Configuration-space Faddeev Formalism

• Configuration-space Faddeev formalism is an ab initio approach that decomposes the three-body wave function into components associated with pair interactions using Jacobi coordinates.
• The method formulates coupled differential or integral equations with specific asymptotic conditions to handle bound states, elastic scattering, and breakup processes.
• Extensions of the formalism include Coulomb-modified variants and Yakubovsky decompositions for four- and five-body systems, enhancing accuracy in scattering and structure calculations.

Configuration-space Faddeev formalism is an ab initio coordinate-space formulation of the nonrelativistic few-body Schrödinger problem in which the total wave function of three interacting particles is decomposed into components associated with the pair interactions. After separation of the center-of-mass motion and transformation to Jacobi coordinates, the method produces coupled differential or integral equations whose asymptotic conditions can be tailored to bound states, elastic and rearrangement scattering, and breakup or recombination. In the few-body literature, the term covers the three-body Faddeev equations proper, Coulomb-modified variants such as the Faddeev–Merkuriev and Faddeev–Noyes–Noble–Merkuriev equations, and the Yakubovsky extension to four- and five-body systems (Guérout, 14 Apr 2026).

1. Hamiltonian structure and coordinate representation

For three particles with masses $m_i$ and laboratory-frame position vectors $r_i$, the intrinsic Hamiltonian is written as

$H=T+V,$

with

$$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$

After separating the center-of-mass motion, one introduces three equivalent sets of mass-scaled Jacobi coordinates $(x_i,y_i)$, where $(ijk)$ are cyclic permutations of $(123)$. In one common convention,

$$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$

and

$$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$

with

$\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$

In these variables the kinetic energy becomes

$r_i$0

Equivalent formulations also use an arbitrary reference mass $r_i$1, leading to the same intrinsic free Hamiltonian in Jacobi coordinates (Guérout, 14 Apr 2026).

For many analytical and numerical purposes, hyperspherical variables are introduced in the $r_i$2 plane,

$r_i$3

together with the angular directions $r_i$4 on the six-dimensional configuration-space sphere. This parametrization is central in treatments of breakup and in numerical schemes that impose asymptotic conditions directly at large hyperradius (Lazauskas et al., 2019).

2. Faddeev decomposition and coupled equations

The defining step is the decomposition of the total wave function into three components,

$r_i$5

each component being driven by one pair potential. In differential form, the coupled equations can be written as

$r_i$6

or, equivalently under the sign convention used in other presentations,

$r_i$7

The total wave function is recovered by summation of the three components. This decomposition avoids double counting of pair correlations by isolating each pair interaction in its own component (Lazauskas et al., 2019).

The operator form employs the free Green’s operator

$r_i$8

and the two-body $r_i$9-operator in channel $H=T+V,$0,

$H=T+V,$1

The three-body Faddeev equations then become

$H=T+V,$2

This integral form is particularly useful when the kernel is expressed after partial-wave projection or when the equations are rewritten as driven scattering equations (Guérout, 14 Apr 2026).

After angular decomposition, each Faddeev component is expanded in bipolar harmonics or equivalent spin-angular basis states. One representative form is

$H=T+V,$3

where $H=T+V,$4 is labeled by angular momenta such as $H=T+V,$5. More general channel labels may include pair spin, total pair angular momentum, spectator orbital momentum, total three-body spin, and isospin. The resulting unknowns are two-dimensional radial amplitudes in the pair and spectator coordinates (Braun et al., 2021).

3. Asymptotic channels, breakup, and the global scattering matrix

The asymptotic content of the formalism depends on the total energy. When only a $H=T+V,$6 channel is open, the component associated with the bound pair $H=T+V,$7 carries an outgoing bound-plane-wave asymptote,

$H=T+V,$8

where $H=T+V,$9 is a bound-state wave function of the pair, $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$0 its energy, and $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$1 satisfies

$$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$2

The coefficients $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$3 are the $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$4 scattering amplitudes. In simpler elastic-channel formulations this asymptote is written as a bound-state factor multiplied by regular and outgoing spectator waves (Guérout, 14 Apr 2026).

When $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$5, breakup is open and all three particles can become free. In each Jacobi set the asymptotic region is $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$6 and $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$7. In hyperspherical variables the free asymptotic equation reads

$$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$8

Separation of variables introduces Delves angular functions $$T=-\sum_{i=1}^3 \frac{1}{2m_i}\nabla_{r_i}^2, \qquad V=V_1(r_2-r_3)+V_2(r_3-r_1)+V_3(r_1-r_2).$$9 with eigenvalue $(x_i,y_i)$0, and the radial dependence is given by Hankel functions $(x_i,y_i)$1, $(x_i,y_i)$2. By standard Bessel asymptotics,

$(x_i,y_i)$3

with breakup amplitude $(x_i,y_i)$4 defined on $(x_i,y_i)$5. The essential boundary condition is therefore the outgoing cylindrical-wave condition in hyperradius, imposed numerically at $(x_i,y_i)$6 by matching both the value and radial derivative to $(x_i,y_i)$7 (Guérout, 14 Apr 2026).

Above breakup, the open channels comprise $(x_i,y_i)$8 discrete $(x_i,y_i)$9 channels and $(ijk)$0 double-continuum channels, so that $(ijk)$1. One constructs $(ijk)$2 linearly independent solutions and collects their asymptotics into the matrix relation

$(ijk)$3

The scattering matrix thus contains elastic, rearrangement, breakup, and recombination processes in a single object.

Block	Process represented
$(ijk)$4	elastic $(ijk)$5
$(ijk)$6	$(ijk)$7 breakup
$(ijk)$8	recombination breakup $(ijk)$9
$(123)$0	elastic $(123)$1

This organization is one of the distinctive features of the double-continuum formulation: all scattering processes, starting from and ending in both single and double continua, are collected in a unique matrix (Guérout, 14 Apr 2026).

4. Coulomb interactions, three-body forces, and related variants

For charged systems, the direct Faddeev decomposition must be modified because the Coulomb interaction is long-ranged. Merkuriev’s construction splits each pair Coulomb term into short- and long-range parts by means of a cutoff function,

$(123)$2

and the Faddeev–Merkuriev equations become

$(123)$3

This rearrangement restores decoupling of the asymptotic channels while retaining the full Coulomb interaction in coordinate space (Deltuva et al., 2012).

A closely related implementation for nucleon–deuteron elastic scattering is the Faddeev–Noyes–Noble–Merkuriev formalism. In that setting one writes the equations in Jacobi and hyperspherical variables, uses matrix elements $(123)$4 for the long-range Coulomb contribution, and imposes Coulomb-distorted asymptotic solutions through the analytic functions $(123)$5, $(123)$6, and phases $(123)$7. The resulting block tri-diagonal algebraic system is then matched at $(123)$8 to determine the scattering amplitudes and recover the $(123)$9-matrix (Suslov et al., 2010).

Genuine three-body forces can be incorporated by adding $$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$0 to the Hamiltonian,

$$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$1

which leads to Faddeev equations of the form

$$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$2

For three-nucleon scattering, the Tucson–Melbourne two-pion-exchange force is represented in coordinate space through radial profile functions

$$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$3

and the operator structures $$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$4, $$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$5, and $$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$6. These terms enter the same configuration-space kernel as the pairwise interactions, together with the spin-isospin recoupling generated by permutations (Braun et al., 2021).

5. Partial-wave reduction and numerical realization

After partial-wave expansion, the coordinate-space Faddeev equations become coupled two-dimensional integro-differential equations for regularized radial amplitudes. One representative form is

$$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$7

with

$$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$8

and $$x_i=\tau_{x_i}(r_j-r_k),\qquad \tau_{x_i}=\sqrt{2\mu_{jk}},\qquad \mu_{jk}=\frac{m_jm_k}{m_j+m_k},$$9 a Jacobi-transform kernel given by a four-angle integral. Equivalent formulations express the right-hand side as recoupling kernels $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$0 or $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$1 after projection onto channel bases (Guérout, 14 Apr 2026).

Several discretization strategies are used in practice. One approach expands the amplitudes on tensor-product cubic Hermite splines $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$2 and collocates at Gauss–Lobatto points, leading to a linear system

$$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$3

where $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$4 and $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$5 are sparse kinetic and potential matrices and $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$6 is the dense Jacobi-kernel coupling between Jacobi sets. In polar coordinates $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$7, the cylindrical-wave breakup boundary condition becomes uniform in $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$8; a typical polar grid may have $$y_i=\tau_{y_i}\Bigl(r_i-\frac{m_jr_j+m_kr_k}{m_j+m_k}\Bigr),\qquad \tau_{y_i}=\sqrt{2\mu_{i,jk}},$$9 with $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$0 points and $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$1 points clustered near $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$2. Large systems are solved with GMRES preconditioned by $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$3, and convergence is reported in a few iterations even for $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$4 unknowns (Guérout, 14 Apr 2026).

An alternative nucleon–deuteron implementation discretizes the hyperangle by quintic Hermite splines and the hyperradius by a generalized Numerov scheme. After collocation at three Gauss points in each angular subinterval, one obtains a three-point hyperradial relation with block matrices $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$5, $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$6, and $\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$7, and the final amplitudes are determined by asymptotic matching and a least-squares fit at the outer boundary (Suslov et al., 2010).

For Coulombic three-cluster reactions, complex scaling rotates the coordinates according to

$\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$8

so that outgoing waves become exponentially decaying and the scattering problem acquires bound-state-like boundary conditions on a finite box. In that representation one again arrives at a large sparse linear system that can be handled with GMRES or similar solvers (Deltuva et al., 2012).

Bound-state calculations also admit a hyperspherical-harmonics realization. In the implementation summarized for three identical spin-$\mu_{i,jk}=\frac{m_i(m_j+m_k)}{m_1+m_2+m_3}.$9 particles, one expands in hyperspherical harmonics $r_i$00, truncates at finite $r_i$01, and solves coupled one-dimensional equations in $r_i$02 or, equivalently, integral equations in hypermomentum. In that work, retaining $r_i$03 captures $r_i$04 of the norm, while the $r_i$05 contribution is $r_i$06 (Kovalchuk, 2016).

6. Extensions, applications, and terminological boundaries

The Faddeev construction extends hierarchically to four and five particles through the Yakubovsky decomposition. For four particles, the six pair Faddeev components are split into $r_i$07 independent Yakubovsky components of $r_i$08-type and $r_i$09-type; for five particles, one obtains $r_i$10 independent FY amplitudes before imposing exchange symmetries. For identical particles, these counts reduce dramatically, for example $r_i$11, $r_i$12, and $r_i$13. After partial-wave projection, the corresponding equations become three- and four-dimensional integro-differential systems in the radial Jacobi variables (Lazauskas et al., 2019).

The formalism is used for both benchmark scattering and structure calculations. The 2026 double-continuum formulation was applied to neutron–deuteron scattering and organized the full set of elastic, rearrangement, breakup, and recombination processes in a unique matrix (Guérout, 14 Apr 2026). In proton– and neutron–deuteron elastic scattering at incident nucleon energy $r_i$14, the Faddeev–Noyes–Noble–Merkuriev approach was implemented with the charge-independent AV14 nucleon–nucleon potential and Coulomb interaction for proton–deuteron scattering (Suslov et al., 2010). In the $r_i$15 model at $r_i$16 deuteron laboratory energy, configuration-space Faddeev–Merkuriev plus complex scaling produced deuteron–$r_i$17 elastic differential cross sections agreeing to better than $r_i$18 with AGS over a wide angular range, and transfer cross sections coinciding within a few percent (Deltuva et al., 2012). For the $r_i$19 cluster model of $r_i$20, the calculated $r_i$21 excitation energy is $r_i$22, while the hyperon binding energy of the bound $r_i$23 state is less than the experimental value, which was noted as possible evidence for violation of the exact three-body cluster structure (Filikhin et al., 2014).

A common source of ambiguity is terminological rather than mathematical. In few-body quantum mechanics, “configuration-space Faddeev formalism” refers to the coordinate-space decomposition of the Schrödinger or Faddeev–Yakubovsky equations. It is distinct from the Faddeev formulation of gravity, where the metric is treated as a composite field built from ten vector fields and one studies continuum or piecewise-flat actions (Khatsymovsky, 2014), and it is also distinct from the Faddeev–Jackiw symplectic treatment of singular Lagrangians in classical mechanics (Fuente et al., 2023). In the few-body context, the defining objects are Jacobi coordinates, pair-resolved wave-function components, and multichannel asymptotic conditions.