Two-Loop Sunrise Integral

• The two-loop sunrise integral is a fundamental multiloop Feynman integral characterized by a double momentum-space integral with three propagators and an elliptic (genus one) structure.
• It is analyzed using Feynman parameterization and Picard–Fuchs differential equations, which reveal its complex analytic and monodromy properties.
• Its study bridges quantum field theory with algebraic geometry and elliptic polylogarithms, providing a framework for multi-loop computations and higher-order corrections.

The two-loop sunrise integral is a foundational multiloop Feynman integral appearing across high-precision quantum field theory computations, particularly as the first non-trivial topology displaying elliptic (genus one) geometry and therefore serving as the canonical example of Feynman integrals beyond the class of multiple polylogarithms. Its analysis connects quantum field theory, algebraic geometry, mixed Hodge theory, and the theory of elliptic polylogarithms, making it a central object for the study of elliptic iterated integrals in perturbative amplitudes.

1. Definition and Feynman Parameter Structure

The general two-loop sunrise (or "sunset") integral is defined in $D$-dimensional Minkowski or Euclidean space as a double momentum-space integral with three propagators: $$S(D; p^2; m_1^2, m_2^2, m_3^2) = \int \frac{d^D k_1}{i \pi^{D/2}} \frac{d^D k_2}{i \pi^{D/2}} \frac{1}{(k_1^2 - m_1^2) (k_2^2 - m_2^2) ((p-k_1-k_2)^2 - m_3^2)}$$ where $p$ is the external momentum and $m_1, m_2, m_3$ are the internal masses (Müller-Stach et al., 2011, Adams et al., 2013, Adams et al., 2015).

Feynman-parameterization expresses the integral as: $$S(D, t) = \Gamma(3-D) \int_{x_i \geq 0, \; \sum x_i = 1} \frac{\mathcal{U}^{3-\frac{3}{2}D}}{\mathcal{F}^{3-D}} \omega$$ with the standard graph polynomials: $$\mathcal{U} = x_1x_2 + x_2x_3 + x_3x_1, \quad \mathcal{F} = -x_1x_2x_3 t + (x_1 m_1^2 + x_2 m_2^2 + x_3 m_3^2) \mathcal{U}$$ and the 2-form $$\omega = x_1 dx_2 \wedge dx_3 + x_2 dx_3 \wedge dx_1 + x_3 dx_1 \wedge dx_2$$ (Adams et al., 2015, Adams et al., 2015). For $D=2$, all divergences are absent, and the focus is on the analytic structure.

Special configurations include:

• Equal-mass case ($m_1=m_2=m_3$): Central for explicit all-order results in the dimensional regulator (Adams et al., 2015).
• Two-mass pseudo-threshold ($q^2 = -m^2$): Appears in non-relativistic QCD and admits explicit elliptic polylogarithmic representation (Kotikov, 2022, Campert et al., 2020).

2. Algebraic and Elliptic-Geometric Structure

The locus $\mathcal{F} = 0$ for generic kinematics defines a family of genus one (elliptic) curves in $\mathbb{P}^2$. After a birational transformation, this can be brought to Weierstrass form $y^2 = 4x^3 - g_2 x - g_3$ (Adams et al., 2013, Adams et al., 2014).

Key features:

• The discriminant of the elliptic curve governs the singularity structure in $t$ ($p^2$).
• The periods $\psi_1, \psi_2$ of the elliptic curve, with ratio $\tau = \psi_2/\psi_1$, are central to the analytic representation of the integral.
• In two-mass sunrise configurations (e.g., pseudo-threshold), the curve is expressed as a quartic $y^2 = (x-a_1)\dots(x-a_4)$, with $t$-dependent branch points (Kotikov, 2022).

3. Differential Equations and Canonical $\epsilon$-Form

The master integrals of the sunrise topology satisfy Fuchsian systems of differential equations. In $D=2$, a single second-order inhomogeneous Picard–Fuchs ODE arises: $$p_0(t)\frac{d^2S}{dt^2} + p_1(t)\frac{dS}{dt} + p_2(t) S = p_3(t)$$ with $p_0(t)$ encoding the location of physical and pseudo-thresholds (Müller-Stach et al., 2011, Adams et al., 2013, Müller-Stach et al., 2012).

For arbitrary masses and $D=2-2\epsilon$:

• The sunrise system can be systematically put into a canonical $\epsilon$-form after an appropriate change of basis, making it amenable to all-order expansions in $\epsilon$, with uniform weight structures (Bogner et al., 2019, Adams et al., 2015).

The pure $\epsilon$-form is expressed in terms of logarithmic one-forms (dlogs) associated to the moduli of the underlying genus-one curve or the marked points on the moduli space $\overline{\mathcal{M}}_{1,3}$ (Bogner et al., 2019, Giroux et al., 2022).

4. Analytic Solution: Elliptic Polylogarithms and Iterated Integrals

The general solution is constructed using the fundamental periods of the associated elliptic curve. For $D=2$, the solution for generic masses is (Adams et al., 2014, Adams et al., 2015): $$S(2,t) = \frac{1}{\pi} \psi_1(t) \sum_{i=1}^3 E_{2;0}\big(w_i(t); -1; -q(t)\big)$$ where $E_{2;0}$ is the elliptic dilogarithm, $w_i(t)$ are algebraic arguments associated to the intersection of the Feynman parameter simplex with the cubic, and $q(t) = \exp(i\pi\tau)$ is the elliptic nome.

For higher $\epsilon$ orders and arbitrary masses, the result involves:

• Elliptic multiple polylogarithms (eMPLs), defined recursively via iterated integrals over simple-pole kernels on the elliptic curve (Campert et al., 2020, Kotikov, 2022).
• Shuffle algebra structures and the complete system of elliptic iterated integrals (as established by Broedel, Duhr, Dulat, and Tancredi).

For $\epsilon$-expansions: $$S(D=2-2\epsilon, t) = S^{(0)}(t) + \epsilon S^{(1)}(t) + \epsilon^2 S^{(2)}(t) + \dots$$ Each $S^{(k)}(t)$ is a finite $\mathbb{Q}$-linear combination of weight-$(k+2)$ eMPLs and their products, with all inhomogeneities in the ODE translating into weight increases in the iterated integration (Adams et al., 2015, Adams et al., 2015).

In the pseudo-threshold kinematics of the two-mass sunrise, each term in the expansion is a sum of eMPLs with kernels derived from the quartic roots and their associated cross ratios (Kotikov, 2022, Campert et al., 2020).

5. Special Cases and Moduli-Space Interpretation

In the equal-mass case ($m_1 = m_2 = m_3$), the situation simplifies considerably:

• The three marked points coincide on the elliptic curve, and all integration kernels are modular forms of a congruence subgroup ($\Gamma_0(6)$ or $SL_2(\mathbb{Z})$).
• The solution reduces to iterated integrals of modular forms (Adams et al., 2015, Bogner et al., 2019).

For generic masses, the moduli space $\overline{\mathcal{M}}_{1,3}$, the tri-punctured torus, naturally parametrizes the external invariants and mass ratios. The solution as iterated integrals over this moduli space provides a geometric interpretation of the function space needed to evaluate the general two-loop sunrise (Bogner et al., 2019).

Boundary conditions are fixed by Feynman parameter evaluations at $t=0$ (the vacuum limit), which involve classical polylogarithms and Clausen-type functions (Adams et al., 2014, Müller-Stach et al., 2011).

6. Analytic Continuation and Physical Regions

The analytic continuation in $t$ is controlled by the monodromy of the periods on the elliptic curve and the branch structure of the elliptic polylogarithms. Singularities occur at physical and pseudo-thresholds $t = (m_1 \pm m_2 \pm m_3)^2$, corresponding to the collision of curve branch points (Adams et al., 2013, Adams et al., 2015).

For $t$ real and above threshold, the branch cuts and their discontinuities match those predicted by Cutkosky rules and unitarity. The function space of solutions with correct monodromy is explicitly characterized by the eMPL shuffle algebra (Campert et al., 2020, Kotikov, 2022).

For the two-mass sunrise at pseudo-threshold, $t>0$ is a real sheet with no $i0$ prescription needed. Analytic continuation to $t<0$ or complex $t$ tracks the eMPL monodromy as the quartic branch points braid in the $x$-plane, implemented via the shuffle algebra (Kotikov, 2022).

7. Impact and Broader Connections

The two-loop sunrise integral established the essential role of elliptic polylogarithms as a function space for Feynman amplitudes beyond multiple polylogarithms, motivating the systematic development of the theory of eMPLs, their functional relations, and modular properties (Adams et al., 2015, Adams et al., 2015, Adams et al., 2014, Bogner et al., 2019).

Applications extend to higher-loop vacuum diagrams, non-relativistic QCD corrections (paradigmatic in pseudo-threshold kinematics), and connections with mirror symmetry, motivic cohomology, and regulators (Bloch et al., 2016).

The Picard–Fuchs/motive-based differential equation method was shown to drastically reduce the complexity of these integrals, producing second-order rather than fourth-order equations compared to standard IBP reduction, and manifesting the geometric origin of the analytic structure (Müller-Stach et al., 2011, Müller-Stach et al., 2012). The structure of the $\epsilon$-form and the unified eMPL framework now serve as templates for all subsequent multi-scale, multiloop computations in elliptic and higher-genus cases.

---

Key References:

• "Sunrise integral in non-relativistic QCD with elliptics" (Kotikov, 2022)
• "Sunrise integral with two internal masses and pseudo-threshold kinematics in terms of elliptic polylogarithms" (Campert et al., 2020)
• "The unequal mass sunrise integral expressed through iterated integrals on $\overline{\mathcal M}_{1,3}$" (Bogner et al., 2019)
• "The iterated structure of the all-order result for the two-loop sunrise integral" (Adams et al., 2015)
• "The two-loop sunrise graph with arbitrary masses" (Adams et al., 2013)
• "A second-order differential equation for the two-loop sunrise graph with arbitrary masses" (Müller-Stach et al., 2011)

These works, along with supporting research on modular and motivic structures, constitute the core literature for the two-loop sunrise integral and its role in the modern analytic theory of Feynman integrals.