Скачать презентацию The role of EXAFS in some condensed matter Скачать презентацию The role of EXAFS in some condensed matter

3b79948b09bc620316292032467386e4.ppt

  • Количество слайдов: 38

The role of EXAFS in some condensed matter systems: Challenges and opportunities for NSLS The role of EXAFS in some condensed matter systems: Challenges and opportunities for NSLS II Vince Harris Department of Electrical and Computer Engineering Northeastern University Boston, MA 02115 USA [email protected] neu. edu www. cm 3 ic. neu. edu Date: Jan. 16, 2008

Many contributions from: • Aria Yang • Scott Calvin • Bruce Ravel • Joe Many contributions from: • Aria Yang • Scott Calvin • Bruce Ravel • Joe Woicik • Trevor Tyson • Dave Pappas • Many studies have benefited from the beamline expertise of John Kirkland Barry Karlin • Maybe as many as 50 -75 coauthors to acknowledgde!

Nanoelectronics • • Other Moore technologies • Minimum feature size of ICs • Magnetic Nanoelectronics • • Other Moore technologies • Minimum feature size of ICs • Magnetic storage media • Digital photography pixel density • CPU speed… Similar Moore’s trend is seen in the minimum feature size. Gate oxide barriers are below 10 nm. 1 nm in some cases. How does one study materials properties at these length scales? How does synchrotron radiation tools help address these needs? – – – Phase ID Atomic fraction measurements Bond lengths Coordination Phase stability Cation site distribution and valency

Scope • The study of thin films – NSLS I • Focused light beamlines Scope • The study of thin films – NSLS I • Focused light beamlines versus nonfocus beamlines – – – Metallization of diamond (Spintronics) Crystallization of thin films (phase change) Strained semiconductors (Nanoelectronics) Buried layers (Nanoelectronics) Cation distribution in oxides (Microwave electronics) – Issues • • • Sample size Concentration Thickness Beam stability Dynamics

PRB Brief Report 52(11) 7890 (1995) Experimental details MBE grown Fe (0. 8 nm) PRB Brief Report 52(11) 7890 (1995) Experimental details MBE grown Fe (0. 8 nm) EXAFS measurements on X 23 B NSLS Sample size is 2 mm x 3 mm TEY, room temp, std. conditions Spot size ~250 mm x 250 mm FEFF modeling (FEFF 3? ) Goal Determine nature of metallization Structure identification Structure stability: bcc vs fcc Bond distances, coordination

PRB Brief Report 52(11) 7890 (1995) Experimental details MBE grown Fe (0. 8 nm) PRB Brief Report 52(11) 7890 (1995) Experimental details MBE grown Fe (0. 8 nm) EXAFS measurements on X 23 B NSLS Sample size is 2 mm x 3 mm TEY, room temp, std. conditions Spot size ~250 mm x 250 mm FEFF modeling (FEFF 3? ) Goal Determine nature of metallization Structure identification Structure stability: bcc vs fcc Bond distances, coordination

Challenges Very thin films (6 -8 Å, 4 -5 ml) Small and very valuable Challenges Very thin films (6 -8 Å, 4 -5 ml) Small and very valuable substrates Synthetic diamond 2 mm x 3 mm substrates Beam spot size: 250 mm x 250 mm If beam walks, this is a nightmare! Results FEFF modeling (FEFF 3. 11 single scattering) determines Fe-Fe and Fe-C bonds 2 -11 Å-1 (k 2) Fe-Fe: 2. 51 +/- 0. 02 Å (~1% in-plane contraction of lattice) Fe-C: 1. 84 +/- 0. 02 Å (~2. 7 atoms) (Fe-C 1. 98Å (Fe 2 C); 2. 14Å (Fe 3 C)) Growth of single crystal austenite PRB Brief Report 52(11) 7890 (1995) What would be great? In situ measurements during growth In-plane vs out-of-plane measurements What is needed? Higher counts Smaller spots Beam stability Linearity in detector circuitry, challenging.

Phys. Rev. B 43, 3, 15 Jan. 1991 (Brief Report) Goals Nature of strain Phys. Rev. B 43, 3, 15 Jan. 1991 (Brief Report) Goals Nature of strain in Ge. Si epitaxial films Experimental Samples are 34 nm Ge. Si (31% Ge) CVD grown films In-plane vs out-of-plane bond distances and lattice parameters EXAFS using FY Sample spinning X 23 A 2 beam does not walk Good linearity and reproducibility Typically JW many scans merged Results Raw data with the first neighbor signal superimposed Bottom image is the fit to the FF c data Very accurate bond distances Strain is distributed in 2 nd and 3 rd neighbors, not so much in the NN • JW often uses 50 -75 mm diameter • wafer substrates with wide beam spot

Appl. Phys. Lett. , Vol. 73, No. 9, 31 August 1998 Goals and Samples Appl. Phys. Lett. , Vol. 73, No. 9, 31 August 1998 Goals and Samples Investigate nature of local distortions in buried layers (In. P caps) Important semiconductor materials Samples are in some cases 5 -6 nm MOCVD processed Experimental Fluorescent yield Spinning samples, wobble In-plane and glancing incidence Results Raw data with best fit of NN bonds

Findings In-As and Ga-As bonds are measured to a high accuracy Epitaxial films experience Findings In-As and Ga-As bonds are measured to a high accuracy Epitaxial films experience tetragonal distortions except for x=0. 53 sample which had a perfect lattice match to the substrate Noteworthy Large substrates Can this be done on small substrates …probably not Appl. Phys. Lett. , Vol. 73, No. 9, 31 August 1998

Appl. Phys. Lett. 68 (15), 8 April 1996 Experimental details Ion beam sputtered Fe Appl. Phys. Lett. 68 (15), 8 April 1996 Experimental details Ion beam sputtered Fe 80 B 20 Metglass™ films 15 nm thick Annealed at different temperatures and different times EXAFS measurements on X 23 B NSLS Sample size is 5 mm x 5 mm TEY, room temp, std. conditions Spot size ~1 mm x 1 mm No FEFF modeling Linear modeling of standards (poor mans PCA) Goal Determine phases present Determine atomic fraction of each Structural stability Onset of crystallization Error bars determined by background removal and counting statistics

Results Mass of sampled data: ~1 nanogram DSC and DTA have 10 -3 -10 Results Mass of sampled data: ~1 nanogram DSC and DTA have 10 -3 -10 -4 gram sensitivity Standards’ data are fit linearly with the following adjustable parameters: -atomic fraction of each phase -Debye-Waller coefficient of the amorphous phase Error bars Determined by fitting the mean, +/- standard deviations Systematic variation of atom fractions until c 2 doubles (this is larger) Counting statistics (10 scans of each sample) result in mean and standard deviations Findings Onset of crystallizations is 200 K below bulk samples Effective technique for the accurate measurement of phases I wish we could have…. In situ measurements would allow for the calculation of activation energies (in situ heating and film growth) The stability of amorphous phases in small volumes are important for understanding the limits of phase change media Appl. Phys. Lett. 68 (15), 8 April 1996

IEEE TRANSACTIONS ON MAGNETICS. VOL. 29. NO. 6. 1993 After 8 minutes the Fe IEEE TRANSACTIONS ON MAGNETICS. VOL. 29. NO. 6. 1993 After 8 minutes the Fe is fully crystallized as DO 3 (bcc Fe. Si ordered) phase of ~12 -15 nm. Experimental Melt spun ribbon samples having excellent soft magnetic properties Amorphous as spun Crystallized by annealing at 550 o. C (opt. T) for various times EXAFS transmission, RT Goals Does the 1% Cu make a difference and how? On the other hand, the Cu atoms never grow beyond clusters that are ~ 1 nm. Clusters have an close packed (fcc) symmetry

The role of 3% Nb Appt. Phys. Lett. 64 (8), 21 February 1994 3% The role of 3% Nb Appt. Phys. Lett. 64 (8), 21 February 1994 3% Nb EXAFS details Ribbons: 25 mm thick, 2 mm wide, 20 mm long Problems with sample uniformity Mapping the transmitted beam we find regions of uniform thickness that are about 300 mm x 1000 mm Fe data is easy but of little interest Cu data is challenging The beam walks a little due to large moment arm EXAFS analysis is largely examining Fourier transforms 0% Nb Findings The presence of Nb limits the nucleation and growth of the Cu atoms Without Nb, the Cu is fully crystallized after 2 minutes at temperature The Cu atoms and clusters act as nucleation sites for the DO 3 Fe. Si grains How to improve measurements… Small spot, more light, stable beam position In situ heating Sometimes samples are what they are….

Goals The goal is to gather information on both the element specific cation distribution Goals The goal is to gather information on both the element specific cation distribution and valency This is true for the manganites as well Magnetic and electronic properties Partial Radial Distribution Function Fourier transformed Fe EXAFS from Fe 3 O 4 standard illustrating pair correlations leading to A- & B-site fingerprinting IEEE Trans. Magn. 31, 6, Nov. 1995 Experimental details Samples are ferrite films prepared by SSP or PLD Size is typically 3 -5 mm on a side Thickness is typically 100 -200 nm EXAFS: FY or TEY

Results As shown the peaks from 2 -4 Å experience systematic changes FEFF 6: Results As shown the peaks from 2 -4 Å experience systematic changes FEFF 6: Multiple scattering fits without FEFFIT First determination of cation site distribution in ferrites (one step better than ME) Findings allow interpretation of M vs y trends with increasing Zn These results have provided a methodology to correlate SRO properties to fundamental magnetic properties in the ferrites What we really need is… Site specific, element specific, cation distribution and valency… Appl. Phys. Lett. 68 (15), 8 April 1996

Other ferrite systems studied by NRL and NEU researchers since 2000 NP Mn-ferrite (polyol) Other ferrite systems studied by NRL and NEU researchers since 2000 NP Mn-ferrite (polyol) C. N. Chinnasamy, et al. , J. Appl. Phys. 101, 09 M 509 (2007). PLD Cu-ferrite Aria Yang, et al. , Appl. Phys. Lett. , 86, 252510 (2005). BM NP Zn-ferrite S. A. Oliver, V. G. Harris, H. Hamdeh and J. C. Ho, Appl. Phys. Lett. , 76 (19), 2761 (2000). NP Mn. Zn-ferrite (reverse micelle) S. Calvin, E. E. Carpenter, B. Ravel, V. G. Harris, S. A. Morrison, Appl. Phys. Lett. , 81(20), 3828 (2002). Sc-doped Ba-hexaferrite A. Yang, Z. Chen, C. Vittoria, and V. G. Harris, J. Appl. Phys. , accepted, 2008. PLD Ni-ferrite C. N. Chinnasamy, et al. , J. Appl. Phys. , 101, 09 M 517 (2007). PLD and ATLAD Mn-ferrite Aria Yang, et al. , IEEE Trans. on Magnetics, 42 (10), 2870 (2006); Aria Yang, et al. , IEEE Trans. On Magn. , vol. 40, pp. 2802 -2805, 2004; Xu Zuo, et al. , Appl. Phys. Lett. , 87, 152505 (2005) NP Co-ferrite (reverse micelle) Sichu Li, et al. , IEEE Trans. On Magn. , vol. 37, No. 4, 2350 -2352 (2001); Sichu Li, et al. , J. Appl. Phys. , 87 (9), 6223 (2000). ATLAD Cu-ferrite Aria Yang, et al. , J. Appl. Phys. , accepted, 2008. NP Co-ferrite K. Giri, E. M. Kirkpatrick, P. Moongkhamklang, S. A. Majetich, and V. G. Harris, Appl. Phys. Lett. 80, 2341 (2002); PLD Co-ferrite G. Hu, V. G. Harris, and Y. Suzuki, IEEE Trans. On Magn. , vol. 37, No. 4, 2347 -2349 (2001); G. Hu, T. K. Nath, C. B. Eom and V. G. Harris, Y. Suzuki, Phys. Rev. B (Rapid Comm. ), 62(2), R 779 (2000).

Nanoparticle ferrites: Opportunities and Challenges Technologies are varied, from biosensor, cancer remediation therapies, MRI Nanoparticle ferrites: Opportunities and Challenges Technologies are varied, from biosensor, cancer remediation therapies, MRI contrast agents, laser printer toner, radar absorbing materials (stealth) Synthesis Nano-particles are formed in the general sequence of nucleation, growth; Ostwald ripening; Aggregation/Agglomeration/Sinteri ng/Coalescence (La Mer’s Law) Experimental Fe. Cl 3. 6 H 2 O =0. 1 M Mn. Cl 2. 4 H 2 O = 0. 05 M Boiling Na. OH Washing

Magnetism and Microstructure studies Fig. TEM micrographs of Mn. Fe 2 O 4 nanoparticles Magnetism and Microstructure studies Fig. TEM micrographs of Mn. Fe 2 O 4 nanoparticles synthesized by using [OH-] molars of (a) 4 M, (b) 2 M, (c) 1 M and (d) 0. 425 M. • The Néel temperature is found to increase with particle size • The Néel temperature is significantly higher (400 C) than the bulk Mn. Fe 2 O 4 (300 C). • Contradiction to earlier reports that indicated an increase in TN with reducing particle size and attributed to a finite size scaling. • The small particles are spherical like morphology and the larger particles consist of a mixture of both cubic and spherical particles. • The particle size increased while we decreased the molar concentration of Na. OH solution C. N. Chinnasamy, et al. , J. Appl. Phys. , 101, 09 M 509 (2007)

EXAFS data, fitting model and Fourier transform fitting results EXAFS Constraints • Stoichiometry Constraint: EXAFS data, fitting model and Fourier transform fitting results EXAFS Constraints • Stoichiometry Constraint: • Multiple scattering paths treatments (Normalized) Guess Parameters: Calculated: • Element dependence of first coordination shells • Constraints for Nanoparticles r: distance to the coordination shell in question R: radius of the particle S. Calvin, et. al. , Phys. Scr. , T 115, 744 (2005) FIG. Real part of Fourier transform of EXAFS data and best fits of manganese edge, and iron edge, for both 4 nm and 50 nm samples. The data was Fourier transformed with the k range of 2. 6 -11. 5Å-1, and k weight of 3.

Table. Results determined from the EXAFS best fits Average particle sizes (diameter) 4 nm Table. Results determined from the EXAFS best fits Average particle sizes (diameter) 4 nm 7. 5 nm 25 nm 50 nm Ceramic standard Lattice parameter (Ǻ) 8. 421(2) 8. 400(2) 8. 387(2) 8. 383(2) 8. 5 Oxygen parameter 0. 388(2) 0. 387(1) 0. 385(1) 0. 3846 EXAFS R factors 0. 0237 0. 0267 0. 027 0. 0199 --- d 62. 91%± 1. 84% 57. 83± 2. 09% 55. 68± 1. 74% 51. 07± 1. 93% 20% Nanoparticle samples have reduced lattice parameters in comparison to the bulk value of 8. 5 Ǻ. Contraction -> Increase of super exchange interaction -> Increase of Néel temperature Results Neel temperature reflects the strength of the exchange Neel temperature increases with particle size but inversely with quench rate What is needed in NP research? Insight into nucleation and growth In situ growth (Mn 1−d. Fed)tet[Mnd. Fe 2 -d]oct. O 4

Automated NP chemical processor Objective: Allowing identical samples to be synthesized on demand in Automated NP chemical processor Objective: Allowing identical samples to be synthesized on demand in synchrotron and the institution, so that the characterization resources of the home institution’s laboratory can be applied immediately to samples synthesized at the beamline. Scott Calvin, Sarah Lawrence College ([email protected] edu); Everett E. Carpenter, Virginia Commonwealth University Department of Chemistry

Argonaut Surveyor™ Chemical Processor Configuration of system tested at NSLS A robotic arm can Argonaut Surveyor™ Chemical Processor Configuration of system tested at NSLS A robotic arm can transfer samples and reagents under inert atmospheres x-ray hutch B C E F G A Monochromator nine sources of reagents ten 25 -m. L reaction vessels temperature controllable from -80°C to 150°C D H E Argonaut processor B Incident x-ray detector F Waste collection C Flow cell G Control computer D Lytle detector H Beamline computer Configuration figure adapted from S. Calvin et al. , Rev. Sci. Instrum. 76, 016103 (2005).

nylon screws Custom flow cell O-ring nylon screws Kapton window bores (offset to ensure nylon screws Custom flow cell O-ring nylon screws Kapton window bores (offset to ensure new aliquot displaces old) Kapton window Flow cell figures adapted from S. Calvin et al. , Rev. Sci. Instrum. 76, 016103 (2005).

Time-resolved mode 172 7. 5 7. 0 6. 5 6. 0 5. 5 4. Time-resolved mode 172 7. 5 7. 0 6. 5 6. 0 5. 5 4. 5 3. 5 4. 0 3. 0 2. 5 2. 0 1. 5 129 86 Time (min. ) 43 0 R (Å) eff Example of time-resolved EXAFS data at the iron edge collected at X 23 B using the Surveyor™ and custom flow cell. Figure adapted from S. Calvin et al. , Rev. Sci. Instrum. 76, 016103 (2005). Details of reaction are given in that reference.

Time-resolved mode Oxygen bonding Cation filling Too complicated to be of value (for now) Time-resolved mode Oxygen bonding Cation filling Too complicated to be of value (for now) 172 7. 5 7. 0 6. 5 6. 0 5. 5 4. 5 3. 5 4. 0 3. 0 2. 5 2. 0 1. 5 129 86 Time (min. ) 43 0 R (Å) eff Example of time-resolved EXAFS data at the iron edge collected at X 23 B using the Surveyor™ and custom flow cell. Figure adapted from S. Calvin et al. , Rev. Sci. Instrum. 76, 016103 (2005). Details of reaction are given in that reference.

Oxygen bonding Time-resolved mode Too slow! Little information to be learned in the area Oxygen bonding Time-resolved mode Too slow! Little information to be learned in the area of Cation filling nucleation. Yes, we can slow the kinetics down some, but then we are not duplicating Too complicated to laboratory conditions. be of value (for now) 172 7. 5 7. 0 6. 5 6. 0 5. 5 4. 5 3. 5 4. 0 3. 0 2. 5 2. 0 1. 5 129 86 Time (min. ) 43 0 R (Å) eff Example of time-resolved EXAFS data at the iron edge collected at X 23 B using the Surveyor™ and custom flow cell. Figure adapted from S. Calvin et al. , Rev. Sci. Instrum. 76, 016103 (2005). Details of reaction are given in that reference.

AT-LAD grown Mn. Fe 2 O 4 films Objective: Allowing identical samples to be AT-LAD grown Mn. Fe 2 O 4 films Objective: Allowing identical samples to be synthesized on demand in synchrotron and the institution, so that the characterization resources of the home institution’s laboratory can be applied immediately to samples synthesized at the beamline. Scott Calvin, Sarah Lawrence College ([email protected] edu); Everett E. Carpenter, Virginia Commonwealth University Department of Chemistry

Properties of AT-LAD grown films X. Zuo et al. , J. Appl. Phys. 97, Properties of AT-LAD grown films X. Zuo et al. , J. Appl. Phys. 97, 10 G 103 (2005)

1 m. Torr 50 m. Torr Percentage of Fe at A sites (calculated) 9% 1 m. Torr 50 m. Torr Percentage of Fe at A sites (calculated) 9% 18% Percentage of Mn at B sites (i. e. Inversion parameter) 18% (4) 35% (4) Lattice Parameter 8. 570(4) 8. 602(2) Oxygen Parameter 0. 392 (1) 0. 3948(8) Debye-Waller factor 0. 010(2)

Inversion parameters • 4 p. Ms decreased as Inversion parameter (percentage of Mn ions Inversion parameters • 4 p. Ms decreased as Inversion parameter (percentage of Mn ions at B sites) increased, which is similar with Cu-ferrite system

X-ray absorption spectra U 4 B, NSLS, BNL • Sample processed above Pox>5 m. X-ray absorption spectra U 4 B, NSLS, BNL • Sample processed above Pox>5 m. Torr more oxidized than samples processed less than 2 m. Torr. • Mn 3+: 4 m. B • Mn 2+: 5 m. B • Fe 3+: 5 m. B • over oxidized Mn ions at B site could reduce magnetization • Could Mn and Fe have different valence at different sites?

DAFS Basics • Diffraction Anomalous Fine Structure A Feature Of Resonant Elastic Scattering Observed DAFS Basics • Diffraction Anomalous Fine Structure A Feature Of Resonant Elastic Scattering Observed Above Absorption Edges in the Energy Dependence of : – Bragg Peaks • DAFS can be used to isolate the fine structure from a subset of resonant atoms based on their long-range order XAFS µ(E) DAFS I(E) f’(E) + if’’(E)

Experimental Setup • Do an XAFS-like energy scan with the sample and the detector Experimental Setup • Do an XAFS-like energy scan with the sample and the detector mounted on a diffractometer – Track the Bragg peak using software or feedback • Advantages: Atom specific, site specific • Bragg reflection chosen: – <222> ~ B sites – <422> ~ A sites – <111> ~ combination of A & B sites (for self-consistency check) – EXAFS Fluorescence (for selfconsistency check)

Table. Oxidation charge of Mn and Fe at specific site for both ATLAD Mn. Table. Oxidation charge of Mn and Fe at specific site for both ATLAD Mn. Fe 2 O 4 films 1 m. T 50 m. T [email protected] +2. 18(5) +2. 31(7) [email protected] +3. 38(8) +3. 42(9) Mn avg. 2. 39 2. 70 [email protected] +3. 35(8) +2. 40(9) [email protected] +2. 92(9) +3. 10(4) Fe avg. 2. 9597 2. 979 Mn edge, 1 m. T Mn edge, 50 m. T

Table. First oxygen neighbor bond distance of Mn and Fe at specific site for Table. First oxygen neighbor bond distance of Mn and Fe at specific site for both ATLAD Mn. Fe 2 O 4 films 1 m. T 50 m. T Standard [email protected] 2. 06(1) 1. 96(3) 1. 98 [email protected] 1. 98(3) 2. 02(3) 2. 05 [email protected] 1. 96(2) 1. 93(4) 1. 98 [email protected] 2. 02(3) 2. 02(5) 2. 05 Figure. The EXAFS FT data and the fit to theoretical standards of the Mn (422) and (222) K edge of the 1 m. Torr sample.

1 m. T 50 m. T Bulk SEM-EDX measured stoichiometry, Mn: Fe ratio 1: 1 m. T 50 m. T Bulk SEM-EDX measured stoichiometry, Mn: Fe ratio 1: 2 7: 13 1: 2 4 p. Ms (G) (20 o. C) 4547 2251 50001 4 p. Ms (G) (4 K) 7212 4321 70001 Predicted 4 p. Ms (G) (0 K) 7447 6837 Net Charge (Mn and Fe ions) +8. 31 +8. 64 +8 These parameters, and nothing short, are needed as input for first principles and MF theories This is underway J. Smit and H. P. J. Wijn, Ferrites, John Wiley & Sons, Philips Technical Library, 157 (1959)

What are the central issues in NSLS-II EXAFS beamlines • Sensitivity to sub nanometer What are the central issues in NSLS-II EXAFS beamlines • Sensitivity to sub nanometer thickness is absolutely essential • These thin layers might be buried • The available surface area may be 4 -6 mm 2 • Dynamics are essential – – – Film growth NP synthesis Magnetic fields Electric fields Lasers What else…. . ?