particles. A pure BaTi,O, has been successfully synthesized by heat-treating the PC powder precursors ir air at reduced temperatures (700-900 °C) for 2 h. The success in lowering the processing temperature for BaTi,O, down to 700 °C implies that the molecular-level mixing of Ba and T seems to be likely throughout the PC process. It has been shown that the PC route is < promising methodology for preparing BaTi,O, with relatively large surface areas that can be subsequently used as highly active nanocomposite BaTi,O,/RuO, photocatalysts for wate decomposition.
Figure 2: Tetragonal ZrO. Evolution During Heat Treatment Te IED IE EEE Figure 6 shows the results of the grain size measurements performed on TEM micrographs. In the calcined samples, all of the compositions showed a grain size of less than ‘ nm. After 24 hours, the 10% and 20% samples had grown the most, and had exceeded the critical grain size for tetragonal ZrO2. Most of the samples with a high percentage of tetragonal ZrO» maintained a grain size of 15-25 nm. However, a few samples had grains as large as 28 nt while maintaining the tetragonal phase. This agrees with the literature values of 15-30 nm as th critical grain size for tetragonal ZrO. Dark field TEM images were used to determine the size « the Al.O3 grains in several samples. The Al,O3 grain size varied widely, from 10-50 nm, with t majority of the grains in the 20-50 nm range. XRD measurements of the grain size proved to be difficult, as the sample quantities were limited, leading to low intensity data and a low degree o confidence in the results.
Figure 2: Tetragonal ZrO. Evolution During Heat Treatment Te IED IE EEE Figure 6 shows the results of the grain size measurements performed on TEM micrographs. In the calcined samples, all of the compositions showed a grain size of less than ‘ nm. After 24 hours, the 10% and 20% samples had grown the most, and had exceeded the critical grain size for tetragonal ZrO2. Most of the samples with a high percentage of tetragonal ZrO» maintained a grain size of 15-25 nm. However, a few samples had grains as large as 28 nt while maintaining the tetragonal phase. This agrees with the literature values of 15-30 nm as th critical grain size for tetragonal ZrO. Dark field TEM images were used to determine the size « the Al.O3 grains in several samples. The Al,O3 grain size varied widely, from 10-50 nm, with t majority of the grains in the 20-50 nm range. XRD measurements of the grain size proved to be difficult, as the sample quantities were limited, leading to low intensity data and a low degree o confidence in the results.
Figure 4: XRD Pattern of 10% Al,O3-90% ZrO, After Heat Treatment at 1100 °C for 120 hour:
Figure 4: XRD Pattern of 10% Al,O3-90% ZrO, After Heat Treatment at 1100 °C for 120 hour:
FIG.3. Surficial micrographic picture of the PZTBS glass-ceramic thick-film annealed at 1100°C for 10 min. The leakage current of the film measured with increasing voltage varying from 1 to 100 V, at 1 V step increments, is shown in Fig. 4. The leakage current increases from around 1 to 100 pA with the applied voltage increasing from 1 to 100 V. The calculated resistivity, based on the measured leakage current and the dimension of the film, is also given in the same figure and the value is around 7x10’? ohm-cm. This resistivity is as high as that of bulk ceramics and is suitable to produce practical ferroelectric devices.
FIG.3. Surficial micrographic picture of the PZTBS glass-ceramic thick-film annealed at 1100°C for 10 min. The leakage current of the film measured with increasing voltage varying from 1 to 100 V, at 1 V step increments, is shown in Fig. 4. The leakage current increases from around 1 to 100 pA with the applied voltage increasing from 1 to 100 V. The calculated resistivity, based on the measured leakage current and the dimension of the film, is also given in the same figure and the value is around 7x10’? ohm-cm. This resistivity is as high as that of bulk ceramics and is suitable to produce practical ferroelectric devices.
FIG. 4. Leakage current and resistivity under increasing voltage for the PZTBS glass-ceramic thick-film annealed at 1100°C for 10 min.
FIG. 4. Leakage current and resistivity under increasing voltage for the PZTBS glass-ceramic thick-film annealed at 1100°C for 10 min.
FIG. 6. The polarization - electric field hysteresis loops of the PZTBS glass-ceramic thick-film annealed at 1100°C for 10 min.
FIG. 6. The polarization - electric field hysteresis loops of the PZTBS glass-ceramic thick-film annealed at 1100°C for 10 min.
Figure 3. Theoretical density versus sintering temperature plot for pure Al,O3 and with MgO and TiO, additives.
Figure 3. Theoretical density versus sintering temperature plot for pure Al,O3 and with MgO and TiO, additives.
Figure 2. Fracture surface of an MgO-doped Al,O3 specimen sintered to 99.9% theoretical density in 10 minutes at 1400 °C.
Figure 2. Fracture surface of an MgO-doped Al,O3 specimen sintered to 99.9% theoretical density in 10 minutes at 1400 °C.
Figure 6. TEM micrograph displaying dislocation generation at the particle interfaces. (bar = 333 nm)
Figure 6. TEM micrograph displaying dislocation generation at the particle interfaces. (bar = 333 nm)
Figure 5: Silica-Encapsulated Silver Clusters Produced Using ACES (These clusters all have diameters smaller than 5 nm. The clusters are stabilized using octyltrimethoxysilane as surfactant.) Size-separation techniques also apply to silica-coated clusters. Figure 5 displays mi- crographs of ultrafine silver-silicon clusters produced in ACES. The clusters have been stabilized by octyltrimethoxysilane dissolved in mesitylene. It is difficult to discern the SiO, coating on such small particles. These particles were formed at an evaporation rate exceeding 1 g/hr. Figure 6 shows micrographs of copper-silica clusters produced using the
Figure 5: Silica-Encapsulated Silver Clusters Produced Using ACES (These clusters all have diameters smaller than 5 nm. The clusters are stabilized using octyltrimethoxysilane as surfactant.) Size-separation techniques also apply to silica-coated clusters. Figure 5 displays mi- crographs of ultrafine silver-silicon clusters produced in ACES. The clusters have been stabilized by octyltrimethoxysilane dissolved in mesitylene. It is difficult to discern the SiO, coating on such small particles. These particles were formed at an evaporation rate exceeding 1 g/hr. Figure 6 shows micrographs of copper-silica clusters produced using the
Figure 6: Silica~-Encapsulated Copper Clusters Produced Using ACES
Figure 6: Silica~-Encapsulated Copper Clusters Produced Using ACES
Figure 1: Volume fraction of gold in particles as a function of time. Points are from optical absorption at 0.44 pum. Line is from Eq. 1 with n = 3/2. Figure 2: Measured optical absorption spectra of small gold particles in glass. Samples irradiated in UV light and heated at 600 °C for time in hours on curves.
Figure 1: Volume fraction of gold in particles as a function of time. Points are from optical absorption at 0.44 pum. Line is from Eq. 1 with n = 3/2. Figure 2: Measured optical absorption spectra of small gold particles in glass. Samples irradiated in UV light and heated at 600 °C for time in hours on curves.
Figure 3: Optical absorption of small gold particles as a function of +. Present work in glass: (O) 12 hr. at 600 C, (C) 68 hr. at 600 C (another melt), absorption displaced for clarity. (A) Ref. 12, in water. In all samples, the particle diameter was about one nm (31 gold atoms). Sa i ee) The 1+ dependence of absorption i in these samples results from the free electron character of the electrons in them. Eq. 9 shows that for a particle with R<</, the effective mean free path, /., is equal to the radius. The small value of 1, leads to a large damping constant, y (Eq. 7) and also of € (Eq. 5). Thus the value of€3 in the denominator of Eq. 3 becomes much larger than (€; + 2n”)’, so for these small clusters:
Figure 3: Optical absorption of small gold particles as a function of +. Present work in glass: (O) 12 hr. at 600 C, (C) 68 hr. at 600 C (another melt), absorption displaced for clarity. (A) Ref. 12, in water. In all samples, the particle diameter was about one nm (31 gold atoms). Sa i ee) The 1+ dependence of absorption i in these samples results from the free electron character of the electrons in them. Eq. 9 shows that for a particle with R<</, the effective mean free path, /., is equal to the radius. The small value of 1, leads to a large damping constant, y (Eq. 7) and also of € (Eq. 5). Thus the value of€3 in the denominator of Eq. 3 becomes much larger than (€; + 2n”)’, so for these small clusters:
Fig. 4. Conversely, the electrical conductivity is increased with a nanoscale SiO, addition. Asymptotic values at long times are lower than for TiO, alone, even with the classical two-phase factor [8] applied here. The chemical diffusivity measurements are shown in Fig. 3 for samples held for days at 1153°C. The plotted data for re-equilibration runs (each 1500 sec in duration) after step changes from 0.02 atm to 0.42 atm pO,. For 1153°C hold times less than 10° sec, each sample with the SiO, nanoscale addition shows markedly slower re-equilibration rates than for the baseline TiO, without SiO2. The largest initial decrements in diffusivity are found for the largest volume fraction additions. The lowest diffusivity values are partly affected by the limit imposed by the re-equilibration measurement run time and the cross-sectional area. The electrical conductivity measurements over the same days of holding at 1153°C are shown in Fig. 4. These conductivities are the based on the conductance values at the end of
Fig. 4. Conversely, the electrical conductivity is increased with a nanoscale SiO, addition. Asymptotic values at long times are lower than for TiO, alone, even with the classical two-phase factor [8] applied here. The chemical diffusivity measurements are shown in Fig. 3 for samples held for days at 1153°C. The plotted data for re-equilibration runs (each 1500 sec in duration) after step changes from 0.02 atm to 0.42 atm pO,. For 1153°C hold times less than 10° sec, each sample with the SiO, nanoscale addition shows markedly slower re-equilibration rates than for the baseline TiO, without SiO2. The largest initial decrements in diffusivity are found for the largest volume fraction additions. The lowest diffusivity values are partly affected by the limit imposed by the re-equilibration measurement run time and the cross-sectional area. The electrical conductivity measurements over the same days of holding at 1153°C are shown in Fig. 4. These conductivities are the based on the conductance values at the end of
Fig. 3. Nanoscale SiO, added to TiO, causes a decrease (arrow) in the interdiffusion coefficient, relative to TiO, alone. The largest drops are for the largest vol.%’s. After 10° sec at 1153°C, the decreases are diminished.
Fig. 3. Nanoscale SiO, added to TiO, causes a decrease (arrow) in the interdiffusion coefficient, relative to TiO, alone. The largest drops are for the largest vol.%’s. After 10° sec at 1153°C, the decreases are diminished.
Fig. 2. SEM micrographs of MgO/BaTiO, composites prepared from (a) 1» m and (b) 300 nm BaTiO, powder.
Fig. 2. SEM micrographs of MgO/BaTiO, composites prepared from (a) 1» m and (b) 300 nm BaTiO, powder.
ig.4 Raman spectra of (a) monolithic BaTiO, ceramics and (b) MgO/10vol% BaTiO, nano- composite prepared from 300 nm BaTiO, powder.
ig.4 Raman spectra of (a) monolithic BaTiO, ceramics and (b) MgO/10vol% BaTiO, nano- composite prepared from 300 nm BaTiO, powder.
Figure 1. TEM: As fired 825°C 8 min with particles smaller than 20 nm.
Figure 1. TEM: As fired 825°C 8 min with particles smaller than 20 nm.
Figure 2. TEM: As fired 875°C 12 min with most particles larger than 20 nm.
Figure 2. TEM: As fired 875°C 12 min with most particles larger than 20 nm.
Figure 1 Schematic diagram showing principle of film stratification to control the propagation of columnar morphology The generalised behaviour summarised in the previous section is modified considerably in stratified heterostructures largely as a result of the disruption in the evolution of colunmar morphology in the film. This disruptive effect is illustrated schematically in figure 1. The realisation of a brick-wall effect is dependant not only on the magnitude of the mismatch in lattice constant but also on the adatom mobility in the individual layers. For example, in the case of a heterostructure made from ZnS and ZnSe, both materials share a cubic zinc-blende lattice with a mismatch in lattice constants of only 5%. This is insufficient to prevent column propagation (as shown in figure 2) largely due to the accommodation of micro-strain effects by twinning. In the case of barium and lead fluoride, the difference in lattice constants is even smaller (4%), but the difference in adatom mobility allows a lath-type morphology to be produced (figure 3). Some of the characteristics of the component dendritic BaF2 are retained, although the high mobility of the PbF2 produces a basic bulding block some 1000 x 50A in dimension. When the extent of lattice mismatch is raised to 15% as in the case of ZnS and BaF, a brick-wall morphology is produced as shown in figure 4. Here the dimensions of each crystallite is close to 100A. The well-defined level of ordering in each crystallite is clearly evident, indicating that the material is not fundamentally amorphous, although selected area diffraction patterns cannot resolve the individual phases. There is no evidence of any porosity and such films are highly stable in the environment.
Figure 1 Schematic diagram showing principle of film stratification to control the propagation of columnar morphology The generalised behaviour summarised in the previous section is modified considerably in stratified heterostructures largely as a result of the disruption in the evolution of colunmar morphology in the film. This disruptive effect is illustrated schematically in figure 1. The realisation of a brick-wall effect is dependant not only on the magnitude of the mismatch in lattice constant but also on the adatom mobility in the individual layers. For example, in the case of a heterostructure made from ZnS and ZnSe, both materials share a cubic zinc-blende lattice with a mismatch in lattice constants of only 5%. This is insufficient to prevent column propagation (as shown in figure 2) largely due to the accommodation of micro-strain effects by twinning. In the case of barium and lead fluoride, the difference in lattice constants is even smaller (4%), but the difference in adatom mobility allows a lath-type morphology to be produced (figure 3). Some of the characteristics of the component dendritic BaF2 are retained, although the high mobility of the PbF2 produces a basic bulding block some 1000 x 50A in dimension. When the extent of lattice mismatch is raised to 15% as in the case of ZnS and BaF, a brick-wall morphology is produced as shown in figure 4. Here the dimensions of each crystallite is close to 100A. The well-defined level of ordering in each crystallite is clearly evident, indicating that the material is not fundamentally amorphous, although selected area diffraction patterns cannot resolve the individual phases. There is no evidence of any porosity and such films are highly stable in the environment.
Figure 3 Cross-sectional transmission electron micrograph of a stratified PbF2/BaF2 film showing the development of a lath-type grain morphology.
Figure 3 Cross-sectional transmission electron micrograph of a stratified PbF2/BaF2 film showing the development of a lath-type grain morphology.
Figure 2 Cross-sectional transmission electron micrograph of stratified a ZnS/ZnSe film highlighting the inability of the lighter contrast ZnS films to prevent propagation of columnar grain morphology of the ZnSe component.
Figure 2 Cross-sectional transmission electron micrograph of stratified a ZnS/ZnSe film highlighting the inability of the lighter contrast ZnS films to prevent propagation of columnar grain morphology of the ZnSe component.
Figure 5a X-ray photoelectron spectra measured for monolayer of BiF3 deposited onto ZnS film. Lower curve as deposited, upper curve measured after 2 days in UHV. Note lack of oxygen uptake as evidence of cleanliness of system.
Figure 5a X-ray photoelectron spectra measured for monolayer of BiF3 deposited onto ZnS film. Lower curve as deposited, upper curve measured after 2 days in UHV. Note lack of oxygen uptake as evidence of cleanliness of system.
Figure 5b Detail of X-ray photoelectron spectra measured for monolayer of BiF3 deposited onto ZnS film. Lower curve as deposited, upper curve measured after 2 days in UHV. Note shift in positioning of Bi 4f lines near 160eV as chemical environment changes from Bi-F to Bi-S. Differences in peak intensity of these lines are due to underlying sulphur 2p peak.
Figure 5b Detail of X-ray photoelectron spectra measured for monolayer of BiF3 deposited onto ZnS film. Lower curve as deposited, upper curve measured after 2 days in UHV. Note shift in positioning of Bi 4f lines near 160eV as chemical environment changes from Bi-F to Bi-S. Differences in peak intensity of these lines are due to underlying sulphur 2p peak.
component forms the regions of small precipitates. The dimensions of the precipitates are very small, typically only a few tens of nanometres, commensurate with the scale of structures produced by the stratification technique. The relationship between the composition of the composite films and relative vapour phase supersaturation of the components during growth is complex, particularly at temperatures significantly above ambient. In general the composition obtained is not simply in proportion to the molecular fluxes of the separate BaF2 and the ZnS components, but rather the presence of the BaF2 has a significant effect on the sticking coefficient of the ZnS. For example, at a Knudsen source temperature of 800°C, a pure ZnS film would deposit at 100°C at a rate of about 0.51m/hr. However once BaF2 is added to the vapour flux, the sticking coefficient of the ZnS is reduced, so that the growth rate of the film becomes controlled more by the BaF2 flux and the resulting film is correspondingly BaF? rich. Significantly higher overpressures of ZnS must be used to achieve compositions above the equimolar. This effect is illustrated in Figure 7.
component forms the regions of small precipitates. The dimensions of the precipitates are very small, typically only a few tens of nanometres, commensurate with the scale of structures produced by the stratification technique. The relationship between the composition of the composite films and relative vapour phase supersaturation of the components during growth is complex, particularly at temperatures significantly above ambient. In general the composition obtained is not simply in proportion to the molecular fluxes of the separate BaF2 and the ZnS components, but rather the presence of the BaF2 has a significant effect on the sticking coefficient of the ZnS. For example, at a Knudsen source temperature of 800°C, a pure ZnS film would deposit at 100°C at a rate of about 0.51m/hr. However once BaF2 is added to the vapour flux, the sticking coefficient of the ZnS is reduced, so that the growth rate of the film becomes controlled more by the BaF2 flux and the resulting film is correspondingly BaF? rich. Significantly higher overpressures of ZnS must be used to achieve compositions above the equimolar. This effect is illustrated in Figure 7.
Figure 8 Variation of refractive index of ZnS/BaF2 nano-composite with mole fraction of film, as deposited at 100°C.
Figure 8 Variation of refractive index of ZnS/BaF2 nano-composite with mole fraction of film, as deposited at 100°C.
The materials engineering approach afforded by nanocomposite materials has been exploited for the fabrication of complex optical filters, particularly those based on inhomogeneous designs. In such structures, reflection bands are produced as a result of interference effects in thin film ensembles with half-wave sinusoidal periodicity in refractive index. The bandwidth of the reflection peak is dependant on the refractive index excursion An around the mean value, whilst the peak reflectivity achieved depends on both An and on the total number of periods present. Multiline designs can be realised by the superimposition of individual sinusoids [13].
The materials engineering approach afforded by nanocomposite materials has been exploited for the fabrication of complex optical filters, particularly those based on inhomogeneous designs. In such structures, reflection bands are produced as a result of interference effects in thin film ensembles with half-wave sinusoidal periodicity in refractive index. The bandwidth of the reflection peak is dependant on the refractive index excursion An around the mean value, whilst the peak reflectivity achieved depends on both An and on the total number of periods present. Multiline designs can be realised by the superimposition of individual sinusoids [13].
the superimposition of the characteristic sinusoids for each reflection band using the inhomogeneou: ZnS/BaF2 nano-composite. 9. ACKOWLEDGMENT
the superimposition of the characteristic sinusoids for each reflection band using the inhomogeneou: ZnS/BaF2 nano-composite. 9. ACKOWLEDGMENT
Fig.5 TEM micrographs of (a) a SiC particle and (b) TiN particles on grain boundaries. Table I Values of some relevant parameters associated with the interface properties. Yo/Vep = Sin®, / sinO,
Fig.5 TEM micrographs of (a) a SiC particle and (b) TiN particles on grain boundaries. Table I Values of some relevant parameters associated with the interface properties. Yo/Vep = Sin®, / sinO,
When combining two materials which form a solid solution, such as TiN and TiC [10], the hardness H follows the rule-of-mixture, i.e. it increases monotonously according to equation (4) (see [10]).
When combining two materials which form a solid solution, such as TiN and TiC [10], the hardness H follows the rule-of-mixture, i.e. it increases monotonously according to equation (4) (see [10]).
creased due to some yet unknown effect. In [9] we have speculated if electronic phenomena related to quantum confinement could be responsible for such effect. However, a detailed theo- retical treatment of problem is still lacking but highly desirable. Fig. 4: A typical HRTEM image of a sample with 33 mole % of SisN, showing a direct image of the (111) and (200) lattice planes of those TiN nanocrystals which have the correct orienta- tion with respect to the electron beam. The defect free nc-TiN crystallites of a fairly regular, almost spherical shape are imbedded in the amorphous a-Si3N, matrix. (See text.)
creased due to some yet unknown effect. In [9] we have speculated if electronic phenomena related to quantum confinement could be responsible for such effect. However, a detailed theo- retical treatment of problem is still lacking but highly desirable. Fig. 4: A typical HRTEM image of a sample with 33 mole % of SisN, showing a direct image of the (111) and (200) lattice planes of those TiN nanocrystals which have the correct orienta- tion with respect to the electron beam. The defect free nc-TiN crystallites of a fairly regular, almost spherical shape are imbedded in the amorphous a-Si3N, matrix. (See text.)
Figure 1: FE-SEM micrograph of fracture surface (a) and bright field TEM micrograph (b) of TiN specimen obtained at T=1200°C and P=4 GPa. Turning to the analysis of Table I it seems to be important to keep in mind that increase of processing temperature has resulted in hardness increase in spite of possible recrystallization effect. It can be connected not only with a small density increase but also with “strengthening” of gtain boundaries as we have pointed out [21] and also noted by Watanabe [22]. It seems very likely that a further reduction of TiB2 powder size and an optimization of high pressure sintering parameters or other high-energy consolidation methods will further increase the hardness level and the foregoing data are not the limiting values.
Figure 1: FE-SEM micrograph of fracture surface (a) and bright field TEM micrograph (b) of TiN specimen obtained at T=1200°C and P=4 GPa. Turning to the analysis of Table I it seems to be important to keep in mind that increase of processing temperature has resulted in hardness increase in spite of possible recrystallization effect. It can be connected not only with a small density increase but also with “strengthening” of gtain boundaries as we have pointed out [21] and also noted by Watanabe [22]. It seems very likely that a further reduction of TiB2 powder size and an optimization of high pressure sintering parameters or other high-energy consolidation methods will further increase the hardness level and the foregoing data are not the limiting values.
"Methods of preparation: DCMNRS - d.c.magnetron non-reactive sputtering; DCMRS - d.c. ma- gnetron reactive sputtering; RFMNRS - r.f.magnetron “Sa sputtering; CAPD - cathodic arc plasma deposition; EBIP - electron beam ion plating. ? A depth of nanoindenter penetration. Table II. Vickers Microhardness (H,) of TiN/TiB2 Films
"Methods of preparation: DCMNRS - d.c.magnetron non-reactive sputtering; DCMRS - d.c. ma- gnetron reactive sputtering; RFMNRS - r.f.magnetron “Sa sputtering; CAPD - cathodic arc plasma deposition; EBIP - electron beam ion plating. ? A depth of nanoindenter penetration. Table II. Vickers Microhardness (H,) of TiN/TiB2 Films
Table I: Film deposition conditions.
Table I: Film deposition conditions.
Table II: Peak shifts due to alloying and calculated lattice constants in 5.0 nm/layer films.
Table II: Peak shifts due to alloying and calculated lattice constants in 5.0 nm/layer films.
Figure 3: XRD patterns of (a) TiC/ZrC and (b) mixed ZrC/VC alloy with 5.0 nm layer periods.
Figure 3: XRD patterns of (a) TiC/ZrC and (b) mixed ZrC/VC alloy with 5.0 nm layer periods.
Transmission electron microscopy was used to investigate the ZrC/VC films with the 0.3 nm and 25 nm nominal layer thicknesses. A TEM plan-view image of a ZrC/VC 0.3 nm film is shown in Figure 5. In this film, an alloy formed with an amorphous matrix, as shown by the first broad ring in the diffraction pattern. Small (~ 7 nm) crystalline particles were observed in the matrix. Crystalline particles of a much larger characteristic size were also observed and are believed to be particulates. Energy-dispersive analysis of the film was also conducted and veri- fied the presence of Zr, V and C, but a specific composition analysis was not conducted. A TEM cross-section image of a ZrC/VC multilayer (25 nm) film is shown in Figure 6. This film was also deposited on Si (111) at 4000C/ 5 mTorr Ar, but with an increased dwell time on each target.
Transmission electron microscopy was used to investigate the ZrC/VC films with the 0.3 nm and 25 nm nominal layer thicknesses. A TEM plan-view image of a ZrC/VC 0.3 nm film is shown in Figure 5. In this film, an alloy formed with an amorphous matrix, as shown by the first broad ring in the diffraction pattern. Small (~ 7 nm) crystalline particles were observed in the matrix. Crystalline particles of a much larger characteristic size were also observed and are believed to be particulates. Energy-dispersive analysis of the film was also conducted and veri- fied the presence of Zr, V and C, but a specific composition analysis was not conducted. A TEM cross-section image of a ZrC/VC multilayer (25 nm) film is shown in Figure 6. This film was also deposited on Si (111) at 4000C/ 5 mTorr Ar, but with an increased dwell time on each target.
B72 UV Deposition of films in this study was carried out using separate targets of each carbide. This suggests that distinct carbide phases should result in the absence of any diffusion or atomic mixing processes. The deposition temperature of 400 C represents a homologous temperature of 0.2Tm for TiC, with similar values for ZrC and VC. Consideration of the diffusion of C in Ti (D ~ 10-13 cm?/s [5]) gives a diffusion distance of 144 nm after 1800 seconds. Therefore, the depo- sition temperature does not allow for significant long-range diffusion. However, the PLD process produces a significant fraction of highly energetic depositing species [6] of energies on the order of 200-400 eV. The high energies of these species can enhance atomic mixing and surface diffu- sion processes at the depositing film, which contrasts with other deposition processes where most of the depositing species are at thermal energies. eee: ame amen eA ge9 - oe . -—=—=. se ~~. ows oe»
B72 UV Deposition of films in this study was carried out using separate targets of each carbide. This suggests that distinct carbide phases should result in the absence of any diffusion or atomic mixing processes. The deposition temperature of 400 C represents a homologous temperature of 0.2Tm for TiC, with similar values for ZrC and VC. Consideration of the diffusion of C in Ti (D ~ 10-13 cm?/s [5]) gives a diffusion distance of 144 nm after 1800 seconds. Therefore, the depo- sition temperature does not allow for significant long-range diffusion. However, the PLD process produces a significant fraction of highly energetic depositing species [6] of energies on the order of 200-400 eV. The high energies of these species can enhance atomic mixing and surface diffu- sion processes at the depositing film, which contrasts with other deposition processes where most of the depositing species are at thermal energies. eee: ame amen eA ge9 - oe . -—=—=. se ~~. ows oe»
Fig. 2 TEM photograph of as-deposited Pt/ TiO? nanocomposite film with a Pt/Ti atomic ratio of 0.18. Figure 1 shows the Pt/Ti atomic ratio in the deposited nanocomposite films as a function of Pt wire length placed on the TiO2 target. The Pt/Ti atomic ratio was calculated from Ti2p and Pt4f peak areas in X-ray photoelectron spectra of nanocomposite films. The Pt/Ti atomic ratio in the deposited nanocomposite films increased with an increase in Pt wire length, suggesting that the content of Pt in the deposited Pt/TiO2 nanocomposite can be easily controlled by the amount of Pt wire placed on the TiO2 target. The Pt/Ti atomic ratio of 0.18 corresponds to 30 wt% of Pt in the
Fig. 2 TEM photograph of as-deposited Pt/ TiO? nanocomposite film with a Pt/Ti atomic ratio of 0.18. Figure 1 shows the Pt/Ti atomic ratio in the deposited nanocomposite films as a function of Pt wire length placed on the TiO2 target. The Pt/Ti atomic ratio was calculated from Ti2p and Pt4f peak areas in X-ray photoelectron spectra of nanocomposite films. The Pt/Ti atomic ratio in the deposited nanocomposite films increased with an increase in Pt wire length, suggesting that the content of Pt in the deposited Pt/TiO2 nanocomposite can be easily controlled by the amount of Pt wire placed on the TiO2 target. The Pt/Ti atomic ratio of 0.18 corresponds to 30 wt% of Pt in the
Fig. 1 Pt/Ti atomic ratio in Pt/TiO2 nanocomposite films as a function of Pt wire length on sputtering target of TiO.
Fig. 1 Pt/Ti atomic ratio in Pt/TiO2 nanocomposite films as a function of Pt wire length on sputtering target of TiO.
TableI XPS Pt4f7/2 binding energies for various Pt species. films with Pt/Ti atomic ratios of 0.35 and 0.18 were 71.1 eV and 74.8 eV, respectively. These data indicate that the chemical states of Pt nanoparticles in the nanocomposite films with Pt/Ti atomic ratio of 0.35 and 0.18 were almost always Pt metal and PtO2, respectively. In contrast, Pt4f7/2 binding energy of the nanocomposite film with Pt/Ti atomic ratio of 0.21 was 72.4 eV, which is close not to Pt metal, PtO and PtO2 but to Pt-O-Ti, indicating that some complex oxide species (Pt- O-Ti) is present in the nanocomposite film. Ti2p3/2 binding energy in as-deposited nanocomposite films was observed in an energy range from 458.1 to 458.6 eV which is very close to the Ti2p3/2
TableI XPS Pt4f7/2 binding energies for various Pt species. films with Pt/Ti atomic ratios of 0.35 and 0.18 were 71.1 eV and 74.8 eV, respectively. These data indicate that the chemical states of Pt nanoparticles in the nanocomposite films with Pt/Ti atomic ratio of 0.35 and 0.18 were almost always Pt metal and PtO2, respectively. In contrast, Pt4f7/2 binding energy of the nanocomposite film with Pt/Ti atomic ratio of 0.21 was 72.4 eV, which is close not to Pt metal, PtO and PtO2 but to Pt-O-Ti, indicating that some complex oxide species (Pt- O-Ti) is present in the nanocomposite film. Ti2p3/2 binding energy in as-deposited nanocomposite films was observed in an energy range from 458.1 to 458.6 eV which is very close to the Ti2p3/2
Fig. 3 XPS spectra of Pt4f level in as-deposited Pt/TiO2 nanocomposite films with various Pt/Ti atomic ratios; (a)0.35, (b)0.21, (c) 0.18. Sea aera RE OTN NE TNC CAMELS NAY Fa NER TORTI ON AAT? CENTS UTM ANON V EN Serve eg STANTS ETE TE Sere tinte Me Mme niin! ier Figure 3 shows XPS spectra of the Pt4fsy2 : and Pt4f7/2 levels i in as-deposited Pt/Ti02 nanocomposite films. These peak positions in binding energy apparently changed among the nanocomposite films with different P/Ti atomic ratios. Table I shows XPS Pt4f7/2 binding energies for several Pt species [11]. As shown in Fig. 3, Pt4f7/2 binding energies of the nanocomposite
Fig. 3 XPS spectra of Pt4f level in as-deposited Pt/TiO2 nanocomposite films with various Pt/Ti atomic ratios; (a)0.35, (b)0.21, (c) 0.18. Sea aera RE OTN NE TNC CAMELS NAY Fa NER TORTI ON AAT? CENTS UTM ANON V EN Serve eg STANTS ETE TE Sere tinte Me Mme niin! ier Figure 3 shows XPS spectra of the Pt4fsy2 : and Pt4f7/2 levels i in as-deposited Pt/Ti02 nanocomposite films. These peak positions in binding energy apparently changed among the nanocomposite films with different P/Ti atomic ratios. Table I shows XPS Pt4f7/2 binding energies for several Pt species [11]. As shown in Fig. 3, Pt4f7/2 binding energies of the nanocomposite
binding energy of TiO2 (458.7 eV) [12], indicating that the chemical state of Ti in as-deposited films was almost always TiO? with a slight reduction during the sputter deposition process in argon atmosphere. Fig. 4 XRD patterns of heated TiO (a) and Pt/TiO2 (b-e) nanocomposite films at various temperatures in air; (a), (b) 700 °C, (c) 600 °C, (d) 500 °C, (e) 400 °C.
binding energy of TiO2 (458.7 eV) [12], indicating that the chemical state of Ti in as-deposited films was almost always TiO? with a slight reduction during the sputter deposition process in argon atmosphere. Fig. 4 XRD patterns of heated TiO (a) and Pt/TiO2 (b-e) nanocomposite films at various temperatures in air; (a), (b) 700 °C, (c) 600 °C, (d) 500 °C, (e) 400 °C.
Fig. 5 TEM photographs of heated Pt/TiO2 nanocomposite films with a Pt/Ti atomic ratio of 0.18 at (a) 500°C and (b) 600°C. temperatures. It is well known that PtO2 decomposes into Pt metal at 585 °C [13]. Thus, the Pt metal particles rapidly grow via decomposition of PtO2 at temperatures higher than 600 °C. In heated nanocomposite films with a Pt/Ti atomic ratio of 0.35 at 600 °C, the nanoparticle size ranged from 10 to 100 nm. The size of these Pt nanoparticles in the heated nanocomposite films increased with an increase in the Pt/Ti atomic ratio of the films. PTT Nkok ren rare ichen be, Se laa ceca: fee ce tain lee Snancnwa nee Pase se Sis ae i aa i a BK. i ae ie ee eS
Fig. 5 TEM photographs of heated Pt/TiO2 nanocomposite films with a Pt/Ti atomic ratio of 0.18 at (a) 500°C and (b) 600°C. temperatures. It is well known that PtO2 decomposes into Pt metal at 585 °C [13]. Thus, the Pt metal particles rapidly grow via decomposition of PtO2 at temperatures higher than 600 °C. In heated nanocomposite films with a Pt/Ti atomic ratio of 0.35 at 600 °C, the nanoparticle size ranged from 10 to 100 nm. The size of these Pt nanoparticles in the heated nanocomposite films increased with an increase in the Pt/Ti atomic ratio of the films. PTT Nkok ren rare ichen be, Se laa ceca: fee ce tain lee Snancnwa nee Pase se Sis ae i aa i a BK. i ae ie ee eS
nm pores the contribution trom the first term 1s much more visible. Lhe fitting parameters corresponding to curves (2), (3) and (4) in Fig. 1 are: (2) - z=2.3, t2=0.04 ms (exponential decay is neglected); (3) - 71=0.07 ms, z=3.6, 79==3s; (4) - 713=0.06 ms, z=4, T2=23s. We found that the relaxation time of the slow process for 5CB in 10 nm pores strongly increases when temperature decreases from 300 K up to 270 K varying from 1.7 x 1074s to 14 s. The data analysis shows that the temperature dependence of the relaxation times for 5CB in 10 nm random pores, in the temperature interval (270-300) K, follows the Vogel-Fulcher law: + = Toexp(B/(T ~ To)) with parameters: 7) = 1.4-107l's, B = 847K and Ty = 246K. The dynamical behavior of 5CB confined in cylindrical and large random pores (100 nm) was closer to the bulk behavior as we expected. In these matrices the decay due to director fluctuations dominates. However a slow decay, origin of which is under question appears in the anisotropic phase of LC (Fig.1).
nm pores the contribution trom the first term 1s much more visible. Lhe fitting parameters corresponding to curves (2), (3) and (4) in Fig. 1 are: (2) - z=2.3, t2=0.04 ms (exponential decay is neglected); (3) - 71=0.07 ms, z=3.6, 79==3s; (4) - 713=0.06 ms, z=4, T2=23s. We found that the relaxation time of the slow process for 5CB in 10 nm pores strongly increases when temperature decreases from 300 K up to 270 K varying from 1.7 x 1074s to 14 s. The data analysis shows that the temperature dependence of the relaxation times for 5CB in 10 nm random pores, in the temperature interval (270-300) K, follows the Vogel-Fulcher law: + = Toexp(B/(T ~ To)) with parameters: 7) = 1.4-107l's, B = 847K and Ty = 246K. The dynamical behavior of 5CB confined in cylindrical and large random pores (100 nm) was closer to the bulk behavior as we expected. In these matrices the decay due to director fluctuations dominates. However a slow decay, origin of which is under question appears in the anisotropic phase of LC (Fig.1).
For bulk 5CB in the nematic phase one relaxational process associated with the molecular rotation around it’s short axes was observed [9]. The characteristic frequency of this process is in MHz range. The temperature dependence of the corresponding relaxation times obeys the Arrhenius formula: t = toexp(U/kT), where U is the activation energy and k is the Boltzmann constant. In bulk nematic phase the value of U; is 0.61 eV. The dielectric behavior of 5CB dispersed in porous matrices is different from it’s bulk behavior. In confined 5CB we observe at least three identified relaxational processes. The first low frequency process (1 Hz - 10 KHz) which does not exist in the bulk is illustrated in Fig. 3 for 5CB in 10 nm pores (curves 1 and 1a) and in 100 nm pores (curves 2 and 2a). The second process is very clearly seen in the MHz frequency range and the last one, which is less visible appears in the frequency range f >30MHz. Both processes are shown in Fig. 3, curves 1b, 1c (10 nm pores) and 2b, 2c (100 nm pores). The solid lines in Fig.3 represent the results
For bulk 5CB in the nematic phase one relaxational process associated with the molecular rotation around it’s short axes was observed [9]. The characteristic frequency of this process is in MHz range. The temperature dependence of the corresponding relaxation times obeys the Arrhenius formula: t = toexp(U/kT), where U is the activation energy and k is the Boltzmann constant. In bulk nematic phase the value of U; is 0.61 eV. The dielectric behavior of 5CB dispersed in porous matrices is different from it’s bulk behavior. In confined 5CB we observe at least three identified relaxational processes. The first low frequency process (1 Hz - 10 KHz) which does not exist in the bulk is illustrated in Fig. 3 for 5CB in 10 nm pores (curves 1 and 1a) and in 100 nm pores (curves 2 and 2a). The second process is very clearly seen in the MHz frequency range and the last one, which is less visible appears in the frequency range f >30MHz. Both processes are shown in Fig. 3, curves 1b, 1c (10 nm pores) and 2b, 2c (100 nm pores). The solid lines in Fig.3 represent the results
- The relaxational process in the MHz range with 7 ~ 1078s is bulk-like and corresponds to the rotation of the molecule around the short axis. In cylindrical pores at frequencies f >1MHz the main contribution to observed dielectric relaxation is due to molecular rotation around short axis, and the process with tT ~ 1079s was less visible [7] than in random pores. The temperature ‘dependence of relaxation times corresponding to the rotation of molecules around short axis for 5CB in 10 nm random pores is presented in Fig. 4. This dependence for 5CB in pores is different from that in the bulk nematic phase. From Fig. .4 we see that ther is no indication of a sharp or well identified nematic-isotropic phase transition. Instead w observe a gradual change of relaxation times in a wide temperature range 36.5 °C< T <30.( °C. The temperature range T' < 36.5°C, where the temperature dependence of Int deviate: from linear dependence, corresponds to the anisotropic phase and Inv is not a linear functio1 of 1/T in this range. However if we consider the temperature regions T < 30.0°C anc T > 36.5°C separately then Inr = f(1/T) in these regions is reasonably well approximatec by a linear function and the corresponding activation energies are U = 0.54eV and Uj, = 0.48eV. The first activation energy U is less than the activation energy of bulk nemati phase U, = 0.6leV. This fact could be considered as an evidence for smectic type orde formation at pore wall - LC interface in this temperature range. Qualitatively for 5CB it cylindrical pores the temperature dependence of relaxation times corresponding to molecula: rotation around short axis is close to that in 10 nm pores. For 5CB in 100 nm random pore: we found that the temperature range (T < 34.5°C) corresponds to the anisotropic phase of 5CB and Inv is not a linear function of 1/T. Again if we consider the temperatur regions 34.5°C < T < 20°C and 19.5°C < T'< 9°C separately then Int = f(1/T) it each of these regions could be reasonably well approximated by a linear function and th corresponding activation energies were found to be U; = 0.74eV and U2 = 0.53eV. The firs!
- The relaxational process in the MHz range with 7 ~ 1078s is bulk-like and corresponds to the rotation of the molecule around the short axis. In cylindrical pores at frequencies f >1MHz the main contribution to observed dielectric relaxation is due to molecular rotation around short axis, and the process with tT ~ 1079s was less visible [7] than in random pores. The temperature ‘dependence of relaxation times corresponding to the rotation of molecules around short axis for 5CB in 10 nm random pores is presented in Fig. 4. This dependence for 5CB in pores is different from that in the bulk nematic phase. From Fig. .4 we see that ther is no indication of a sharp or well identified nematic-isotropic phase transition. Instead w observe a gradual change of relaxation times in a wide temperature range 36.5 °C< T <30.( °C. The temperature range T' < 36.5°C, where the temperature dependence of Int deviate: from linear dependence, corresponds to the anisotropic phase and Inv is not a linear functio1 of 1/T in this range. However if we consider the temperature regions T < 30.0°C anc T > 36.5°C separately then Inr = f(1/T) in these regions is reasonably well approximatec by a linear function and the corresponding activation energies are U = 0.54eV and Uj, = 0.48eV. The first activation energy U is less than the activation energy of bulk nemati phase U, = 0.6leV. This fact could be considered as an evidence for smectic type orde formation at pore wall - LC interface in this temperature range. Qualitatively for 5CB it cylindrical pores the temperature dependence of relaxation times corresponding to molecula: rotation around short axis is close to that in 10 nm pores. For 5CB in 100 nm random pore: we found that the temperature range (T < 34.5°C) corresponds to the anisotropic phase of 5CB and Inv is not a linear function of 1/T. Again if we consider the temperatur regions 34.5°C < T < 20°C and 19.5°C < T'< 9°C separately then Int = f(1/T) it each of these regions could be reasonably well approximated by a linear function and th corresponding activation energies were found to be U; = 0.74eV and U2 = 0.53eV. The firs!
Figure 2: Number average particle size distribution obtained by Dynamic Light Scattering on centrifuged Si(4h) and Si(Sh) samples. ‘igure 1: Bright-field Transmission Electron Micrographs of Si(4h) (A) as-milled, and (B) after separation by centrifugation.
Figure 2: Number average particle size distribution obtained by Dynamic Light Scattering on centrifuged Si(4h) and Si(Sh) samples. ‘igure 1: Bright-field Transmission Electron Micrographs of Si(4h) (A) as-milled, and (B) after separation by centrifugation.
Figure 3: Radial X-ray diffraction profiles of (A) bulk Si, (B) as-milled sample Si(Sh), and (C Si(Sh) after separation by centrifugation.
Figure 3: Radial X-ray diffraction profiles of (A) bulk Si, (B) as-milled sample Si(Sh), and (C Si(Sh) after separation by centrifugation.
Figure 4: UV-Vis absorbance spectra of different Si samples. The UV-Vis spectrum of the buik Si, milled Si powder, and nanoparticles separated by centrifugation are shown in Figure 4. The bulk Si and as-milled powder (both black in color, show an essentially featureless spectra, with absorbance increasing over the wavelength range The centrifuged nanoparticles (with dimensions of 21 and 38 nm) depict spectra unlike those observed from the under 6 nm chemically synthesized nanoparticles.[7,8,13] A sharp peal between 250 and 300 nm is observed followed by a gradual tailing of the absorbance. The peak below 250 nm appears to be an artifact based on the absorption of the solvent (ethanol) in tha! range. The absorbance from the smaller particle dimension [Si(Sh)] is blue shifted. In addition, vibronic signature is visible, especially for the Si(Sh) separated sample. We are currently investigating whether these effects are due to impurities introduced during the nanomilling process phenomenon associated with the Mie optical extinction,{1] or other extraneous effects.
Figure 4: UV-Vis absorbance spectra of different Si samples. The UV-Vis spectrum of the buik Si, milled Si powder, and nanoparticles separated by centrifugation are shown in Figure 4. The bulk Si and as-milled powder (both black in color, show an essentially featureless spectra, with absorbance increasing over the wavelength range The centrifuged nanoparticles (with dimensions of 21 and 38 nm) depict spectra unlike those observed from the under 6 nm chemically synthesized nanoparticles.[7,8,13] A sharp peal between 250 and 300 nm is observed followed by a gradual tailing of the absorbance. The peak below 250 nm appears to be an artifact based on the absorption of the solvent (ethanol) in tha! range. The absorbance from the smaller particle dimension [Si(Sh)] is blue shifted. In addition, vibronic signature is visible, especially for the Si(Sh) separated sample. We are currently investigating whether these effects are due to impurities introduced during the nanomilling process phenomenon associated with the Mie optical extinction,{1] or other extraneous effects.
Figure 5: Index of refraction of a roughly 50/50 (w/w) gelatin/Si(Sh) nanoparticles film prepared by spin coating. Insert compares the nanocomposite with bulk crystalline Si and gelatin, respectively.
Figure 5: Index of refraction of a roughly 50/50 (w/w) gelatin/Si(Sh) nanoparticles film prepared by spin coating. Insert compares the nanocomposite with bulk crystalline Si and gelatin, respectively.
Spectral reflectance FTIR spectra done of 90/10 TIBE/polymer composite films processed in 1.0 M Ba(OH)2 at 80° C for 60 min on a Ag-coated glass substrate. Figure 2 Similar film morphologies were observed using TEM (Fig. 4). Again, the polyme and BaTiO; could be differentiated in the 50/50 TIBE/polymer film whereas the polyme was not evident in the 90/10 TIBE/polymer film. Carbonate formation was not observed within the films, only on the surface. Figure 5 is a magnified view of a BaTiO3 partick and the corresponding electron diffraction pattern from the film shown in Figure 4b The bright filed image and electron diffraction pattern suggest that the BaTiO; cluster: are composed of primary particles (5-10 nm) as observed in the SEM.
Spectral reflectance FTIR spectra done of 90/10 TIBE/polymer composite films processed in 1.0 M Ba(OH)2 at 80° C for 60 min on a Ag-coated glass substrate. Figure 2 Similar film morphologies were observed using TEM (Fig. 4). Again, the polyme and BaTiO; could be differentiated in the 50/50 TIBE/polymer film whereas the polyme was not evident in the 90/10 TIBE/polymer film. Carbonate formation was not observed within the films, only on the surface. Figure 5 is a magnified view of a BaTiO3 partick and the corresponding electron diffraction pattern from the film shown in Figure 4b The bright filed image and electron diffraction pattern suggest that the BaTiO; cluster: are composed of primary particles (5-10 nm) as observed in the SEM.
XRD pattern of 90/10 TIBE/polymer composite film processed in 1.0 M Ba(OH), at 80° C for 60 min on a glass substrate.
XRD pattern of 90/10 TIBE/polymer composite film processed in 1.0 M Ba(OH), at 80° C for 60 min on a glass substrate.
TEM micrographs of 50/50 and 90/10 TIBE/polymer composite film processed in 1.0 M Ba(OH), at 800° C 60 min.
TEM micrographs of 50/50 and 90/10 TIBE/polymer composite film processed in 1.0 M Ba(OH), at 800° C 60 min.
SEM micrographs of a) 25/75 b) 50/50 and c) 90/10 TIBE/polymer composite films 1.0 M Ba(OH)2 at 80 °C for 60 min.
SEM micrographs of a) 25/75 b) 50/50 and c) 90/10 TIBE/polymer composite films 1.0 M Ba(OH)2 at 80 °C for 60 min.
a) Magnification and b) electron diffraction pattern of a BaTiO3 cluster in a 90/10 TIBE/polymer composite film processed in 1.0 M Ba(OH)2 at 80° C for 60 min.
a) Magnification and b) electron diffraction pattern of a BaTiO3 cluster in a 90/10 TIBE/polymer composite film processed in 1.0 M Ba(OH)2 at 80° C for 60 min.
Dielectric constant for 25/75, 50/50, 90/10 TIBE/polymer composite films processed in 1.0 M Ba(OH)2 at 80° C for up to 6 h. Composite BaTiO3/polymer thin films were hydrothermally processed at 80° C from TIBE-polybutadiene/polystyrene block copolymer precursor solutions. The films were composed of BaTiO3 clusters within a polymer. FTIR spectroscopy confirmed the presence of the block copolymer in all of the composite films. The volume fraction of
Dielectric constant for 25/75, 50/50, 90/10 TIBE/polymer composite films processed in 1.0 M Ba(OH)2 at 80° C for up to 6 h. Composite BaTiO3/polymer thin films were hydrothermally processed at 80° C from TIBE-polybutadiene/polystyrene block copolymer precursor solutions. The films were composed of BaTiO3 clusters within a polymer. FTIR spectroscopy confirmed the presence of the block copolymer in all of the composite films. The volume fraction of
Fig. 2 : Scanning backscattered electron micrographs of cobalt-Al,03;2SiO2 composites prepared by method I (a,a'), II (b) and III (c). Materials have been thermally treated at 1100°C under Hp. 1:4:2 matrix with addition of 35% molar of B20; is used (bars : scale in micron).
Fig. 2 : Scanning backscattered electron micrographs of cobalt-Al,03;2SiO2 composites prepared by method I (a,a'), II (b) and III (c). Materials have been thermally treated at 1100°C under Hp. 1:4:2 matrix with addition of 35% molar of B20; is used (bars : scale in micron).
Fig. 4 : Micro-Raman spectra (A=514.5nm spectra) of iron nitrate impregnated aluminosilicate (1:4:2) glass : (a) brown scales, (b) colourless core of monoliths. The used laser power is given. Stars indicate the contribution of the lenses used for collecting the scattered light.
Fig. 4 : Micro-Raman spectra (A=514.5nm spectra) of iron nitrate impregnated aluminosilicate (1:4:2) glass : (a) brown scales, (b) colourless core of monoliths. The used laser power is given. Stars indicate the contribution of the lenses used for collecting the scattered light.
Ee ee x ee eee Fig. 3 : Scheme of the ions diffusion/adsorption mechanism at the pore opening (a) and of the porous structure of the impregnated host matrix before (b) and after (c) thermal treatment. The size of basic polymeric moities (schematic drawn as circle) is close to 1-3nm.
Ee ee x ee eee Fig. 3 : Scheme of the ions diffusion/adsorption mechanism at the pore opening (a) and of the porous structure of the impregnated host matrix before (b) and after (c) thermal treatment. The size of basic polymeric moities (schematic drawn as circle) is close to 1-3nm.
Fig. 5 : Transmission electron photomicrograph of metal-aluminosilicate composites prepared by thermal treatment at 800°C under hydrogen : (a) iron precipitates (dark field), (b,c) cobalt precipitates.
Fig. 5 : Transmission electron photomicrograph of metal-aluminosilicate composites prepared by thermal treatment at 800°C under hydrogen : (a) iron precipitates (dark field), (b,c) cobalt precipitates.
Fig. 6 : Evolution with frequency (0.1 to 10GHz) of real (e') and imaginary (e") permittivity and real (u') and imaginary (u") permeability for a thick film (thickness 2mm) made of a dispersion of cobalt nanoparticles (8% in volume) in an aluminoborosilicate glass (1:4:2 with addition of 35% molar B2Os).
Fig. 6 : Evolution with frequency (0.1 to 10GHz) of real (e') and imaginary (e") permittivity and real (u') and imaginary (u") permeability for a thick film (thickness 2mm) made of a dispersion of cobalt nanoparticles (8% in volume) in an aluminoborosilicate glass (1:4:2 with addition of 35% molar B2Os).
Fig. 1. Water sorption isobars for the SWS-1L material.
Fig. 1. Water sorption isobars for the SWS-1L material.
Fig.2. Universal water sorption isotherm for the CaCl.-H,0 system confined to micro- and mesopores.
Fig.2. Universal water sorption isotherm for the CaCl.-H,0 system confined to micro- and mesopores.
Table I. Isosteric heat of water sorption AHj, for CaCk - H,O system confined to mesopores well as a shoulder at Nx 4. At Py0/Po > 0.15-0.2 the isotherm is a smooth curve which essentially coincides with the proper curve for the bulk aqueous solutions of calcium chloride [7]. It means that the solution confinement to mesoporous silica gel does not change its water sorption properties as compared to the bulk solution. Our DSC measurements (see below) of the specific heat of CaCl, aqueous solutions dispersed in the mesoporous matrix also have not established any changes in the C, value due to the confinement to the pores.
Table I. Isosteric heat of water sorption AHj, for CaCk - H,O system confined to mesopores well as a shoulder at Nx 4. At Py0/Po > 0.15-0.2 the isotherm is a smooth curve which essentially coincides with the proper curve for the bulk aqueous solutions of calcium chloride [7]. It means that the solution confinement to mesoporous silica gel does not change its water sorption properties as compared to the bulk solution. Our DSC measurements (see below) of the specific heat of CaCl, aqueous solutions dispersed in the mesoporous matrix also have not established any changes in the C, value due to the confinement to the pores.
Fig.5. Water sorption isobars for the SWS-1S material.
Fig.5. Water sorption isobars for the SWS-1S material.
Fig. 1. Absorption spectra of Ag/PMMA prepared by various [Ag]o concentrations. ([BPO]p = 2.0 x 102 M, heat-treated at 120 °C for 24 h.) Figure 1 shows the absorption spectra of Ag/PMMA that were prepared by various concentrations of [Ag]o with [BPO]p = 2.0 x 10-2 M, heat-treated at 120 °C for 24h. As expected, an absorption peak was observed around 420 nm. The absorption band grows depending on the [Ag]o and shifts slightly to shorter-wavelength side.
Fig. 1. Absorption spectra of Ag/PMMA prepared by various [Ag]o concentrations. ([BPO]p = 2.0 x 102 M, heat-treated at 120 °C for 24 h.) Figure 1 shows the absorption spectra of Ag/PMMA that were prepared by various concentrations of [Ag]o with [BPO]p = 2.0 x 10-2 M, heat-treated at 120 °C for 24h. As expected, an absorption peak was observed around 420 nm. The absorption band grows depending on the [Ag]o and shifts slightly to shorter-wavelength side.
The formation and growth of the Ag particles were reconfirmed by TEM measurements. Figure 3 is a typical TEM photograph of the yellow colored Ag/PMMA sample, which was made by the following condition; [Ag]o = 2.8 x 10-2 M, [BPO]p = 2.0 x 10-2 M, and heat-treated at 120 Fig. 3. Transmission electron micrograph of a typical Ag/PMMA sample. ([Ag]o = 2.8 x 10° M, [BPO]p = 2.0 x 102 M, heat-treated at 120 °C for 24 h.)
The formation and growth of the Ag particles were reconfirmed by TEM measurements. Figure 3 is a typical TEM photograph of the yellow colored Ag/PMMA sample, which was made by the following condition; [Ag]o = 2.8 x 10-2 M, [BPO]p = 2.0 x 10-2 M, and heat-treated at 120 Fig. 3. Transmission electron micrograph of a typical Ag/PMMA sample. ([Ag]o = 2.8 x 10° M, [BPO]p = 2.0 x 102 M, heat-treated at 120 °C for 24 h.)
Fig. 2. Absorption spectra of Ag/PMMA prepared by various heat-treatment durations at 120 °C. ([Ag]o = 2.8 x 10-2 M, [BPO]p = 2.0 x 102 M)
Fig. 2. Absorption spectra of Ag/PMMA prepared by various heat-treatment durations at 120 °C. ([Ag]o = 2.8 x 10-2 M, [BPO]p = 2.0 x 102 M)
Fig. 4. Absorption spectra of Ag/PMMA prepared by various [BPO]o concentrations. ([Ag]o = 2.8 x 10-2 M, heat-treated at 120 °C for 24 h.)
Fig. 4. Absorption spectra of Ag/PMMA prepared by various [BPO]o concentrations. ([Ag]o = 2.8 x 10-2 M, heat-treated at 120 °C for 24 h.)
Fig. 5. Absorption spectra of cast film of Ag/PMMA: (a) prepared from Ag/PMMA sample; (b) prepared by dissolving AgCF3CO2 in PMMA. ([Ag]o = 2.8 x 10-2 M, [BPO]po = 2.0 x 102 M, heat-treated at 120 °C for 24 h.)
Fig. 5. Absorption spectra of cast film of Ag/PMMA: (a) prepared from Ag/PMMA sample; (b) prepared by dissolving AgCF3CO2 in PMMA. ([Ag]o = 2.8 x 10-2 M, [BPO]po = 2.0 x 102 M, heat-treated at 120 °C for 24 h.)
Figure 1. Detailed LFM image of a PE/PS/XE (45/55/1) sample. PE is brighter and PS darker; black spots in PE phase correspond to CB aggregates. These systems can be characterized in direct space at the nanometer scale by usin Lateral Force Microscopy (or Friction Force Microscopy) [11]. Since their friction coefficient p are different, it is possible to distinguish PE, PS, and CB particles (Figure 1). It appears fro: the collected images that PE is characterized by a higher value of p, which translates on th image to a lighter color on the gray-scale; CB appears in black, since there is almost n interaction between the tip and the CB particles; the 4: value for PS is intermediate and P therefore appears in dark gray on the image. The black spots attributed to CB particles ar observed to be dispersed only in the PE phase of a PE/PS polymer blend. This indicates that these experimental conditions (45% PE, 55% PS, and 1% XE), CB prefers to localize in the P’ phase. For this system, the percolation threshold is about 2.9%. In the case of BP-type CB, th CB particles then localize exclusively in the PS phase and the percolation threshold is abou 1n Qo4
Figure 1. Detailed LFM image of a PE/PS/XE (45/55/1) sample. PE is brighter and PS darker; black spots in PE phase correspond to CB aggregates. These systems can be characterized in direct space at the nanometer scale by usin Lateral Force Microscopy (or Friction Force Microscopy) [11]. Since their friction coefficient p are different, it is possible to distinguish PE, PS, and CB particles (Figure 1). It appears fro: the collected images that PE is characterized by a higher value of p, which translates on th image to a lighter color on the gray-scale; CB appears in black, since there is almost n interaction between the tip and the CB particles; the 4: value for PS is intermediate and P therefore appears in dark gray on the image. The black spots attributed to CB particles ar observed to be dispersed only in the PE phase of a PE/PS polymer blend. This indicates that these experimental conditions (45% PE, 55% PS, and 1% XE), CB prefers to localize in the P’ phase. For this system, the percolation threshold is about 2.9%. In the case of BP-type CB, th CB particles then localize exclusively in the PS phase and the percolation threshold is abou 1n Qo4
migrate slowly from the first polymer to the more “attractive” one. By stopping the mixing at regular time intervals and recording an LFM image, we can determine the optimal time when the CB particles are mainly located at the PE/PS interface. Figure 2 gives the corresponding LFM image for the optimized mixing time for the system (45% PE, 55% PS and 1% of BP- CB): there clearly appears to be a CB layer between PE and PS. The thickness of this layer is about 100 nm. For this system, the percolation threshold value is as low as 0.60%.
migrate slowly from the first polymer to the more “attractive” one. By stopping the mixing at regular time intervals and recording an LFM image, we can determine the optimal time when the CB particles are mainly located at the PE/PS interface. Figure 2 gives the corresponding LFM image for the optimized mixing time for the system (45% PE, 55% PS and 1% of BP- CB): there clearly appears to be a CB layer between PE and PS. The thickness of this layer is about 100 nm. For this system, the percolation threshold value is as low as 0.60%.
Figure 3. LFM image of a PE/PS/CB polymer blend for which the CB particle pH is intermediate. case, a CB layer about 120 nm thick (appearing in black on the LFM image) is clearly seen a the PE/PS interface. The percolation threshold in this system is 0.46%, a remarkably low value.
Figure 3. LFM image of a PE/PS/CB polymer blend for which the CB particle pH is intermediate. case, a CB layer about 120 nm thick (appearing in black on the LFM image) is clearly seen a the PE/PS interface. The percolation threshold in this system is 0.46%, a remarkably low value.
EXPERIMENTAL
EXPERIMENTAL
Figure 2. Surface areas of the aerogel- and xerogel-derived materials at various calcination temperatures (for 5 hours). ‘igure 3. Pore size distributions of aerogel-derived and xerogel-derived materials at various calcination temperatures (for 5 hours).
Figure 2. Surface areas of the aerogel- and xerogel-derived materials at various calcination temperatures (for 5 hours). ‘igure 3. Pore size distributions of aerogel-derived and xerogel-derived materials at various calcination temperatures (for 5 hours).
igure 4. Surface areas of the unsubstituted and substituted materials calcined at various temperatures (for 5 hours). Figure 4 illustrates the effect of calcination temperature on the surface areas of the high temperature aerogels. A dramatic decrease in surface area occurred for calcinatior temperatures higher than 1000°C and 800°C for the unsubstituted and substituted materials, respectively. This surface area loss appears to be a consequence of phase transformations which will be discussed in the next section. The Mn-substituted hexaaluminates displayed better heat resistance and higher surface areas than the Co-substituted materials when the calcination temperatures were higher than 1000°C. This result suggests that Mn is a better substitution cation than Co.
igure 4. Surface areas of the unsubstituted and substituted materials calcined at various temperatures (for 5 hours). Figure 4 illustrates the effect of calcination temperature on the surface areas of the high temperature aerogels. A dramatic decrease in surface area occurred for calcinatior temperatures higher than 1000°C and 800°C for the unsubstituted and substituted materials, respectively. This surface area loss appears to be a consequence of phase transformations which will be discussed in the next section. The Mn-substituted hexaaluminates displayed better heat resistance and higher surface areas than the Co-substituted materials when the calcination temperatures were higher than 1000°C. This result suggests that Mn is a better substitution cation than Co.
Figure 5. X-ray diffraction patterns of the Mn-substituted high temperature aerogel-derived materials.
Figure 5. X-ray diffraction patterns of the Mn-substituted high temperature aerogel-derived materials.
The second reaction is very slow and requires long calcination times at temperatures above 1200°C (9). The results indicate that the dramatic loss in surface area at temperatures between 800 and 1200°C was partly due to phase transformations.
The second reaction is very slow and requires long calcination times at temperatures above 1200°C (9). The results indicate that the dramatic loss in surface area at temperatures between 800 and 1200°C was partly due to phase transformations.
Table I Phases present in the high temperature aerogel-derived materials following calcination at various temperatures for 5 hours.
Table I Phases present in the high temperature aerogel-derived materials following calcination at various temperatures for 5 hours.
ities Figure 3. Temperature dependant conductivitie: ured (0) of: MEEP*NaMont measured parallel to th (@), composite layers (A), VIF fit (—); MEEP-NaMon ured perpendicular to the composite layers (W), VTF fi (-—); Pristine NaMont (o)m measured perpendicula to the layers.
ities Figure 3. Temperature dependant conductivitie: ured (0) of: MEEP*NaMont measured parallel to th (@), composite layers (A), VIF fit (—); MEEP-NaMon ured perpendicular to the composite layers (W), VTF fi (-—); Pristine NaMont (o)m measured perpendicula to the layers.
In this paper we study the fracture mechanisms and mechanical properties of montmo- rillonite based nanocomposites with polystyrene. Although these are by far not examples of systems with improved mechanical properties they provide very good model systems that enable us to test the methods we use, as well as to compare against the failure behaviour of the pure polymer and its conventional filled systems.
In this paper we study the fracture mechanisms and mechanical properties of montmo- rillonite based nanocomposites with polystyrene. Although these are by far not examples of systems with improved mechanical properties they provide very good model systems that enable us to test the methods we use, as well as to compare against the failure behaviour of the pure polymer and its conventional filled systems.
The fracture toughness of the same samples was measured by Chevron-notched samples in a 3-point bending geometry. Interestingly, the load-deformation curve (figure 2) shows a failure qualitatively similar to that of the pure polymer, although the K;¢ decreased (Table II). At this point we should mention that there are ways to increase the fracture toughness of the nanocomposites [4] but our aim in the present study is to explore methods for mechanical characterization of the nanocomposites and identify their failure mechanisms, as it will become obvious by the TEM results. Moreover, by observing the fractured surfaces with an SEM (figure 3) it is revealed that the fractured surface of the nanocomposite is Table I The longitudinal and shear Young moduli for selected nanocomposites are shown.
The fracture toughness of the same samples was measured by Chevron-notched samples in a 3-point bending geometry. Interestingly, the load-deformation curve (figure 2) shows a failure qualitatively similar to that of the pure polymer, although the K;¢ decreased (Table II). At this point we should mention that there are ways to increase the fracture toughness of the nanocomposites [4] but our aim in the present study is to explore methods for mechanical characterization of the nanocomposites and identify their failure mechanisms, as it will become obvious by the TEM results. Moreover, by observing the fractured surfaces with an SEM (figure 3) it is revealed that the fractured surface of the nanocomposite is Table I The longitudinal and shear Young moduli for selected nanocomposites are shown.
In order to get more insight on the fracture behaviour, we carried out TEM observations of thin nanocomposite films under strain. This can be achieved by mounting m-thin slices of the nanocomposite —obtained either by microtoming the tensile bars or by spin casting— on a ductile copper grid with a 1xlmm mesh size [10]. Subsequently, we strain the system with a very small strain rate to a 5-10% deformation (figure 4). The ductile copper remains deformed and can keep the composite material under strain inside the TEM. In order to minimize as much as possible failure mechanisms that originate from surface properties of ultra-thin films, we used samples thicker than 0.5 wm and up to 1.5 ym, beyond which it becomes difficult to observe them with our TEM.
In order to get more insight on the fracture behaviour, we carried out TEM observations of thin nanocomposite films under strain. This can be achieved by mounting m-thin slices of the nanocomposite —obtained either by microtoming the tensile bars or by spin casting— on a ductile copper grid with a 1xlmm mesh size [10]. Subsequently, we strain the system with a very small strain rate to a 5-10% deformation (figure 4). The ductile copper remains deformed and can keep the composite material under strain inside the TEM. In order to minimize as much as possible failure mechanisms that originate from surface properties of ultra-thin films, we used samples thicker than 0.5 wm and up to 1.5 ym, beyond which it becomes difficult to observe them with our TEM.
Figure 3 Scanning Electron Microscopy (SEM) imaging of the fractured surfaces of the samples in figure 2. Left: polystyrene, right: intercalated nanocomposite.
Figure 3 Scanning Electron Microscopy (SEM) imaging of the fractured surfaces of the samples in figure 2. Left: polystyrene, right: intercalated nanocomposite.
Summarizing, ultrasonic and 3-point bend techniques were successfully employed for the mechanical characterisation of nanocomposite materials. Furthermore, TEM studies of strained composite films revealed the failure mechanisms located mainly at the polymer- silicate interface. A way to improve the toughness of these materials is through the strength- ening of the polymer-surface binding (4, 6]
Summarizing, ultrasonic and 3-point bend techniques were successfully employed for the mechanical characterisation of nanocomposite materials. Furthermore, TEM studies of strained composite films revealed the failure mechanisms located mainly at the polymer- silicate interface. A way to improve the toughness of these materials is through the strength- ening of the polymer-surface binding (4, 6]
Fig. 1. Schematic structure of layered trioctahedral clays cal This paper discusses recent progress toward forming sheet silicate structures that might be used as components of electrochemical devices such as sensors and batteries. of the silicate sites are occupied there is no possibility of forming stable Alkali-O-A! positions so lithium diffusion through the lattice should be enhanced. It's formula can be represented by Li,(Mg3.xLi,)(Si4010)F2 ¢nH20. Layered silicates have been of interest recently for use as electroluminescent displays, polyelectrolytes [5-8], and organic/inorganic composites for creating stress relieving points in engineered materials [9-11]. In order to incorporate polymers within the layers it is necessary to swell the mica-like layers greater than that found for the alkali metal alone. Several types of silicates, such as montmorrilonite and fluorohectorite which have low charge densities, will expand almost continuously when placed in aqueous solutions. This generally leads to a controlled amount of polymer adsorbing onto the layers by controlling the component ratios [12]. However, there are no reports of the swelling of fluorophlogopite so a means must be found to incorporate polymers between the sheets. This paper discusses recent progress toward forming sheet.silicate structures that
Fig. 1. Schematic structure of layered trioctahedral clays cal This paper discusses recent progress toward forming sheet silicate structures that might be used as components of electrochemical devices such as sensors and batteries. of the silicate sites are occupied there is no possibility of forming stable Alkali-O-A! positions so lithium diffusion through the lattice should be enhanced. It's formula can be represented by Li,(Mg3.xLi,)(Si4010)F2 ¢nH20. Layered silicates have been of interest recently for use as electroluminescent displays, polyelectrolytes [5-8], and organic/inorganic composites for creating stress relieving points in engineered materials [9-11]. In order to incorporate polymers within the layers it is necessary to swell the mica-like layers greater than that found for the alkali metal alone. Several types of silicates, such as montmorrilonite and fluorohectorite which have low charge densities, will expand almost continuously when placed in aqueous solutions. This generally leads to a controlled amount of polymer adsorbing onto the layers by controlling the component ratios [12]. However, there are no reports of the swelling of fluorophlogopite so a means must be found to incorporate polymers between the sheets. This paper discusses recent progress toward forming sheet.silicate structures that
Fig. 2 X-ray diffraction pattern of sodium fluorophlogopite The initial Na-fluorophlogopite was indexed and found to have a triclinic unit cell with dimensions of a=5.308(1), b=9.179(2), c=9.788(2), alpha=99.223(5), beta=91.473(5), and gamma=90.025(8). It’s powder x-ray pattern is shown in figure 2. A WDS elemental analysis on the electron microprobe led to the formula of NaMg3(AISi3O10)F2 by using an average over five different crystals which had a mesh size of 200-300. A thermogravimetric analysis carried out in nitrogen showed negligible weight loss, less than 0.5%, up to 1000°C therefore demonstrating no interlayer water or structural hydroxyl groups.
Fig. 2 X-ray diffraction pattern of sodium fluorophlogopite The initial Na-fluorophlogopite was indexed and found to have a triclinic unit cell with dimensions of a=5.308(1), b=9.179(2), c=9.788(2), alpha=99.223(5), beta=91.473(5), and gamma=90.025(8). It’s powder x-ray pattern is shown in figure 2. A WDS elemental analysis on the electron microprobe led to the formula of NaMg3(AISi3O10)F2 by using an average over five different crystals which had a mesh size of 200-300. A thermogravimetric analysis carried out in nitrogen showed negligible weight loss, less than 0.5%, up to 1000°C therefore demonstrating no interlayer water or structural hydroxyl groups.
Fig. 3. Lattice spacing of copper fluorophlogopite on intercalation of long chain amines. Closed circles correspond to an interleaved monolayer, whereas open circles are probably associated with a bilayer of amines, and approach the spacing observed in vermiculite [14]. Reaction of the copper fluorophlogopite with alkylamines caused a substantia expansion of the silicate lattice as shown in figure 3. The lower slope line corresponds t a single layer of amine with the nitrogen adjacent to the oxygen of the silicate and th hydrocarbon chains interleaved. The higher expansions are typical of that for a bilaye of amine, as found for vermiculite and other readily swellable clays. The slope correspond to an angle of the chains of around 60° to the silicate layer. Unlike th vermiculite where an ammonium salt is formed, in this material the amine is acting as | solvent for the cations. TGA showed that around 0.3 or 0.7 amine groups wer incorporated per (Si,Al)4019 group for the two configurations shown in the figure again consistent with single and bilayer amine configurations. Lithium Taeniolite
Fig. 3. Lattice spacing of copper fluorophlogopite on intercalation of long chain amines. Closed circles correspond to an interleaved monolayer, whereas open circles are probably associated with a bilayer of amines, and approach the spacing observed in vermiculite [14]. Reaction of the copper fluorophlogopite with alkylamines caused a substantia expansion of the silicate lattice as shown in figure 3. The lower slope line corresponds t a single layer of amine with the nitrogen adjacent to the oxygen of the silicate and th hydrocarbon chains interleaved. The higher expansions are typical of that for a bilaye of amine, as found for vermiculite and other readily swellable clays. The slope correspond to an angle of the chains of around 60° to the silicate layer. Unlike th vermiculite where an ammonium salt is formed, in this material the amine is acting as | solvent for the cations. TGA showed that around 0.3 or 0.7 amine groups wer incorporated per (Si,Al)4019 group for the two configurations shown in the figure again consistent with single and bilayer amine configurations. Lithium Taeniolite
Table 1. Conductivity of Silicates
Table 1. Conductivity of Silicates
X-ray diffraction measurements provide a quick measure for nanocomposite formation. Generally intense reflections in the range 26 = 3 - 9° indicate either an ordered intercalated hybrid or an immiscible system. The former, however, shows an increase in the d-spacing corresponding to the intercalation of the polymer chains in the host galleries while in the latter the d-spacing remains unchanged. In delaminated hybrids, on the other hand, XRD patterns with no distinct features in the low 20 range are anticipated due to the introduced disorder and the loss of structural registry in the silicate layers. TABLE I
X-ray diffraction measurements provide a quick measure for nanocomposite formation. Generally intense reflections in the range 26 = 3 - 9° indicate either an ordered intercalated hybrid or an immiscible system. The former, however, shows an increase in the d-spacing corresponding to the intercalation of the polymer chains in the host galleries while in the latter the d-spacing remains unchanged. In delaminated hybrids, on the other hand, XRD patterns with no distinct features in the low 20 range are anticipated due to the introduced disorder and the loss of structural registry in the silicate layers. TABLE I
X-ray Diffraction Analysis of PEI-10 Nanocomposites
X-ray Diffraction Analysis of PEI-10 Nanocomposites
Figure 5. Fire test of PEI and PEI nanocomposite. Figure 3. Isothermal at 450 °C in air.
Figure 5. Fire test of PEI and PEI nanocomposite. Figure 3. Isothermal at 450 °C in air.
Fig. 2. X-ray diffraction 3D plot of the dehydration of DTA,H,V 190 2g*8H,0, showing the growth in the higher angle peak as the lower one decreases. stability. This study, then, provided adequate data for structure determination. The structure of DTA,H,V 19079*8H20O is shown in figure 3. In this material, decavanadate clusters are surrounded by water molecules, 2 of which are probably in the form of H,0+. These hydronium ions and 4 DTA chains provide charge balance for the V 19029 clusters. The DTA chains are stacked such that the head groups of neighboring chains are in opposite directions. The additional 6 water molecules are hydrogen bonded to the decavanadate clusters, as can be seen in figure 4. The single crystal data gave a triclinic cell with a space group of P 1, and dimensions a=9,895(1)A, b=11.596(1)A, c=21.924(1)A, o=95.15(1)°, B=93.78(1)°, and y=101.36(1)°.
Fig. 2. X-ray diffraction 3D plot of the dehydration of DTA,H,V 190 2g*8H,0, showing the growth in the higher angle peak as the lower one decreases. stability. This study, then, provided adequate data for structure determination. The structure of DTA,H,V 19079*8H20O is shown in figure 3. In this material, decavanadate clusters are surrounded by water molecules, 2 of which are probably in the form of H,0+. These hydronium ions and 4 DTA chains provide charge balance for the V 19029 clusters. The DTA chains are stacked such that the head groups of neighboring chains are in opposite directions. The additional 6 water molecules are hydrogen bonded to the decavanadate clusters, as can be seen in figure 4. The single crystal data gave a triclinic cell with a space group of P 1, and dimensions a=9,895(1)A, b=11.596(1)A, c=21.924(1)A, o=95.15(1)°, B=93.78(1)°, and y=101.36(1)°.
Fig. 4. Water hydrogens bonding the decavanadate clusters together [15]. Fig. 3. Structure of DTA,H, V;)O2g*8H,O showing the arrangement of the organic chains [15].
Fig. 4. Water hydrogens bonding the decavanadate clusters together [15]. Fig. 3. Structure of DTA,H, V;)O2g*8H,O showing the arrangement of the organic chains [15].
Fig. 6. X-ray diffraction pattern of the manganese chloride surfactant material [2] showing the layer-like nature of the compound.
Fig. 6. X-ray diffraction pattern of the manganese chloride surfactant material [2] showing the layer-like nature of the compound.
The data obtained by chemical analysis (table I), IR and optical spectroscopy, XRD give the following formulas for the synthesized compounds: [LiAl(OH).]}.[Cuedta]-4H20, [LiAl(OH).]2[Coedta]-4H,O and [LiAl(OH)¢][Niedta]-4H,O. Edta is an anion of ethylenediaminetetraacetic acid forming stable complexes with transition metals. Thermal decomposition of LADH-X materials leads to dehydration and dehydroxylation of the hydroxide matrix and to the collapse red/ox process of the complex anion with the appearence of transition metal phase. DTA and mass spectral data show that the process of decomposition is a sequence of the following stages. Up to 200°C elimination of adsorbed and interlayer H2O from the hydroxide prevails. The process of dehydration of interstitially located water molecules proceeds in two stages approximately at 60-90°C (2 molecules) and 140-180°C (2 molecules) for all Medta- intercalates. Heating above 200°C results in the destruction of the hydroxide matrix. The process occurs before the decomposition of organic anion excepting LADH-[Cuedta], where both processes start near-concurrently. On a further increase in temperature, the process of thermolysis proceeds differently. For the nickel-containing intercalate at temperatures up to 300°C, the decomposition process goes with the predominent release of water molecules into the gas phase, which is caused by the destruction of hydroxide layeres. In the case of the copper and cobalt containing intercalates, water removal is accompanied by the simultaneous formation of gaseous products with 28 m/e and 44 m/e, which were referred to CO and CO, (fig.1). At temperatures above 300°C, for all compounds, one can observe not only the release of CO, CO. and H20 (18 m/e), but also the appearance of NH; and H; (16 and 2 m/e), which indicates a deep destruction of edta. The considerable content of carbon is determined in the products of vacuum decomposition, which points to the carbonization of the samples in the course of thermolysis (table I), The C/M and N/M ratios show that approximately one half of the carbon and nitrogen content is removed from the initial intercalates upon thermolysis.
The data obtained by chemical analysis (table I), IR and optical spectroscopy, XRD give the following formulas for the synthesized compounds: [LiAl(OH).]}.[Cuedta]-4H20, [LiAl(OH).]2[Coedta]-4H,O and [LiAl(OH)¢][Niedta]-4H,O. Edta is an anion of ethylenediaminetetraacetic acid forming stable complexes with transition metals. Thermal decomposition of LADH-X materials leads to dehydration and dehydroxylation of the hydroxide matrix and to the collapse red/ox process of the complex anion with the appearence of transition metal phase. DTA and mass spectral data show that the process of decomposition is a sequence of the following stages. Up to 200°C elimination of adsorbed and interlayer H2O from the hydroxide prevails. The process of dehydration of interstitially located water molecules proceeds in two stages approximately at 60-90°C (2 molecules) and 140-180°C (2 molecules) for all Medta- intercalates. Heating above 200°C results in the destruction of the hydroxide matrix. The process occurs before the decomposition of organic anion excepting LADH-[Cuedta], where both processes start near-concurrently. On a further increase in temperature, the process of thermolysis proceeds differently. For the nickel-containing intercalate at temperatures up to 300°C, the decomposition process goes with the predominent release of water molecules into the gas phase, which is caused by the destruction of hydroxide layeres. In the case of the copper and cobalt containing intercalates, water removal is accompanied by the simultaneous formation of gaseous products with 28 m/e and 44 m/e, which were referred to CO and CO, (fig.1). At temperatures above 300°C, for all compounds, one can observe not only the release of CO, CO. and H20 (18 m/e), but also the appearance of NH; and H; (16 and 2 m/e), which indicates a deep destruction of edta. The considerable content of carbon is determined in the products of vacuum decomposition, which points to the carbonization of the samples in the course of thermolysis (table I), The C/M and N/M ratios show that approximately one half of the carbon and nitrogen content is removed from the initial intercalates upon thermolysis.
Fig.2. HREM microphotograph of LADH-[Niedta] decomposed in vacuum at 400°C.
Fig.2. HREM microphotograph of LADH-[Niedta] decomposed in vacuum at 400°C.
Fig.3. Powder X-ray diffraction patterns of LADH-[Cuedta] before (1) and after heating (2) in air at 250°C during 20 min. Heating in air for several hours gives rise to formation of metal oxide phase and to subsequent production of spinel forms at high temperatures (above 500°C). Since the destruction of the organic anion for the ([Cuedta]-containing intercalate occurs near-concurrently with dehydroxilation of matrix, it was interesing to obtain nanophase particles in the layered hydroxide system in the case of suitable temperature and period of heating. For 20 minutes at 250°C the result was retention of the layered hydroxide and formation not of the CuO composite, but of the metal copper one (fig.3). It is attributable to the slow process of dehydroxilation of the matrix which protects the generated copper phase against Oo.
Fig.3. Powder X-ray diffraction patterns of LADH-[Cuedta] before (1) and after heating (2) in air at 250°C during 20 min. Heating in air for several hours gives rise to formation of metal oxide phase and to subsequent production of spinel forms at high temperatures (above 500°C). Since the destruction of the organic anion for the ([Cuedta]-containing intercalate occurs near-concurrently with dehydroxilation of matrix, it was interesing to obtain nanophase particles in the layered hydroxide system in the case of suitable temperature and period of heating. For 20 minutes at 250°C the result was retention of the layered hydroxide and formation not of the CuO composite, but of the metal copper one (fig.3). It is attributable to the slow process of dehydroxilation of the matrix which protects the generated copper phase against Oo.
Figure 1. (a) The initial and (b) final structures of the nanocrystal with the diameter 6 at T* = 0.05.
Figure 1. (a) The initial and (b) final structures of the nanocrystal with the diameter 6 at T* = 0.05.
Figure 2. (a) The initial and (b) final structures of the nanocrystal with the diameter 14 at T* = 0.05. = 0.05 as shown in figure 3(a). Then the thermal stability of this sample was examined by the similar calculation at T* = 0.10, 0.15 and 0.20 and the obtained structures are shown in figures 3(b), (c) and (d). At T* = 0.10, the nanocrystal structure was stable but the orientation of one of the particles changed to make more stable interface. The rearrangement of the atom in the particle occurred at the grain boundary but systematic rotation of the particle could not be observed. At T* =0.15 and 0.20, the nanocrystal structure was not stable any more. The structure changed into single crystal structure by the same mechanism as the case of 7* = 0.10.
Figure 2. (a) The initial and (b) final structures of the nanocrystal with the diameter 14 at T* = 0.05. = 0.05 as shown in figure 3(a). Then the thermal stability of this sample was examined by the similar calculation at T* = 0.10, 0.15 and 0.20 and the obtained structures are shown in figures 3(b), (c) and (d). At T* = 0.10, the nanocrystal structure was stable but the orientation of one of the particles changed to make more stable interface. The rearrangement of the atom in the particle occurred at the grain boundary but systematic rotation of the particle could not be observed. At T* =0.15 and 0.20, the nanocrystal structure was not stable any more. The structure changed into single crystal structure by the same mechanism as the case of 7* = 0.10.

Related papers

The Apparent Thermal Conductivity of Liquids Containing Solid Particles of Nanometer Dimensions: A Critique
Georgia Tertsinidou, William Wakeham

International Journal of Thermophysics, 2015

There have been conflicting statements in the literature of the last twenty years about the behavior of the apparent thermal conductivity of two-or three-phase systems comprising solid particles with nanometer dimensions suspended in fluids. It has been a feature of much of the work that these multi-phase systems have been treated as if a single-phase fluid and that the thermodynamic characteristics of the system have varied even though the systems have been given the same name. These so-called nanofluids have been the subject of a large number of investigations by a variety of different experimental techniques. In the current paper we critically evaluate the studies of seven of the simplest particulate/fluid systems; Cu, CuO, Al 2 O 3 , and TiO 2 suspended in water and ethylene glycol. Our conclusion is that when results for exactly the same thermodynamic system are obtained with proven experimental techniques, the apparent thermal conductivity of the nanofluid exhibits no behavior that is unexpected and inconsistent with a simple model of conduction in stationary, multiphase systems. The wider variety of behavior that has been reported in the literature for these systems is therefore attributed to ill-characterization of the thermodynamic system and/or the application of experimental techniques of unproven validity.

View PDFchevron_right
Effective thermal conductivity of wet particle assemblies
František Štěpánek

International Journal of Heat and Mass Transfer, 2004

The effective thermal conductivity of mono-and poly-dispersed random assemblies of spherical particles and irregular crystals, both dry and partially or fully saturated by wetting and non-wetting liquids, has been determined computationally by numerical solution of the FourierÕs law on 3-D reconstructed media and experimentally by the transient hot wire method. The effect of spatial distribution and volume fractions of the vapour, liquid, and solid phases on effective thermal conductivity was systematically investigated. A power-law correlation for estimating the effective conductivity, valid over a wide range of phase volume fractions and relative conductivities of components, has been proposed.

View PDFchevron_right
Effect of nanomaterial properties on thermal conductivity of heat transfer fluids and nanomaterial suspension
111 Rohit Mahesh Khedkar

Energy has been rated as the single most important issue facing humanity in the current as well as next 50 years. Securing clean energy has become the top priority of most developed countries. Considering the rapid increase in energy demand worldwide, intensifying the heat transfer process and reducing energy loss due to ineffective use have become an increasingly important task. Fundamentally, energy conversion and transportation occur at atomic or molecular levels, Nanoscience and nanotechnology are expected to play a significant role in revitalizing the traditional energy industries and stimulating the emerging renewable energy industries. Nanofluid is a modern engineering heat transfer fluid with superior potential for enhancing the heat transfer performance of conventional fluids such as water, ethylene glycol and oils. It is consisting of solid nanoparticles with sizes typically of 1-100 nm suspended in base fluids. Many attempts have been made to investigate its important thermal properties, i.e. thermal conductivity; however, no definitive agreements and idea have emerged about this property. This article reports the effect of different nanomaterial on the thermal conductivity enhancement of nanofluids experimentally. TiO 2 , Fe 3 O 4 and Al 2 O 3 nanoparticles dispersed in water and ethylene glycol with volume concentration of 1-7.5 vol. % is used in the present study. A transient hot-wire apparatus (KD2 pro) is used for measuring the thermal conductivity of nanofluids. The results show that all the heat transfer fluids show an increase in thermal conductivity with the addition of nanoparticles in it. The measured thermal conductivity of nanofluids increased as the particle concentrations increased and are higher than the values of the base liquids. This confirms the effect of volume concentration of nanoparticles on the thermal conductivity enhancement.

View PDFchevron_right
Enhancing thermal conductivity of fluids with nanoparticles
Jeffrey Eastman

2009

View PDFchevron_right
Thermal conductivity of nanoparticle suspensions
jeremy johnson

2002

We describe an optical beam deflection technique for measurements of the thermal diffusivity of fluid mixtures and suspensions of nanoparticles with a precision of better than 1%. Our approach is tested using the thermal conductivity of ethanol-water mixtures; in nearly pure ethanol, the increase in thermal conductivity with water concentration is a factor of 2 larger than predicted by effective medium theory. Solutions of C 60-C 70 fullerenes in toluene and suspensions of alkanethiolate-protected Au nanoparticles were measured to maximum volume fractions of 0.6% and 0.35 vol %, respectively. We do not observe anomalous enhancements of the thermal conductivity that have been reported in previous studies of nanofluids; the largest increase in thermal conductivity we have observed is 1.3% ± 0.8% for 4 nm diam Au particles suspended in ethanol.

View PDFchevron_right
Experimental investigation of viscosity and thermal conductivity of suspensions containing nanosized ceramic particles
Tavman Sebnem, Ismail Tavman, Alpaslan Turgut

2008

Purpose: In this study we report measurements of effective thermal conductivity by using 3ω method and effective viscosity by vibro-viscometer for SiO 2 -water and Al 2 O 3 -water nanofluids at different particle concentrations and temperatures.

View PDFchevron_right
Thermal Conductivity of Suspension of Aggregating Nanometric Rods
Rodolphe Heyd

Entropy

Enhancing thermal conductivity of simple fluids is of major interest in numerous applicative systems. One possibility of enhancing thermal properties consists of dispersing small conductive particles inside. However, in general, aggregation effects occur and then one must address systems composed of dispersed clusters composed of particles as well as the ones related to percolated networks. This papers analyzes the conductivity enhancement of different microstructures scaling from clusters dispersed into a simple matrix to the ones related to percolated networks exhibiting a fractal morphology.

View PDFchevron_right
Thermal transport phenomena in nanoparticle suspensions
Pietro Asinari

Journal of Physics: Condensed Matter, 2016

Nanoparticle suspensions in liquids have received great attention, as they may offer an approach to enhance thermophysical properties of base fluids. A good variety of applications in engineering and biomedicine has been investigated with the aim of exploiting the above potential. The multiscale nature of nanosuspensions raises several issues in defining a comprehensive modelling framework, incorporating relevant molecular details and much larger scale phenomena, such as particle aggregation and their dynamics. The objectives of the present topical review is to report and discuss the main heat and mass transport phenomena ruling their macroscopic behaviour, arising from molecular details. Relevant experimental results are included and properly put in the context of recent observations and theoretical studies, which solved longstanding debates about thermophysical properties enhancement. Major transport phenomena are discussed and in-depth analysis is carried out for highlighting the role of geometrical (nanoparticle shape, size, aggregation, concentration), chemical (pH, surfactants, functionalization) and physical parameters (temperature, density). We finally overview several computational techniques available at different scales with the aim of drawing the attention on the need for truly multiscale predictive models. This may help the development of next-generation nanoparticle suspensions and their rational use in thermal applications.

View PDFchevron_right
Thermal Diffusivity and Viscosity of Suspensions of Disk-Shaped Nanoparticles
Yogesh Joshi

Industrial & Engineering Chemistry Research, 2013

In this work we conduct a transient heat conduction experiment with an aqueous suspension of nanoparticle disks of Laponite JS, a sol forming grade, using laser light interferometry. The image sequence in time is used to measure thermal diffusivity and thermal conductivity of the suspension. Imaging of the temperature distribution is facilitated by the dependence of refractive index of the suspension on temperature itself. We observe that with the addition of 4 volume % of nano-disks in water, thermal conductivity of the suspension increases by around 30%. A theoretical model for thermal conductivity of the suspension of anisotropic particles by Fricke as well as by Hamilton and Crosser explains the trend of data well. In turn, it estimates thermal conductivity of the Laponite nanoparticle itself, which is otherwise difficult to measure in a direct manner. We also measure viscosity of the nanoparticle suspension using a concentric cylinder rheometer. Measurements are seen to follow quite well, the theoretical relation for viscosity of suspensions of oblate particles that includes up to two particle interaction. This result rules out the presence of clusters of particles in the suspension. The effective viscosity and thermal diffusivity data show that the shape of the particle has a role in determining enhancement of thermophysical properties of the suspension.

View PDFchevron_right
Enhancement of Thermal Conductivity of Liquid by Dispersing Ultrafine Particles of Fly Ash
ANOJ MEENA

International Journal of Technical Research & Science

The design of energy-efficient heat transfer equipment, as well as the research for enhancing the thermal capability of conventional fluids, contributes to the effort for better energy management. An apparatus is designed and fabricated with heat transfer fluid by an ultrafine particle of fly ash to measure the Effective thermal conductivity of liquids under various temperature ranges. Natural convection heat transfer has been experimentally investigated to find out the heat loss to the atmosphere. The apparatus is a completely different form of thermal conductivity apparatus which has been widely applied in the normal laboratory The objective of the present work is to prepare HT fluid of FLY ASH and nano/microparticles with distilled water as a base fluid, measurement of thermo-physical properties (thermal conductivity, viscosity and density), FLY ASH/distilled water HT fluid, working principles of various instruments like KD2 Pro, Ubbelhode viscometer and Pycnometer have been discussed. Description of the experimental setup for measurement of heat transfer coefficient with its schematic diagram, experimental procedure and equations used to calculate the heat transfer coefficient are also discussed. For the preparation of FLY ASH/distilled water with a surfactant, Gum Arabic is used as a surfactant. This surfactant was added in distilled water first and then FLY ASH nano/ were added in it and sonicated for 2.5 hours continuously. But still, the nano/ were not dispersed properly. It remains stable for 1-2 hours only and then nano/ start to settle down. Results show that the thermal conductivity of alumina/distilled water HT fluid increases up to 2.40% for 0.1 volume% and 0.328% for 0.5 volume% of alumina/distilled water as compared to distilled water.

View PDFchevron_right

Related papers

HEAT TRANSFER ENHANCEMENT OF NANO FLUIDS –A REVIEW
Editor IJRET
View PDFchevron_right
Thermal Conductivity Calculations for Nanoparticles Embedded in a Base Fluid
Soran Mamand

Applied Sciences, 2021

View PDFchevron_right
A Numerical Technique for Computing Effective Thermal Conductivity of Fluid–Particle Mixtures
Sunil Kumar

Numerical Heat Transfer, Part B: Fundamentals, 2005

View PDFchevron_right
Investigation of particle size effect on thermal conductivity enhancement of distilled water-Al2O3 nano fluids
Yogesh Kokate

Fluid Mechanics Research International Journal, 2019

View PDFchevron_right
An Experimental Investigation into the Thermal Properties of Nano Fluid
Honey Sinha, Tushar Anand
View PDFchevron_right
Mixing effect on the enhancement of the effective thermal conductivity of nanoparticle suspensions (nanofluids)
Calvin Li

2007

View PDFchevron_right
The effect of Fe 3 O 4 nanoparticles on the thermal conductivities of various base fluids
Cem Altan

Nanotechnology, 2011

View PDFchevron_right
Mechanisms of heat flow in suspensions of nano-sized particles (nanofluids)
P. Keblinski

International Journal of Heat and Mass Transfer, 2002

View PDFchevron_right
Heat Transfer Improvement by Nano Fluids
Sumit Wani

2016

View PDFchevron_right
The Interfacial Layer Effect on Thermal Conductivity of Nano-Colloidal Dispersions
Sunita Srivastava

International Journal of Applied Physics and Mathematics, 2014

View PDFchevron_right
Heat conduction mechanisms in nanofluids and suspensions
Jinwei Gao

Nano Today, 2012

View PDFchevron_right
Thermal conductivity in spherical nanocolloids
Majles Ara

International Journal of Thermal Sciences, 2013

View PDFchevron_right
The effect of nanoparticles aggregation on the thermal conductivity of nanofluids at very low concentrations: Experimental and theoretical evaluations
Solaimani Nazar, mohsen motavassel, mohsen motevasel
View PDFchevron_right
Study of Nano Particles for Enhanced Heat Transfer Characteristics of Base Fluids for Cool Thermal Energy System
Promit Choudhury

2014

View PDFchevron_right
Measuring Thermal Conductivity of Fluids Containing Oxide Nanoparticles
Jeffrey Eastman

Journal of Heat Transfer, 1999

View PDFchevron_right
Physical and Chemical Properties of CaCl2/H2O and LiBr/H2O Systems Confined to Nanopores of Silica Gels
Giovanni Restuccia

MRS Proceedings, 1996

View PDFchevron_right
Effects of Particle Surface Charge, Species, Concentration, and Dispersion Method on the Thermal Conductivity of Nanofluids
Raghu Gowda

Advances in Mechanical Engineering, 2010

View PDFchevron_right
Thermal Conductivity of Fluids Containing Quasi-Spherical Nanoinclusions
Kliment Hadjov
View PDFchevron_right
Regimes of heat transfer in particle suspensions
ali yousefi

2020

View PDFchevron_right
A Review on Heat Transfer Augmentation in Nano Fluids
International Journals for Researchers [ER Publication, WOAR Journals, IJEAS and IJEART]
View PDFchevron_right
Pool boiling characteristics of nano-fluids
Putra We

International Journal of Heat and Mass Transfer, 2003

View PDFchevron_right

Related topics

Engineering