Presentation is loading. Please wait.

Presentation is loading. Please wait.

Tony Leggett Department of Physics

Similar presentations


Presentation on theme: "Tony Leggett Department of Physics"โ€” Presentation transcript:

1 Quasiparticles in Normal and Superfluid Fermi Liquids (more questions than answers โ€ฆ)
Tony Leggett Department of Physics University of Illinois Urbana-Champaign The David Pines Symposium on Superconductivity Today and Tomorrow Urbana, IL 30 March 2019

2 Quasiparticles in Normal Phase
Landau (1956), Noziรจres, Theory of Interacting Fermi Systems (1964): Start from noninteracting Fermi gas, then energy eigenstates specified by ๐‘› ๐’‘,๐œŽ ,๐‘› ๐‘,๐œŽ =0 or 1 groundstate has ๐‘› ๐’‘,๐œŽ =ฮ˜ ๐‘ ๐น โˆ’ ๐’‘ , ๐‘ ๐น =โ„ 3 ๐œ‹ 2 ๐‘› 1/3 switch on inter-particle interaction ๐‘‰ adiabatically: ๐‘‰ ๐‘ก = ๐‘‰ exp ๐›ผ๐‘ก ๐‘ก<0, ๐›ผโ†’0 ๐‘‰ ๐‘ก=0 = ๐‘‰ provided perturbation theory, converges, states of fully interacting system can be labelled by the noninteracting states ๐‘› ๐‘,๐œŽ from which they evolved. then define โ€œno. of quasiparticles in state ๐‘,๐œŽโ€ as n ๐’‘,๐œŽ (๏ƒž Luttinger theorem trivial) ๐‘„ = ๐‘๐œŽ ๐‘ž ๐‘๐œŽ ๐›ผ ๐‘๐œŽ + ๐›ผ ๐‘,๐œŽ Suppose ๐‘„ = ๐‘๐œŽ ๐‘ž ๐‘๐œŽ ๐‘› ๐‘,๐œŽ . then in original noninteracting system ๐‘„ = ๐‘๐œŽ ๐‘ž ๐‘๐œŽ ๐‘› ๐‘,๐œŽ ? Is it true that in fully interacting system also Answer: yes, if and only if ๐‘„ , ๐ป ๐‘ก =0

3 ๐ฝ ๐œŽ โ‰  ๐‘๐œŽ ๐‘ ๐‘š ๐œŽ๐‘› ๐‘๐œŽ What quantities are conserved ๐‘„ , ๐ป ๐‘ก =0 ?
(a) liquid 3 ๐ป๐‘’ : ๐‘ = ๐‘๐œŽ ๐›ผ ๐‘๐œŽ + ๐›ผ ๐‘๐œŽ yes total number ๐‘† โ‰ก ๐‘๐œŽ ๐œŽ ๐›ผ ๐‘๐œŽ + ๐›ผ ๐‘๐œŽ yes total spin ๐‘ฑ = ๐‘š โˆ’1 ๐‘๐œŽ ๐‘ ๐›ผ ๐‘๐œŽ + ๐›ผ ๐‘๐œŽ yes total current ๐‘ฑ ๐œŽ โ‰ก ๐‘š โˆ’1 ๐‘๐œŽ ๐‘๐œŽ ๐›ผ ๐‘๐œŽ + ๐›ผ ๐‘๐œŽ no total spin current ๏ƒž in real liquid 3 ๐ป๐‘’ , ๐‘† = ๐‘๐œŽ ๐œŽ๐‘› ๐‘๐œŽ (etc.) ๐ฝ ๐œŽ โ‰  ๐‘๐œŽ ๐‘ ๐‘š ๐œŽ๐‘› ๐‘๐œŽ (b) metallic system (e.g. cuprates) ๐‘,๐‘† conserved but ๐‘ฑ not conserved (even after transformation to Bloch states, because of U-processes).

4 ๐œ”= v ๐น ๐‘˜๏‚ฏ ๐ผ๐‘š ๐‘๐‘ฮ› ๐‘˜๐œ” v ๐น ๐‘˜ ๐œ”โ†’ extends to ฯ‰โ‰ซv ๐น ๐œ… ๏‚ฏ v ๐น ๐œ… ๐œ”โ†’
Consequences of conservation for response functions Consider ๐œ’ ๐‘„๐‘„ ๐‘˜,๐œ” โ‰ก๐น๐‘‡ of โ‰ช ๐‘„ 0,0 ๐‘„ ๐‘Ÿ๐‘ก โ‰ซ If ๐‘„ is conserved, then in limit ๐‘˜โ†’0, support of ๐ผ๐‘š ๐œ’ ๐‘„๐‘„ ๐‘˜๐œ” comes entirely from quasiparticle states and is limited to ๐œ”โ‰ฒ v ๐น ๐‘˜ Ex: ๐‘„ = ๐‘ (density response function) of liquid 3 ๐ป๐‘’ ๐œ”= v ๐น ๐‘˜๏‚ฏ ๐ผ๐‘š ๐‘๐‘ฮ› ๐‘˜๐œ” ๏‚ฌzero-sound peak v ๐น ๐‘˜ ๐œ”โ†’ If ๐‘„ is not conserved, e.g. ๐‘„ = ๐ฝ ๐œŽ in 3 ๐ป๐‘’ , quasiparticle states do not exhaust sum rule for ๐œ’ ๐‘„๐‘„ ๐‘˜๐œ” and even in limit kโ†’0 get incoherent background. extends to ฯ‰โ‰ซv ๐น ๐œ… ๏‚ฏ ๐ผ๐‘š ๐œ’ ๐‘„๐‘„ ๐‘˜๐œ” v ๐น ๐œ… ๐œ”โ†’ In liquid 3 ๐ป๐‘’ , can infer quasiparticle contribution from Landau parameter ๐น ๐‘Ž 1 (measurable in spin-echo experiments). Conclusion: In 3 ๐ป๐‘’ , incoherent background contributes >80% of sum rule!

5 Moral: Even if system is a โ€œdecentโ€ Fermi liquid, correlation function of non-conserved quantity can have large contribution from incoherent background. Application to cuprates (and maybe other SCES): ๐‘,๐‘† conserved but ๐‘ฑ not conserved (because of U-processes) sum rule for ๐œ”๐ผ๐‘š ๐œ’ ๐ฝ๐ฝ ๐œ” =๐‘…๐‘’ ๐œŽ ๐œ” can have large contribtion from incoherent background (MIR peak). Are the optimally doped and underdoped cuprates โ€œbadโ€ Fermi liquids? (cf. e.g. Berthod et al., PR B 87, (2013)).

6 energy relative to Fermi energy
QUASIPARTICLES IN THE SUPERFLUID STATE 1. Andreev reflection ฮ” 0 ฮ” ๐‘ง ๐œ€ ๐‘˜ ? e.g.: > energy relative to Fermi energy ๐‘ง๏‚ฎ ๏‚ฌ ๐ฟโ‰ซ ๐‘˜ ๐น โˆ’1 ๏‚ฎ Incident particle, 0< ๐œ€ ๐‘˜ < ฮ” 0 for normal reflection on needs ฮ”๐‘˜~2 ๐‘˜ ๐น so amplitude ~expโˆ’2 ๐‘˜ ๐น ๐ฟโ‰ช1, and Fermi sea blocked ๏ƒž reflected as hole: by conservation of energy ๐ธ ๐‘˜ = ๐œ– ๐‘˜ โ‡’ hole energy is โˆ’๐œ– ๐‘˜ . What is momentum transfer? ๐œ– ๐‘˜ ~โ„ ๐‘ฃ ๐น ๐‘˜โˆ’ ๐‘˜ ๐น โ‡’ ฮ”๐‘= 2 ๐œ– ๐‘˜ ๐‘ฃ ๐น โ‰ช ๐‘ ๐น (normal incidence) Is there direct experimental evidence for this? Yes! Buchanan et al., (PRL 57, 341 (1986) measure terminal velocity of 3 ๐ป๐‘’ ๐ดโˆ’๐ต interface. ๐‘ฃ ๐‘ก๐‘’๐‘Ÿ๐‘š =ฮ” ๐บ ๐ด๐ต /ฮ“ ๏‚ฌ frictional force due to reflection of qps If we assume reflection is โ€œnormalโ€, ๐‘ฃ ๐‘ก๐‘’๐‘Ÿ๐‘š โ‰ฒ1 mm/sec Experimentally, ๐‘ฃ ๐‘ก๐‘’๐‘Ÿ๐‘š ~ 0.1โˆ’1 m/sec ๏ƒž Andreev reflection (S-K Yip and AJL, PRL 57, 345 (1986))

7 2. The โ€œZeeman-dimpleโ€ problem
Note: spatial variation of gap ฮ” ๐‘ง not a necessary condition for AR! Andreev reflection Can alternatively result from spatial variation of โ€œdiagonalโ€ potential ๐‘‰ ๐‘Ÿ , provided this is the same for particle and hole (e.g. Zeeman potential โˆ’ ๐œ‡ ๐ต ๐œŽ๐ต ๐‘ง ) Ex*: neutral Fermi superfluid with Zeeman coupling to external field ๐ต ๐‘ง with โ€œdimpleโ€ ๐ต 0 โ‰ชฮ”/ ๐œ‡ ๐ต ๐ต ๐‘ง โ†‘ ๏‚ฌ๐ฟโ‰ซ ๐œ‰ 0 โ†’ Even (number) โ€“ parity ground state ฮจ 0 has ฮ” ๐‘ง =๐‘๐‘œ๐‘›๐‘ ๐‘กโ‰กฮ” to linear order in ๐ต. What is nature of lowest-energy odd-parity state? Answer: Single Bogoliubov quasiparticle trapped in โ€œdimpleโ€. Extra spin localized in/close to dimple = 1. *Y.-R. Lin and AJL, JETP 119, 1034 (2014)

8 ๐ธ โ„Ž๐‘œ๐‘™๐‘’ = ๐ธ ๐‘๐‘Ž๐‘Ÿ๐‘ก๐‘–๐‘๐‘™๐‘’ โ‡’ ๐œ– โ„Ž๐‘œ๐‘™๐‘’ =โˆ’ ๐œ– ๐‘๐‘Ž๐‘Ÿ๐‘ก๐‘–๐‘๐‘™๐‘’
What is extra charge? particle hole In quasiclassical approximation with only Andreev reflection: ๐ธ โ„Ž๐‘œ๐‘™๐‘’ = ๐ธ ๐‘๐‘Ž๐‘Ÿ๐‘ก๐‘–๐‘๐‘™๐‘’ โ‡’ ๐œ– โ„Ž๐‘œ๐‘™๐‘’ =โˆ’ ๐œ– ๐‘๐‘Ž๐‘Ÿ๐‘ก๐‘–๐‘๐‘™๐‘’ but in formula ๐œ“ ๐‘ž๐‘ = ๐‘ข ๐‘˜ ๐‘Ž ๐‘˜โ†‘ + โˆ’ ๐‘ฃ ๐‘˜ ๐‘Ž โˆ’๐‘˜โ†“ + | ฮจ 0 ๐‘ข ๐‘˜ = ๐œ€ ๐‘˜ / ๐ธ ๐‘˜ , ๐‘ฃ ๐‘˜ = โˆ’ ๐œ€ ๐‘˜ / ๐ธ ๐‘˜ so ๐œ€ ๐‘˜ โ†’โˆ’ ๐œ€ ๐‘˜ โ‡’ ๐‘ข ๐‘˜ โ‡„ ๐‘ข ๐‘˜ . Also, ๐œ ๐‘˜ = โ„ โˆ’1 ๐œ•๐ธ ๐‘˜ ๐œ•๐‘˜ ๐‘˜ =โ„ โˆ’1 ๐ธ ๐‘˜ / ๐ธ ๐‘˜ ๐œ• ๐ธ ๐‘˜ /๐œ•๐‘˜ , so to the extent that N -state spectrum is particle-hole symmetric, ๐œ• ๐œ– ๐‘˜ /๐œ•๐‘˜=โ„ ๐‘ฃ ๐น =๐‘๐‘œ๐‘›๐‘ ๐‘ก. , extra charge = 0

9 Further complication: in this approximation, ground state of odd-number-parity sector is doublet related by time reversal! hole particle particle hole โ‡” T โ€œNormalโ€ (non-Andreev) reflection splits doublet into even and odd combinations with exponentially small splitting. However, this does not change situation with regard to C-symmetry. ๏ƒž zero extra charge is not robust. (even in quasiclassical approximation)

10 3. Effect of taking particle number conservation seriously
With assumption of SBU(1)S ๏‚ฌ spontaneously broken U(1) symmetry standard formula for creation of Bogoliubov quasiparticle from even-parity groundstate | ฮจ 0 ~ ๐ถ ๐‘/2 | ๐‘ฃ๐‘Ž๐‘ is ๐œ“ ๐‘ข = ๐‘ข ๐‘˜ ๐‘Ž ๐‘˜โ†‘ + โˆ’ ๐œ ๐‘˜ ๐‘Ž โˆ’๐‘˜โ†“ | ฮจ 0 or more generally (BDG) Bogoliubov-de Gennes ๐œ“ ๐‘ข = ๐›พ โ€  | ฮจ 0 ๐›พ โ€  = ๐‘ข ๐‘Ÿ ๐œ“ โ€  ๐‘Ÿ +๐‘ฃ ๐‘Ÿ ๐œ“ ๐‘Ÿ ๐‘‘๐‘Ÿ This does not conserve particle number. Remedy: ๐›พ โ€  = ๐‘ข ๐‘Ÿ ๐œ“ โ€  ๐‘Ÿ +๐‘ฃ ๐‘Ÿ ๐œ“ ๐‘Ÿ ๐ถ ๐‘‘๐‘Ÿ ๏‚ญ creates extra Cooper Pair Question 1: Is the โ€œextraโ€ pair the same as those in the even-parity GS? Question 2: Irrespective of answer to 1, does it matter?

11 Conjecture: for โ€œusualโ€ case (e. g
Conjecture: for โ€œusualโ€ case (e.g. Zeeman-dimple problem), effect is nonzero but probably small. but for case where Cooper pairs have โ€œinterestingโ€ properties (e.g. intrinsic angular momentum) effect may be qualitative. The crunch case: Majorana fermions in (p+ip) Fermi superfluid (Sr2RuO4?): does extra Cooper pair change results of โ€œstandard theory (e.g. Ivanov 2001) qualitatively? -the $64K (actually $6.4M!) questionโ€ฆ


Download ppt "Tony Leggett Department of Physics"

Similar presentations


Ads by Google