Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.##...
Transcript of Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.##...
![Page 1: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/1.jpg)
Superconductivity in low-dimensional systems
Marco Grilli Dipar,mento di Fisica
Lecture 2: SC in low (actually 2) dimensions. What’s up with SC? Breaking news and old revisited stuff
Aim: show how the learned concepts and dogmas are used/challenged in present research
![Page 2: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/2.jpg)
Why (quasi) 2D superconductors are interes,ng and “hot”?
• Layered or 2D systems are rela,vely easy to manipulate: -‐they can be built and designed with various techniques (MBE, etching, ...) like Josephson-‐Junc,on arrays, thin films, ..... -‐electron density can be varied in various ways [interlayer doping (e.g. high-‐Tc cuprates), field-‐effect (oxide interfaces like LAO/STO, example of E-‐tuned SC),...];
• Compe,ng mechanisms in 2D are weaker (like AF-‐spin order and charge-‐ordering in cuprates) or stronger (like disorder in films and interfaces)
• Interes,ng objects for fundamental physics: -‐role of fluctua,ons in low D, physics of vor,ces, Berezinski-‐Kosterliz-‐Thouless
(BKT) transi,on,... -‐SC proper,es can be tuned (and e.g. Tc →0) to study the SC-‐Insulator or SC-‐Metal quantum phase transi,on -‐different ways disorder and interac,ons can affect SC ...
![Page 3: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/3.jpg)
Short premise: quantum phase transi,ons (quick survey) [Sondhi]
€
Z(β) = nn∑ e−βH n The density-‐matrix operator similar to ,me evolu,on
provided
€
e−βH
€
e−iHτ /
€
τ = −iβ
The Feynman path integral describing the (imaginary) ,me evolu,on of each degree of freedom becomes equivalent to the par,,on func,on in D+1 dimensions
![Page 4: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/4.jpg)
A simple example: the 1D Josephson-‐Junc,on array
Δ >>T : in each grain electrons are well paired, but L~ξ and the phase can fluctuate
ϑj ϑj+1
€
H =C2
V j2 − EJ
j∑ cos ϑ
∧
j −ϑ∧
j+1⎛ ⎝ ⎜
⎞ ⎠ ⎟
where is the voltage on the j-‐th junc,on naturally related to the number of Cooper pairs on grain j (remember that
€
V j ≡ −i2eC
∂∂ϑ j
€
n j = −i ∂∂ϑ j
EJ
![Page 5: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/5.jpg)
The sum of all the imaginary-‐,me evolu,ons to get Z(β) is equivalent to the par,,on func,on in 1+1 dimensions with
where with being the capaci,ve charging energy:
(small K: small Ec or large EJ) ⇒ charge can fluctuate a lot and phases can be more “rigid” (large K: largel Ec or small EJ) ⇒ charge cannot fluctuate a lot and phases are more loosely bound between the sites. Thus Small K: ordered phases, large K: disordered phase 1+1 dimension can be generalized to D+1 dimensions.
€
HXY =1K
cos ϑ i −ϑ j( )ij∑
€
K ≈ EC /EJ
€
EC =2e( )2
C
T Tc(K)
Kc
QUANTUM CRITICAL
QUANTUM DISORDERED
(a) (c) (b)
QCP
![Page 6: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/6.jpg)
K
T Tc(K)
Kc
QUANTUM CRITICAL
QUANTUM DISORDERED
(a) (c) (b)
QCP
At T=0 both ξ(K) and ξτ(K) diverge for δ≡(K-‐Kc)→0 as
with z dynamical cri,cal index
define the typical energy scale of flucts.
At finite T, in the QD region flucts. are mostly quantum
In the QC region flucts. are both thermal and quantum T is the most important parameter ruling the physics
€
ξ ~ δ −ν ,ξτ ~ ξz
€
ξ ~ δ −ν
€
ξ ~ T−1z
Some more “pills” of quantum cri,cality
€
ξτ−1 ~ ξ−z ~ 1
τ fl
~ ω fl
![Page 7: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/7.jpg)
A first “lab” of quasi 2D superconductors: high-‐Tc cuprates
1) Weakly coupled CuO2 planar structures Strong anisotropy quasi-2D systems 2) When one hole per CuO2 cell is present (half-filling) the cuprates are insulating (and AF)
e-e interactions are strong X doping
T
AF
metal
SC
Charge reservoir
CuO2
CuO2
J. Bednorz, K.A. Müller, 1986, ...+26 years of fran,c ac,vity.....
![Page 8: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/8.jpg)
What are the “hot” physical issues in cuprates?
TTc(X)
Xc
QUANTUM CRITICAL
QUANTUM DISORDERED
QCP X
Tc
T
|U| BCS regime
Δ∼T* Incoherent bound pairs
condensate of pairs
unbound electrons
X
Two dis,nct lines of thought: In the PG phase i) there are only incoherent bound pairs (anomalies due to strong e-‐e correla,ons and pairs ii) there is a “hidden” ordered phase and a QCP underneath the SC dome around op,mal doping Xc Low dimensionality weakens the “ordered phase” and leaves dynamical quantum fluctua,ons (source of
pairing?)
i)
ii)
Remember the nega,ve-‐U Hubbard model?
Tc
T
|U| BCS regime
Δ∼T* Incoherent bound pairs
condensate of pairs
unbound electrons
X
![Page 9: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/9.jpg)
Short list of hot SC-‐related issues in cuprates
• What is the “normal” state: FL of Landau QPs or anomalous Non-‐FL state?
• Superconduc,ng mechanism;
• Physics of SC transi,on: BCS or preformed pair condensa,on?
• role of phase fluctua,ons: any signature of BKT transi,on? Can vortex-‐an,vortex pairs affect the normal-‐state proper,es?
• Quantum cri,cality: -‐Is there any compe,ng phase (and related dynamical low-‐energy modes)? -‐SC-‐Insulator transi,on: role of disorder, polaronic effects, inhomogenei,es, compe,ng phases,....
Cuprates are a big “cocktail” of several “revolu,onary” issues : all the “dogmas” of the BCS paradigm are challenged
![Page 10: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/10.jpg)
All the very basic ingredients of BCS are ques,oned in cuprates....
High Tc cuprates Normal state: Strong interactions (both repulsive and attractive)+ low dimensionality+ small EF
What happens at low energy? FL or NFL? Just preformed pairs below T* or there is a competing phase (like, e.g., charge-ordered) with low-energy critical dynamical flucts.? In this case how they interact with the QPs?
Superconducting state Attraction and repulsion of the same order (non-phononic pairing?) Strong pairing small pair size
€
ξ0 ≈ 3 − 5aMean-‐field is ques,onable Large phase flucts. Likely occurrence of preformed incoherent pairs below T*
Low Tc (standard metals) Normal state: High energy e-e repulsion (~EF~10eV)+ fast screening processes Nearly free electron gas at low energy (Landau QP’s)
Superconducting state Attraction (pairing) at low energy from phonons (small expansion parameter,
Migdal )
Large pair size : weak phase flucts. and rigid w.f. €
ωD
EF
<<1
€
ξ0 >> a
![Page 11: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/11.jpg)
The doping is set near the SIT and then the hole density is changed by VG
€
x − xc ~ ξ−1ν ~ T
−1z
⎛
⎝ ⎜
⎞
⎠ ⎟
−1ν
Bollinger et al., Nature 2011. electric-‐field driven SC
SC-‐Insulator transi,on in strongly underdoped cuprates
![Page 12: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/12.jpg)
How many SC-‐I, SC-‐M-‐I, SC-‐M transi,ons?
T
ρ ρc
ρ
T T
ρ ρc ρc1
ρc2
The type of transi,on, the universality class, and so on depend on:
• The morphology of the system (homogenus, granular, honeycomb,....)
• The way the transi,on is induced (film width, amount of disorder, magne,c field, electron density by doping or ga,ng,....)
Universal value h/4e2?
![Page 13: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/13.jpg)
How many SC-‐I, SC-‐M, SC-‐M-‐I transi,ons?
1. Percola,on: a “geometrical” effect
2. Granular superconductors: R-‐shunted JJA: What ma|ers are phase fluctua,ons, and transport due to fermionic QPs
3. The fermionic scenario: SC dies because pairing is spoiled
4. the Bosonic scenarios:Fisher & Co., Feigel’man,Ioffe,Mezard,Kravtsov,...SC dies, but fermions remain paired.
![Page 14: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/14.jpg)
1. Percola,on Pieces of coherent SC add to a coherent cluster (e.g. by pouring more charge into the system). When the growing cluster crosses the whole sample, global SC sets in....
SC-‐I or SC-‐M depend on the matrix embedding the SC cluster
Biscaras et al., PRL 2012 Biscaras et al., Nat. Commun. 2010
LaTiO3/SrTiO3
![Page 15: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/15.jpg)
2. Granular SCs: the R-‐shunted JJA
SC arises inside each grain
Well locked phases
Badly locked phases
![Page 16: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/16.jpg)
QPs are freed and start conduc,ng jumping from one grain to the other....
J. Valles talk, Leiden
![Page 17: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/17.jpg)
ie ϕψ ψ=
)1log(g241
0 ετπλλ −=
Con,nuous decrease of Tc
Cooper pairing dies out at the SIT
M.P.A. Fisher et al., PRL 64, 587 (1990) M.P.A. Fisher, PRL 65, 923 (1990)
SIT due to long wave length phase fluctua,ons
Insulator made of localized Cooper pairs
3. vs 4. The Fermionic vs. Bosonic scenarios of SIT
![Page 18: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/18.jpg)
€
bk,bk '+[ ] = δk,k' 1− nk↑ + nk↓( )[ ]
bk,bk '[ ] = bk+,bk'
+[ ] = 0
Spin models for SC
have the same algebra of spin-‐1/2 operators P.W. Anderson, Phys. Rev. 110, 827 (1958); 112, 1900 (1958)
€
Skz =121− nk↑ + nk↓( )[ ]
Sk+ ≡ Sk
x + iSky = bk
+, Sk− ≡ Sk
x − iSky = bk
Remember also the XY model represen,ng the 1+1 Josephson-‐Junc,on array
EJ
![Page 19: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/19.jpg)
€
H = −4 ξiSiz
i∑ − g Si
+S j− + Si
−S j+( )
i, j∑
ξiare (random) variables: the energy of a spa,ally localized CP
Two different bosonic scenarios depending on rela,ve importance of disorder (W) and Coulomb rep. (remember in the XY model for JJA...)
ξ
P(ξ)
W -‐W
€
g ~ 1 K ~ EJ EC
Coulomb-‐driven SC-‐I trans (Fisher and Co.) Disorder not the main issue long range Coulomb (enhanced by disorder, see Altshuler Aronov) competes with pair hopping ⇒ charge rigidity and vortex of xy spin component destroy SC.
Pair hopping ~ EJ
SC means a finite magne,za,on on the XY plane
Disorder-‐driven SC-‐I transiAon (Feigel’man, Ioffe, Kravtsov, Mezard,...) When disorder is too strong (pairs very localized, W>>g) the local field orients the spins along ±z and destroys the “magne,za,on” (i.e. SC) on the xy plane
Seminal paper: Ma, Lee, PRB 1985
![Page 20: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/20.jpg)
Coulomb-‐driven bosonic SC-‐I transi,on
Disorder-‐driven bosonic SC-‐I transi,on
![Page 21: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/21.jpg)
An example from the fermionic class: a-‐Bi/Ge
![Page 22: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/22.jpg)
TiN, Sacepe et al. PRL
Map of the spectral gap in sample TiN1
B. Sacépé et al. PRL 101, 157006 (2008)
TiN films
![Page 23: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/23.jpg)
0exp TRT
⎛ ⎞∝ ⎜ ⎟
⎝ ⎠ T0
Tc
Δ = 1.76
Tc
D. Shahar and Z. Ovadyahu, Phys. Rev. B 46, 10917 (1992)
V. F. Gantmakher et al., JETP 82, 951 (1996)
InOx films
Sacepe et al, Nat. Phys. 2011
![Page 24: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/24.jpg)
Oxide interfaces
Caviglia et al.,Nature 2008
Biscaras et al., PRL 2012 Biscaras et al., Nat. Commun. 2010
LaAlO3/SrTiO3
LaTiO3/SrTiO3
![Page 25: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/25.jpg)
![Page 26: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/26.jpg)
What do we learn: SC can die in many different ways and all of them can be realized in
different systems (or maybe some also in the same system)....
A rich variety of physical systems gives rise to a huge variety of phenomena. Difficult classifica,on and organiza,on of various
effects in different materials.
It’s s,ll confusing but it’s also big fun
![Page 27: Superconductivity in low-dimensional systems1+1#dimension#can#be#generalized#to#D+1#dimensions.## € H XY = 1 K cos(ϑ i −ϑ j) ij ∑ € K≈ E C /E J € E C = (2e)2 C T Tc(K)](https://reader034.fdocuments.in/reader034/viewer/2022050502/5f94799c6d8add35f6489e39/html5/thumbnails/27.jpg)
Bibliography • G. Deutscher, Chance and Ma+er, Les Houches, Session XLVI, NATO ASI, 1986, North-‐Holland
• [Sondhi]S. L. Sondhi, M. Girvin, J. P. Carini, and D. Shahar, Rev. Mod. Phys. 69, 315 (1997)
• [Chaikin] Chaikin, P.M., and Lubensky T.C., Principles of Condensed Ma+er Physics (Cambridge University, New York), 1995.
• [Goldman] A. M. Goldman, Int. J. Mod. Phys. B 24, 4081 (2010) Review on SC-‐I transi,ons
• h|p://www.lorentzcenter.nl/lc/web/2011/449/presenta,ons/index.php3?wsid=449&type=presenta,ons