Fermion Quantum Monte Carlo based on the idea of sampling “graphs” Ali Alavi University of Cambridge Alex Thom James Spencer EPSRC Overview Introduction and motivation Paths integrals and the Fermion sign problem FSP as a problem in “path counting” A useful combinatorial formula From path-sums to graph-sums Applications to molecular systems Towards application to periodic systems Essence of idea Express the many-electron path integral in a finite Slater Determinant basis Resum the path integral over exponentially large numbers of paths to convert path-sums => graph-sums k i2 iP-1 i1 iP l i j The graphs are much more stable entities which can then be sampled. A graphical, or diagrammatic, expansion of the partition function Q w [G] (n) n Q G + + + + + …. G 2-vertex 3-vertex Each vertex is a Slater determinant Each graph represents the sum over all paths of length P which visit all verticies of the graph Non-pertubative expansion Path Integrals Consider the (thermal) density matrix: ˆ ˆ ( ) e H In terms of the eigenstates of the Hamiltonian: ˆ i e E i i i 0 0 for (i.e. zero temperatur e) The energy can be calculated from: Q Tr [e Hˆ ] Hˆ ln Q E Hˆ Tr [e ] Tr [ He ] The density matrix can be represented in real space For a single electron located at x: x x' x e Q dx x e e Hˆ Hˆ Hˆ x' x ( / P ) Hˆ ( / P ) Hˆ ( / P ) Hˆ e .e .... e P f actors: or " timeslices" x x' dx2 ...dxP x e( / P ) H x2 x2 e( / P ) H x3 ... xP e( / P ) H x' ˆ xe ( / P )(Tˆ Vˆ ) ˆ x' e ( mP / 2 )( x x ') 2 ˆ .e ( / 2 P )(V ( x ) V ( x ')) PE terms KE terms:harmonic spring In the limit P the path denoted by x3 x x2 ... xP x x2 tends to a continuous function x xP x( ), 0 , with bc : x( ) x(0) Q dx Dx( )e S [ x ( )] path intergral S [ x( )] 1 2 0 2 mx ( ) V [ x( )]d KE along path PE along the path One can simulate an electron as a ring-polymer, moving in the external field (which itself can be dynamic). Polarons [Parrinello Rahman] For N electrons X ( x1 , x2 ,..., x N ) X ( ) x1 x1 x2 x2 x2 x1 x1 x2 Describes closed paths which can exchange identical particle coordinates X ( ) Pˆ X (0) ( xi1 , xi2 ..., xiN ) 1 2 S[ X ( )] d mi xi V [ X ( )] U [ X ( )] i 2 Coulomb interaction 0 1 Pˆ S / Q dX DX ( )( 1) e N ! Pˆ Odd permutations subtract from the Partition function: Fermion sign problem As N or increases, there is an exponential cancellation of contributions arising from even and odd paths. Slater Determinant space Let Di be a Slater determinant composed out of N orthonormal spin-orbitals [e.g. Hartree-Fock orbitals, Kohn-Sham, etc] chosen out of a set of 2M: Di Dn1n2 .. nN {u1 , u2 ,..., u2 M } 1 det[un1 un2 ...unN ] N! 2M N det N e.g. M 100, N 10 N det 1017 The Di form a orthonormal set of antisymmetric functions. They are solutions to a non-interacting, or uncorrelated (mean-field) Hamiltonian H0: H 0 Di Ei Di 0 Full problem: H T U V , N N 1 2 1 [T i ,U ,V v(ri )] 2 i i j ri r j i H E , ca i Di i Exact solutions are linear superpositions of uncorrelated determinants Paths in Slater determinant space A closed path in S.D. space Di1 Di 2 ... Di P Di1 i2 iP-1 i1 iP w( P ) [ Di1 , Di 2 ,...., Di P , Di1 ] Di1 e H / P Di 2 Di 2 e H / P Di 3 ... Di P e H / P Di1 A path of P steps in SD space Q Tr[e H ] Di e H Di i ... w( P ) [i1 , i 2 ,..., i P , i1 ] i1 i2 iP A given S.D. can occur multiple times along a path ij Di e H / P Dj ij is a computable , diagonally - dominant and extremely sparse matrix : ij ij P H ij O[( / P ) 2 ] (Primitive approximat ion) and a better approximat ion is : ij e ( E i ( 0 ) E j( 0 ) ) / 2 P [ H H ( 0 ) H (1) ] ( ij P H (1) ij ) O[( / P ) 2 ] Hamiltonian matrix elements (Slater-Condon rules) Since H contains at most 2-body interactions: Di H Dj 0 if Di and Dj differ by more than 2 spin - orbitals Di Dj i j 1 a b 1 Di U Dj ij r12 ab ij r12 ba Hamiltonian connects only single and double excitations: Maximum connectivity N ( N 1)(2M N )( 2M N 1) / 4 N 2 M 2 Spin selection rule: Di H Dj 0 if S z [ Di ] S z [ Dj ] Cost of calculation Other symmetries may also exist Hubbard model: translational invariance; Molecules:point group symmetry Di H Di O( N 2 ) Di H Dj O( N ) if Di and Dj differ by 1 spin - orbitals Di H Dj O(1) if Di and Dj differ by 2 spin - orbitals Search for a power series in ii ij Di e H / P n2 Dj j Q wi Two-hop i i n1 P 2 P 2 n1 wi ii P j i n1 0 n2 0 P 3 P 3 n1 P 3 n1 n2 k j i n1 0 n2 0 n3 0 P-2-n1-n2 ii ij jj ji ii n1 n2 P 2 n1 n2 ii ij jj jk kk ki ii n1 n2 n3 j 3-hop k i P 2 n1 n2 n3 ... Rearranging Two - hop terms : P 2 P 2 n1 j i n1 0 n2 0 ii n ij jj n ji ii P 2 n n 1 2 1 2 Nested sum n2 P 2 P 2 n1 P2 jj ii ij ji n1 0 n2 0 ii j i transitionmatrix elements 3 - Hop terms : P 3 P 3 n1 P 3 n1 n2 k j i n1 0 n2 0 n3 0 ii n ij jj n jk kk n ii P 3 n n ii P 3 1 k j i 2 3 1 2 P 3 P 3 n1 P 3 n1 n2 ij jk ki n1 0 n2 0 n3 0 jj ii n2 kk ii n3 Define the nested sum: Zh ( P) P h P h n1 ( x1 , x2 ..., xh ) n1 0 n2 0 which appears in the h-hop term P h h1 ni i 1 ... nh 0 n1 n2 x1 x2 ..xh nh The “hop” series ij ji ( P ) jj wi ii 1 Z 2 (1, ) 2 ii j i ii P j k i k j i j i kj l k i ij jk ki Z3 ii ij jk kl li 3 ii 4 ( P) jj kk (1, , ) ii ii Z4 j k i l ( P) jj kk ll (1, , , ) .... ii ii ii Using induction, one can show: 1 z 1 1 ( x1 , x2 ..., xh ) 2i C z 1 ( z x ) P Zh ( P) x1 x2 x3 AJW Thom and A Alavi, J Chem Phys, 123, 204106, (2005) Residue Theorem gives For x1 x2 x3 .. Zh ( P) x 1 1 ( x1 , x2 ..., xh ) i xi 1 ( xi x j ) i h P j i e.g . x 1 ( x1 ) 1 x1 1 P ( P) 1 Z 1 x 1 1 x 1 ( P) 2 Z 2 ( x1 , x2 ) 1 x2 1 ( x2 x1 ) x1 1 ( x1 x2 ) P P Some useful properties of Z-sums Zh ( P) ( x1 , x2 ..., xh ) 0 if h P for all x1 , x2 ..., xh Replace upper limit of sums over h to 1 z 1 dz 0 2i z 1 P Z0 ( P) Symmetry: Zh ( P) ( x1 , x2 ..., xh ) Z h ( P) ( xi1 , xi2 ..., xih ) From “hop”-expansion to “vertex” expansion Consider the 4-hop terms: j k 4-vertex i l j=l j k 3-vertex i “chain” i=k j=l “Star” 2-vertex i=k l Analytic summation over alternating series j j k Z3 ( P) i k Z6 ... i n ij jk ki jj kk Z 3n (1, S , ) 3 ii ii ii n 0 ( P) “Cycle function” ij jk ki Define : Aijk ( z ) ( z ii ii )( z ii jj )( z ii kk ) 1 z P 1 1 z P 1 1 n S A ijk 2 i z 1 2 i z 1 1 Aijk ( z ) n 0 C C Solve : 1 Aijk ( z ) 0 feed solutions into the residue th. Eg. A 2-vertex graph j ii jj a ij b i b2 b2 A( z ) 2 ( za a)( za a) a ( z 1) 2 (For simplicity) Next compute S2: 1 z P 1 1 1 z P 1 S2 2i z 1 1 A( z ) 2i z 1 1 b2 1 2 a ( z 1) 2 1 z P 1 ( z 1) 2 1 ( z 1) P z 1 2i z 1 ( z 1) 2 (b / a ) 2 2i ( z 1) 2 (b / a ) 2 1 ( z 1) P z 1 2i ( z 1 b / a )( z 1 b / a ) (1 b / a) 1 (b / a ) (1 b / a ) P 1 (b / a ) 2(b / a ) 2(b / a ) P 1 ( a b) P ( a b) P 1 1 P P 2 a a Star graphs j j k 2 1 i l Z5 ( P) 3 1 k 2 1 i Z8 ( P) ... G1 l Gg For a star-graph with g-spokes, G1,G2,…Gg attached to i S star G2 G3 n1 n2 ... ng 1 z P 1 n1 n2 ng AG1 AG2 .... AGg n1 , n2 ,..., ng 2i C z 1 n1 , n2 .., n g 1 z P 1 1 2i C z 1 1 AG1 AG2 .... AGg Chains graphs 1 1 G3 G1 G2 Z7 ( P) 1 1 1 G3 G1 2 G2 Z10 ( P) ... n1 1 n2 n2 1 n3 .... AG1 n1 AG2 n2 .... AGg ng S chain n1 n2 n1 , n2 .., n g 1 z P 1 2i C z 1 1 AG1 1 AG 2 1 ... 1 1 AGg General 3-vertex graph Unfolded representation: Each spoke represents an Independent circuit on the graph k Folded representation j j k j k i k i k 1 z P 1 S3 (ijk ) 2i c z 1 j 1 1 2 Aijk 1 Ajk Aij 1 Ajk Aik 1 Akj 1 Ajk 1 z P 1 2i c z 1 1 2 Aijk Aij Aik Ajk Denominator is cubic polynomial in z i.e. there are 3 residues Unfolded 4-vertex graph Denominator is a quartic polynomial in z A graphical, or diagrammatic, expansion of the partition function Q w [G] (n) n Q G Dijab + Dijab + + abcd Dijkl Dijab Dija 'b ' + + …. Dia' j'b' ' G 2-vertex 3-vertex Each vertex is a Slater determinant Each graph represents the sum over all paths of length P which visit all verticies of the graph 2,3, and 4-vertex graphs + + + + Monte Carlo sampling of graphs The energy can be obtained from: E ln Q 1 1 w( n ) [G ] ( n ) ( n ) .w [G ] Q n G w [G ] If graphs can be sampled with an un-normalised probability given by w(n) [G], then the energy estimator is: ln w( n ) [G ] ~ (n) E [G ] i.e. ~ (n) E E [G ] ( n ) w [G ] ~ (n) sign ( w [G]) E [G] (n) E sign ( w( n ) [G]) | w| | w| For this to be useful, the denominator has to be well-behaved as i.e. the number of positive sampled graphs should exceed the number of negative sampled graphs in such a way that this difference is finite and does not vanish. Monitor fraction of sampled graphs which are trees, positive cyclic and negative cyclic graphs. f [T ] [G T ] |w| , f [C ] [G C ] sign ( w( n ) [G]) |w| |w| , f [C ] [G C ] f [T ] f [C ] f [C ] |w| For graphs that contain the HF determinant: wHF (n) w [G] n G HF ln wHF ~ EHF ~ lim EHF EHF [Hartree-Fock energy] 0 ~ lim EHF E0 [Ground state energy] Approximation:Truncate series at 2-vertex, 3-vertex or highervertex graphs. wHF (v) v (n) w [G] n G HF 2 vertex: Double-excitations 3 vertex: Quadruple excitations 4 vertex: Hexatuple excitations Number of graphs= [N2M2] [N4M4] [N6M6] N2 molecule ~ EHF N2 molecule in VDZ basis Types of sampled graphs (4-vertex level) sign ( w( n ) [G]) |w| N2 sampled energies (4-vertex level) N2 binding curve [sampling graphs which contain the HF determinant] Applications to periodic systems Taking a plane-wave PP code (CPMD) which can solve for (i) KS orbitals and potential (ii) KS virtuals -> Use these as the basis for the vertex series KS Hamiltonian becomes the reference (single-excitations now contribute) Need 2-index and 4-index integrals, which are computed on-the-fly using FFTs (time consuming part) Advantage: (i) Treatment of periodic systems (ii) No BSSE (iii) Can be used as a post-DFT method Graphite (4 atom) primitive cell. 16el, BHS PP (Ec=90Ry) 2-vertex Conclusions and outlook Development of QMC methods based on graphs gives a method to combat the Fermion sign problem Proof of concept for small molecular systems Major effort is now being expended on developing a periodic code….. …..perhaps to return to surface problems in due course! Advantage of graph-sampling algorithm O(N2) scaling! The observed stability at the 4-vertex level is extremely encouraging. Current work: (1) Extension to higher order graphs (2) Improved Monte Carlo sampling (3) Applications to large systems Graphs A graph a set of n distinct elements (in no particular order) with a given connectivity G Di , Dj , Dk ..... n distinct determinants Connectivity of graph is determined by ij k m k i j l i j Ga Gb Compactly expressed: Q w( n ) [G] n G Di , Dj , Dk ..... (n) w [G] G A set of n connected determinants Sum over all paths which visit all the determinants in G Each graph represents a sum over exponentially large numbers of paths its weight can be expected to be much better behaved than that of individual paths. A graph, G, is an object on which we can represent the paths which visit all the vertices in G The weight of a given graph is the sum over all paths of length P which visit all the vertices of the graph: w( n ) [G ] ... ' w( P ) [i1 , i 2 ,..., i P , i1 ] i1G i 2 G i P G The prime ‘ indicates that the summation indicies must be chosen in Such a way that each vertex in G is visited at least once. This condition ensures that the weights of two different graphs Ga and Gb (I.e. two graphs that differ in at least one vertex) do not double-count paths which visit only Ga Gb k m k i j l i j Ga w[Ga ] w[Gb ] Gb will not double-count paths which visit w[Ga Gb ] Quantum Chemical applications Dissociation of diatomic molecules: Multiple-bond dissociation, e.g. the N2 molecule, is a major challenge to any ab initio method. Use HF orbitals generated from MOLPRO Gaussian basis set [cc-pVDZ or VTZ] Two-electron primitive integrals read in from MOLPRO output and matrix constructed on the fly. Cost of the calculations 2-vertex 3-vertex 4-vertex <1 s 150 secs 1 week [over 109 4-vertex graphs to sum] On a pentium 4 processor [2003 vintage] How to make 4-vertex (and eventually higher vertex) calculations practical? So if on step t of an MC simulation consisting of K steps we are at graph Gt 1 E lim K K ~ (n) E [Gt ] t In order to perform a Metropolis MC simulation, one needs to ensure that microscopic reversibility is satisfied. In the present implementation, we generate fresh graphs at each step according to an algorithm to be shortly described. In addition one needs to compute the generation probability of a graph using this algorithm, in order to unbias the Metropolis MC acceptance ratio. w( n ) [G ' ] Pgen [G ] Pacc [G ' | G ] min( 1, ) (n) Pgen [G ' ] w [G ] Tree graphs are graphs that do not contain cycles j l k i The weight of trees is positive definite at all Exactly diagonalised by Krogh, Olsen CPL 344, 578, (2001), and by Chan, Kallay and Gauss, JCP, 121, 610 (2004) N2 VDZ 15820024220 determinants -108 0 1 2 3 4 5 -108.2 E/a.u. -108.4 FCI RHF -108.6 v=2 b=5 -108.8 v=3 b=5 all v=4 b=5 -109 -109.2 -109.4 r/a.u. To summarize: wi ii 1 S 2 (ij) S3 (ijk ) S 4 (ijkl ) ... k j i l k j i j i P Approximation:Truncate series at 2, 3 or higher vertex terms. 2 vertex: Double-excitations [N2M2] 3 vertex: Quadruple excitations [N4M4] 4 vertex: Hexatuple excitations [N6M6] By Comparison: CCSD: N2(M-N)4 Nit CCSDT: N3(M-N)5 Nit CCSDTQ: N4(M-N)6 Nit CCSDTQ56: N6(M-N)8 Nit An iIlustration of the Monte Carlo: 8x8 Hubbard lattice with 6 eMomentumspace basis 8 site system at or near half-filling is strongly open-shell (1,1) (0,0) (1,0) (2,0) +4 12 495 4 -4 (k x , k y ) 2t cos[( k x k y ) / 2] cos[( k x k y ) / 2] N det 16 8008 6 Dijab DHF + Dijab + + Dia' j'b' ' abcd Dijkl Dijab Dija 'b ' + + …. Finite T We wish to compute the energy at a finite -1=kT as Tr [ He H ] E H Tr [e ] D He D e H i i H i Di Di i Where the trace is taken over all Ndet determinants. Problem is that these sums are not “Monte Carlo-able”. Sampling Slater determinants Letting wi Di e H Di E H w D He Di / wi i i i w i i Since e H is a positive definite operator, its diagonal matrix elements are positive wi pi 0 is a probability in the usual sense of the word wi i (i.e. non - negative and normalised ) Noting : Di He H Di wi ln Di e H Di ln wi one can write the energy in a form suitable for a Monte - Carlo experiment : ln wi ~ writing : Ei ~ wi Ei ~ ~ i E p i Ei Ei wi i wi i Where the expectation value is taken over an ensemble of determinants sampled with probability pi. Perform Metropolis sampling of Di chosen according to wi 1 ~ E t it K K where i t is the determinan t on step t of the MC E lim simulation Di Dj , Pacc min(1, wj /wi ) Problem: the weight itself is a path-integral! Define: ij Di e H / P Dj [High-temperature DM] / P 10 3 10 4 wi Di e H Di ( ) ii P P 10 4 105 k wi ij jk ... li j,k ..., l P f actors Discrete path integral: wildly oscillatory integrand. Can’t use Monte Carlo! Hopeless to calculate by brute-force! j i l Generation of graphs with a computable generation probability We adopt a Markov chain algorithm is which successive determinants are added to a list until the desired size of graph is reached. However, since the connectivity of each determinant is not uniform, such an algorithm can produce a non-uniform generation probability. k j3 i j j2 Start at i, and selected a connected determinant, j, with probability pij. This results in a 2-vertex graph, G={i,j}. Next, select k, connected to j, with probability pjk. If k is distinct, then add k to the list: G={i,j,k}. Otherwise, select a new determinant from the current Position (i.e. the last visited determinant). Continue this process until n distinct verticies have been visited. The generation probability can be calculated by examining all possible ways of generating G according to this algorithm. For example, for a 3-vertex graph, G={i,j,k}: Pgen [G ] ( pij p ji ) n pij p jk ( pij p ji ) n pik n 0 n 1 ( pik pki ) n pik pkj ( pik pki ) n pij n 0 n 1 pij ( p jk p ji pik ) 1 pij p ji pik ( pkj pki pij ) 1 pik pki This procedure can be generalised for a n-vertex graph (n>3). The general case is most compactly expressed in matrix notation. Let us call our n verticies G={i1,i2,…,in}, all distinct, with i1=i. Consider the generation probability of G in the given order (i1,i2,…,in). According to this algorithm, we visit i2 for the first time from i1, i3 for the first time from either i1 or i2, etc. In general we visit ik for the first time from any of the previous visited k-1 verticies. The algorithm terminates when we first visit the n-th vertex. This is a first-passage problem in Markov chain theory. 0 p P ( k ) [i1 , i2 ,..., ik ] i2i1 ... 0 pi1i2 0 ... pi2i3 0 0 pi1ik pi2ik 1 We will construct a series of transition probability matrices in which vertex ik is an absorbing state: 0 pi2i1 (k ) P [i1 , i2 ,..., ik ] ... 0 Note that: pi1i2 0 ... pi2i3 0 0 pi1ik pi2ik 1 [ P ( k ) ]n ik 1 ,ik represents the probability of arriving at ik in exactly n steps given we Started from ik-1, passing through some or all of (i1,i2,…., ik-1). So therefore the total probability of arriving at ik, is simply the geometric series: n 1 [ P ( k ) ]n ik 1 ,ik (k ) I P k 1,k Therefore the probability to generate the graph G in the sequence: i1 i2 ... ik 1 1 1 Pgen [i1 , i2 ,..., ik ] ... ( 2) ( 3) (k ) I P 1, 2 I P 2,3 I P k 1,k The probability to generate G in any order is given by the sum over all n! permutations: Pgen [G ] Pgen [ Pˆ (i1 , i2 ,..., ik )] Pˆ In current implementation we choose pij 1 Ni j3 i j1 j2 Where Ni is the number of determinants connected to i. In other words we do not introduce an energetic bias in the selection of determinants. Conclusions A new approach for Fermion Monte Carlo is being developed, based on sampling Slater determinant space with weights computed according to a novel path counting scheme. The mathematics of path-counting needs further investigation. The scheme has been applied to the Hubbard model and the N2 problem with encouraging results. Topology of graphs j j j k i Cyclic i,j, and k all must be single or doubleexcitations of each other. k k i Star (tree) j and k all must be s- or d-excitations of i, but not necessarily of each other i Chain (tree) j must be a s- or dexcitation of i, and k is a triple or quadruple of i. Future work Technical Sampling graphs to counter the scaling problem Calculation of electron density Parallelisation of code Systems: Hubbard models [e.g. stability of striped phases] Dispersion interactions (e.g. graphitic systems) Contribution to the weights wi ii 1 S 2 (ij) S3 (ijk ) S 4 (ijkl ) ... k j i l k j i j i P 10-site, N=10, U=4 weights Contribution to the Weight 9 8 7 momentum b=1 6 UHF b=1 5 UHF b=2 4 UHF b=5 3 UHF b=10 2 1 0 0 1 2 3 Vertex 4 5 6 Tentative conclusion is: for the Hubbard model with U=4, the 3-vertex approximation is not perfect, although it is nevertheless an improvement over UHF: Captures about 20-50% of the correlation energy. Can we estimate the contribution of the higher order graphs through a MC sampling? => work in this direction is in progress Distribution of terms among the 2,3 and 4 vertex graphs [UHF basis] Convergence of Êi with the vertex approximation N=10, U=4, beta=1 -5 -5.5 0 1 2 3 4 5 6 -6 Energy vertex [mom] -6.5 Exact GS UHF -7 vertex UHF b=1 -7.5 RHF -8 -8.5 -9 Vertex 1010 Hubbard Model 184756 determinants at N=10. Exactly diagonalisable with effort on P4 Half-filled system is closed-shell +4 +1 -1 -4 Two important questions How good is the 3-vertex approximation? What is the best one-particle basis to use? j A 3-determinant star a b b b a 0 (a 2b), 3 a b 0 a 1 / 2 1/ 2 0 c 1 / 2 , c 1 / 2 , c 3 1 / 2 1/ 2 1 / 2 1 / 2 (a 2b) P (a 2b) P w 2 2 Contour integral solution: b2 Aij ( z ) Aik ( z ) 2 a ( z 1) 2 i k 1 z P 1 1 1 z P 1 S3 2i z 1 1 Aij ( z ) Aik ( z ) 2i z 1 1 2b 2 1 2 a ( z 1) 2 1 z P 1 ( z 1) 2 1 ( z 1) P z 1 2i z 1 ( z 1) 2 2(b / a ) 2 2i ( z 1) 2 2(b / a ) 2 1 ( z 1) P z 1 2i ( z 1 2b / a )( z 1 2b / a ) (1 2b / a ) P 1 ( 2b / a ) (1 2b / a ) P 1 ( 2b / a ) 2 2 (b / a ) 2 2 (b / a ) 1 (a 2b) P (a 2b) P 1 1 P P 2 a a Therefore w a P (1 S3 ) 1 (a 2b) P (a 2b) P 2 Again in exact agreement with the diagonalisation result j Fully connected 3-vertex graph k i Via diagonalisation: a b b b a b 1 (a 2b), 2,3 (a b) b b a 1 3 2/ 6 0 1 1 3 , 2 1 2 , 3 1 6 1 2 1 3 1 6 (a 2b) P 2 wi ( a b) P 3 3 Via the contour integral: 1 Ajk 1 z P 1 S3 2i z 1 1 2 Aijk Aij Aik Ajk b2 1 2 1 z P 1 a ( z 1) 2 2b 3 3b 2 2i z 1 1 3 2 3 a ( z 1) a ( z 1) 2 [multiply top and bottom by (z-1)3] 1 ( z 1) 2 (b / a ) 2 P z 1 2i ( z 1) 3 2(b / a )3 3(b / a ) 2 ( z 1) 1 ( z 1 b / a )( z 1 b / a ) P z 1 2i ( z 1 b / a ) 2 ( z 1 2b / a ) 1 ( z 1 b / a) P z 1 2i ( z 1 b / a )( z 1 2b / a ) (1 b / a) 1 (2b / a ) (1 2b / a ) P 1 (b / a ) (3b / a ) 3(b / a ) P 2 (a b) P 2 1 (a 2b) P 1 P P 3 a 3 3 a 3 [Factorise] [Cancel factors] [Evaluate two residues at z=1-b/a, and z=1+2b/a] Therefore 1 2 w a P (1 S3 ) (a b) P (a 2b) P 3 3 N2 VDZ -107.6 0 2 4 6 8 10 -107.8 FCI RHF CCSD CCSD(T) CCSDT v=2 b=5 v=3 b=5 all v=3 b=5 spec v=3 b=5 spec3 v=4 b=5 fullsum spec3 -108 E/a.u. -108.2 -108.4 -108.6 -108.8 -109 -109.2 -109.4 r/a.u. 8-site Hubbard model with N=6 electrons 3-vertex weights against exact weights j Some simple examples. ii jj a (a) A two-determinant system Exact solution via diagonalisation: ij b i 1 / 2 1/ 2 a b , (a b), b a 1 / 2 1/ 2 2 P ( a b) P ( a b) P wi k Di k 2 2 k Solution via Contour integral formula: First define A(z): b2 b2 A( z ) 2 ( za a)( za a) a ( z 1) 2 Next compute S2: 1 z P 1 1 1 z P 1 S2 2i z 1 1 A( z ) 2i z 1 1 b2 1 2 a ( z 1) 2 1 z P 1 ( z 1) 2 1 ( z 1) P z 1 2i z 1 ( z 1) 2 (b / a) 2 2i ( z 1) 2 (b / a) 2 1 ( z 1) P z 1 2i ( z 1 b / a)( z 1 b / a ) (1 b / a) 1 (b / a) (1 b / a) P 1 (b / a) 2(b / a) 2(b / a) P 1 ( a b) P ( a b) P 1 1 P P 2 a a Therefore 1 w a P (1 S 2 ) (a b) P (a b) P 2 In exact agreement with diagonalisation result Residue theorem 1 f ( z )dz (sum of enclosed residues ) 2i c Residue at pole of order m at z0 1 d ( m 1) m a1 ( z z ) f ( z) 0 ( m 1) (m 1)! dz Exact ground-state energy, UHF and lowest Êi vs particle number U=4 10-site, 10-site, U=4 -6 -6 -6 -6 000 -6.5 -6.5 -6.5 0 -6.5 222 2 44 4 4 66 6 6 88 8 8 10 10 10 10 12 12 12 12 -7 -7 -7 -7 Energy/t Energy/t Energy/t -7.5 -7.5 -7.5 -7.5 -8 -8 -8 GroundEE Ground Ground E EUHF Ground E EUHF EUHF v=3 [momentum] v=3 [momentum] v=3 [momentum] -8.5 -8.5 -8.5 -9 -9 -9 v=3 UHF -9.5 -9.5 -9.5 -9.5 -10 -10 -10 -10 -10.5 -10.5 -10.5 -10.5 -11 -11 -11 -11 Nel Nel Nel Nel The electron correlation problem How to account for the fact that electrons move around in a correlated fashion? Quantum chemistry approach is: Start from Hartree-Fock and try to improve systematically Hartree-Fock [mean-field theory, N3 ~ N4] HF=D0 Coupled Cluster [CCSD(T), N7] + perturbation theory Full-configuration interaction [eN] =eT HF = HF+ j cjDj Expansions in Antisymmetric functions Essential feature of HF theory: maintains an orbital (one-particle) picture of electronic structure HF det[u1 (x1 )u2 (x 2 )...u N (x N )] Quantum Monte Carlo QMC refers to stochastic methods to solve the Schrödinger Equation (or sample path integrals) based on interpreting the S.E. as a “diffusion equation in imaginary time”: 1 2 H V 2 is interpreted as a probability distribution. Long-time stochastic propagation [diffusion+life/death processes] leads to sampling the nodeless eigenstate of H. Application to Fermion systems is severely hampered by “sign problems”. Unconstrained sampling of the configuration space of Fermions leads to Boson Catastrophe. QMC can be stabilised by the introduction of constraints: -Fixed node approximation in diffusion MC [J Anderson] -Restricted path integral MC (fixing nodes of density matrix) [Ceperley] -Positive projection and constrained path MC for auxillary field QMC [Fahy and Hamman, Zhang] Why not use antisymmetric spaces? We would like to explore the possibility of using an antisymmetrized space as the basis for quantum monte carlo. Can we sample a set of Slater determinants in such way that we can extract meaningful physical quantities (eg energy) at the end of the simulation? This strategy avoids the Boson catastrophe from the outset without imposition of fixed-node type approximations. We will show that (1) Such a method is indeed numerically stable (2) The MC weights are obtained by summing over many paths of fluctuating sign. (3) Method depends on combinatorial ideas for path counting -> So far it is not exact (4) Applications to (a) Hubbard model and (b) Dissociating molecules A major conceptual advantage is that it allows to build directly on the one-particle picture of mean-field theory. What’s the problem? 2M N det N e.g. M 100, N 10 N det 1017 N2 VDZ 15820024220 determinants (exactly diagonalised by Krogh, Olsen CPL 344, 578, (2001), and by Chan, Kallay and Gauss, JCP, 121, 610 (2004) ) -107.6 0 2 4 6 8 10 12 -107.8 -108 FCI Energy/Hartrees -108.2 RHF v=2 b=5 -108.4 HF v=3 -108.6 SPEC2 SPEC3 -108.8 v=4 b=5 1 HF -109 g SPEC 2 -109.2 SPEC3 -109.4 r/a.u. 5 7 g u 1 N2 ( 1 g ) N ( 4S ) N ( 4S ) 5 g 7 g u 4su(2pz) 2g 1u su(2pz) g u su(2pz) g u 3sg (2pz) 2su 1sg sg (2pz) su sg sg (2pz) su sg Formally speaking, in the eigenvalue basis of H: wi e Ea Di a 2 a ~ Ei E e Ea a 2 Di a a e Ea Di a 2 a ~ lim Ei E0 when Di 0 0 Hubbard Model H t s [cs,i cs , j h.c.] U n ,i n ,i , i, j i Model of itinerant magnetism for narrow-band systems. Intensively studied since the mid-80’s in the context of High Tc. U Partition function Q Tr[e H ] Di e H Di i ... w( P ) [i1 , i 2 ,..., i P , i1 ] i1 i2 iP Note that the sign of w(P) is a very poorly behaved quantity: Depends on the product of P matrix elements. Therefore small variations in the path can lead to wild fluctuations in the sign of the path. Exact Diagonalisation H T U V N N 1 2 1 T i ,U , V v(ri ) 2 i i j ri r j i H E , ca i Di i Di H Dj ca j E ca i : a linear eigenvalue problem j Exact solutions are expressed as linear superpositions of uncorrelated Determinants. Conjecture: the n-vertex graph gives rise to a polynomial of degree n in the denominator of the contour integral Contour integrals reduce to a sum over n residues Q dx1dx2 ...dxP x1 e ( / P ) Hˆ x2 x2 e ( / P ) Hˆ x3 ... xP e ( / P ) Hˆ x1 x3 x2 x1 xP One can simulate an electron as a ring-polymer, moving in the external field (which itself can be dynamic). Polarons [Parrinello Rahman] Harmonic springs hold together a “ring polymer” E 1 ( / P ) Hˆ ( / P ) Hˆ dx dx ... dx E [ x , x ,.. x ] x e x x e x3 ... 1 2 P 1 2 P 1 2 2 Q ˆ xP e ( / P ) H x1 Motivation The development of stable fermion QMC algorithms which do not require fixed-node approximations, but which maintain a ~ N2 or N3 scaling. Does working in antisymmetric spaces (eg Slater determinants) help? Intuitively, Slater determinant spaces are the “right” spaces to be dealing with fermions: i.e. one should build in the fermion antisymmetry in the outset in any N-particle representation. Computational cost of electronic structure methods Hartree Fock MP2 - MP4 Coupled Cluster CCSD-(T) FCI DFT QMC N3 N4-N7 N6-N7 eN N3 N2-N3 Electron correlation is ubiquitous in chemistry e.g. in molecular dissociation su sB sA sg D0 ( x1 , x2 ) s g (1)s g (2) 1 ( s A sB ) 2 1 (1 2 1 2 ) 2 1 s g (1) s g (1) 2 s g (2) s g (2) D0 ~ 1 ( s A sB ) 2 Slater determinant: an uncorrelated wavefunction 1 ( s A (1) sB (1))( s A (2) sB (2)) 2 1 ( s A (1) s A (2) s A (1) s B (2) s A (2) sB (1) sB (1) sB (2)) 2 H H H . H . H . H . H H Incorrect dissociation Configuration Interaction Consider the doubly excited determinant: D2 corr ~ 1 s u (1) s u (1) 2 s u (2) s u (2) sB 1 ( D0 D2 ) 2 1 ( s A (1) s B (2) s A (2) s B (1)) 2 H . H . H . H . sA Correlated wavefunction Correct dissociation What is the problem with configuration interaction? -Slowly convergent with respect to short and intermediate range correlation. Must include many determinants: the problem grows exponentially with number of electrons and the number of virtuals -(linear) Truncated CI lacks size consistency: Coupled cluster methods are nowadays preferred.