QENS in the Energy Domain: Backscattering and Time-of of-flight Alexei Sokolov Department of Polymer Science, The University of Akron Outline Soft Matter and Neutron Spectroscopy Using elastic scattering and employing H/D contrast Quasielastic scattering spectra, susceptibility presentation Q-dependence: diffusive vs local processes Geometry of the motion from EISF Use of coherent scattering Spectrometers Soft Matter Characteristics of Soft Materials: -Variety of states and large degree of freedom, metastable states; -Delicate balance between Entropic and Enthalpic contributions to the Free Energy; -Large thermal fluctuations and high sensitivity to external conditions; -Macroscopic softness. Polymers Colloids Liquid Crystals Foams and Gels Biological Systems 1
Frequency map of polymer dynamics Chain dynamics: Rouse and terminal relaxation Structural relaxation, α-process Secondary relaxations Fast relaxation Vibrations Mechanical relaxation G (ν) Dielectric Spectroscopy, ε*(ν) Quasi-optics,TDS Traditional dielectric spectroscopy IR-spectr. FREQUENCY, ν 10-1 10 10 4 10 6 10 8 10 10 10 1 10 14 Hz Light Scattering, I ij (Q,ν) Photon Correlation Spectroscopy Interferometry Raman spectroscopy Neutron Scattering, S(Q,ν) Spin-Echo Back-sc. Time-of-Flight Inelastic X-ray Scattering, S(Q,ν) High-Resolution IXS Scattering techniques have an advantage due to additional variable wave-vector Q Beauty of Neutron Spectroscopy Frequency [Hz] 10 13 10 1 10 11 10 10 10 9 10 8 10 7 10 6 10 3 10 10 1 10 0 10-1 Light Scattering Length [nm] 10-10 -1 10 0 10 1 10 Q [nm -1 ] Inelastic X-Ray Neutron Scattering 10-14 10-13 10-1 10-11 10-10 10-9 10-8 10-7 10-6 Time [s] Measures characteristic times (frequency) and geometry of the motions. Covers broad frequency and Q-range in the most important for microscopic dynamics region. Current X-ray technology cannot compete! Most of the soft materials contain hydrogen atoms, use of D/H contrast. Direct comparison to results of MD-simulations.
Using H/D Contrast An example of elastic scan measurements of PI [Frick, Fetters, Macromol. 7, 1994]. Decrease of elastic intensity marks onset of a relaxation process. Various deuteration of the polymer allows separate methyl group and main-chain motion. The onset of methyl groups rotation at temperatures below Tg is clearly seen. PI-h8 all H, PI-d8 with all D. PI-d5 PI-d3 PI d3 PI d5 Methyl Group Dynamics in Proteins Significant part of H-atoms in proteins are on methyl groups. I el (Q,T)/I el (Q,10K)/DW 1.00 0.95 0.90 0.85 0.80 E 0 ~16.6kJ/mol, E~5.8kJ/mol E~16.6 kj/mol, E~0 Q=1.6 A -1 0.75 0 50 100 150 00 50 300 T [K] Decrease of the elastic intensity in dry lysozyme can be described assuming a Gaussian distribution of energy barriers, g(e i ) exp[-(e i -E 0 ) / E ], with E 0 ~16.6 kj/mol and E~5.8 kj/mol in good agreement with earlier NMR data [J.H.Roh, et al. Biophys.J. 91, 573 (006)]. I el τ i T, ω ~ 0) = DW T) 1 pm + pm Smet ω') R( ω ω' ) dω' DW T) const( + R( ω ω') = 0 g( E ) de dω' ω i i ω= 0 1+ ω' τ Here τ i =τ 0 exp(e i /kt) 0 i 3
Mean-squared Displacements <r > In rough approximation, for an isotropic motion: Q S inc exp < r( > 3 This approximation works well only at low Q. The estimated <r > depends on the selected Q-range and the resolution function of the spectrometer. 1.5 Lysozyme in D O WET 1.00 <r > [Å ] 0.75 0.50 0.5 DRY 0.00 0 50 100 150 00 50 300 T [K] Analysis of <r > helps to identify interesting temperature ranges. However, <r > is an integrated quantity (includes vibrations, rotation, diffusion, etc.) and analysis of spectra is required for understanding the dynamics. Quasielastic Scattering Spectrum Usual approximation is a Lorentzian function: Γ( S( Q, E) E + Γ( In most cases or more Lorentzians are used for the fit of the spectra. This approximation assumes single exponential relaxation: S exp( t / τ ) However, many relaxation processes in soft matter are strongly stretched S exp t [ ( /τ ) ] β S(Q,ν) [arb.un.] 1000000 100000 10000 1000 Relaxations (quasielastic scattering) 90K 10K DNA/D O, 75% r.h. Vibrations (inelastic scattering) 0 500 1000 1500 000 500 3000 ν [GHz] So, approximation by Lorentzians can give misleading quantitative results 4
Susceptibility presentation Susceptibility presentation of scattering spectra has a few advantages: - can be directly compared to ε (ν), G (ν); - each relaxation process appears as a maximum at πντ~1; - slopes of the tails give estimate of stretching exponents. χ"(q,ν)=s(q,ν)/n B (ν) [arb. un.] 10-3 single Lorentzian 10-4 Slow Process Fast relaxation BS data 95K 150K TOF data 10 0 10 1 10 10 3 ν [GHz] Susceptibility spectra of wet lysozyme The spectra of proteins show two relaxation processes. Both processes are strongly stretched (can not be described by a single exponential relaxation). Q-dependence: Diffusion For regular diffusion: r( Dt In frequency domain: In that case: N N Γ S inc exp( Q D = exp( Γ S inc ( ω, = exp( iω exp( Γ t ) dt = π π Γ + ω 1.0 0.8 Q S(Q,ω) Q 1 > Q S(Q, 0.6 0.4 Q 1 Q 0. Q 1 0.0 t ω An exponential decay for S(Q,, with decay rate Γ Q 0 In the case of sub-diffusive regime: with Γ Q /β. r( β ( D => S( Q, exp Q ( D β β [ ] exp[ ( Γ ] Diffusion-like motions exhibit strong dependence of the decay rate Γ (or relaxation time τ 1/Γ) on Q. 5
Q-dependence: a Local Relaxation Process S inc Let s assume that there are two equal positions and molecule makes jumps between r 1 and r positions. In isotropic case: [ EISF( + A ( exp( t / )]; 1 τ d = r r sin Qd sin Qd EISF( = (1/ ) 1 + ; A1 ( q) = (1/ ) 1 Qd Qd EISF( is the Elastic Incoherent Structure Factor. It contains information on geometry of the motion. In the frequency domain: S inc ω) S(Q,ω) 1 τ N EISF( δ ( ω) + A1 ( π 4 + ω τ = 1 S(Q, 1.0 0.8 0.6 0.4 0. 0.0 EISF A 1 Qd~3π/ Q 0 ω For a local relaxation process: S(Q,ω) has two component elastic and quasielastic; Characteristic time scale τ (or Γ) is independent of Q (at least, at large. EISF in dry protein: Methyl Group Dynamics Analysis of elastic incoherent structure factor, EISF(=I el (/[I el (+I QES (], can be done: (i) assuming a single exponential relaxation (single Lorentzian); (ii) taking into account a distribution of τ i or energy barriers g(e i ): S τ i ' ω) const( + R( ω ω' ) g( Ei ) de dω i [ 1 j ( 3) ] pm EISF( = 1 pm + + 0 QR 3 0 1+ ω' τ i EISF HFBS ( 0.0 0.3 0.6 0.9 1. 1.5 1.8 Analysis of the first data set (single Lorentzian) gives mobile fraction of H-atoms p m =0.14 and radius R~1.3 A, while analysis of the second set gives p m =0.5 and radius R~1.3 A. For methyl groups R~1.1 A and p m =0.6 in lysozyme [J.H.Roh, et al. Biophys.J. 91, 573 (006)]. 1.0 0.9 0.8 Single Lorentzian Distribution of τ Dry Lysozyme T=95K Q [Å -1 ] The first approximation overestimates EISF. Fit of the EISF to a 3-site jump model [J.H.Roh, et al. Biophys.J. 91, 573 (006)]: 6
Segmental and Secondary Relaxations in Polymers Homogeneous vs Heterogeneous Dynamics a) Heterogeneous: Normal diffusion with distribution of diffusion coefficient D: β [ ] 1 ( Q D d( ln D ) exp ( Q D 1 S = g(ln D )exp τ Q - Segmental relaxation time τ S exhibits strong Q-dependence, τ S Q -/β, indicating stretched diffusive-like process (β - KWW stretching parameter). b) Homogeneous: Sublinear diffusion in time, <r (> t β : β ( Q r ( / 6) exp[ Q ( D ] S = exp τ Q -/β Colmenero, et al., J.Phys.Con.Matter 11, A363 (1999). Polybutadiene (PB): Split of Segmental and Secondary Relaxations segmental secondary Dielectric data Q dependence of τ self change sharply when T approaches ~00 K. Also scaling with the viscosity time scale τ η fails. This behavior is ascribed to the split of segmental and secondary (local) relaxations. S. Kahle, et al. Appl.Phys. A 74, S371 (00) 7
Coherent Scattering inter-molecular Polybutadiene intramolecular NSE data measured at different T for deuterated PB scaled with the viscosity time scale τ η : -Master curve for the data measured at Q~1.5A -1 ; -No master curve for the data at Q~.7A -1. Q=1.5A -1 Conclusions: Segmental relaxation involves inter-molecular motions; Secondary relaxation involves intra-molecular motion, rotation about the double-bond. Q=.7A -1 A. Arbe, et al. PRE 54, 3853 (1996). Instruments: Back-Scattering Spectrometer HFBS Wavelength Neutron Energy Neutron Flux at Sample Energy range Energy resolution at ± 36 µev Analyzer Span 6.71 Å.08meV 3 x 10 5 n cm - s -1 ± 36 µev About 1 µev 165 Q range 0.5 Å -1 1.75 Å -1 8
Instruments: Back-Scattering Spectrometer HFBS Instruments: TOF Spectrometer DCS The DCS is a direct geometry time-of-flight spectrometer, the only instrument of its kind in North America. The DCS is primarily used for studies of low energy excitations and diffusive motions in a wide variety of materials. The DCS is an extremely versatile instrument. Useful incident wavelengths range from < Å to at least 9Å; correspondingly the elastic energy resolution (FWHM) varies from ~1500 to ~15eV. 9
Instruments: TOF Spectrometer DCS For any experiment try to optimize intensity vs resolution. Conclusions Neutron Spectroscopy is well positioned for analysis of dynamics of Soft Materials. Analysis of elastic scattering and use of H/D contrast allows to identify molecular units involved in the motion, geometry of the motion and interesting temperature ranges. Analysis of the Q-dependence differentiate diffusive and local processes and provide additional information on geometry of molecular motions. Analysis of the energy-resolved spectra provides information on characteristic relaxation times and vibrational frequencies, their distribution and temperature dependence. Coherent scattering provides additional information on cooperativity and geometry of molecular motion. However, analysis of the coherent scattering is more complex than analysis of incoherent scattering. 10
Hands-on Exercises Using DAVE program and provided experimental data (3 sets of data) perform the following tasks: Mean-squared displacement <r > in dry protein (HFBS data from J.H.Roh, et al. Biophys.J. 91, 573 (006)): -Analyze temperature dependence of <r > using HFBS data from elastic scan (Doppler stopped). QENS spectrum of dry protein (HFBS data from J.H.Roh, et al. Biophys.J. 91, 573 (006)): -Analyze Q-dependence of the characteristic relaxation time (decay rate); -Analyze EISF( (assuming Lorentzian spectrum). QENS spectrum of water of polypeptide hydration (DCS data from D. Russo, et al. J.Phys.Chem. B 109, 1966 (005)): - Analyze Q-dependence of characteristic relaxation time (decay rate) 11