D-term scalar potential ($V_D$)

Eliminating the real auxiliary field $D$ gives a nonnegative D-term scalar potential. For a $U(1)$ theory with charges $q_i$ and a Fayet–Iliopoulos parameter $\xi$, one convention gives $V_D=\frac{1}{2}[g{\sum }_i{q}_i|{\varphi }_i{|}^2+\xi {]}^2$.

For one charged chiral superfield, canonical kinetic terms, no nonconstant gauge-invariant superpotential, and $g>0$, the sign of $q\xi$ selects the vacuum. If $q\xi >0$, the minimum has $\varphi =0$, positive vacuum energy, unbroken internal gauge symmetry and a gaugino Goldstino; the scalar squared mass is $gq\xi$ while its chiral fermion is massless. If $q\xi <0$, a scalar condensate cancels the auxiliary field, preserving supersymmetry while Higgsing the gauge group. These are classical statements; a consistent quantum charged spectrum must cancel its gauge anomalies.

For canonical kinetic terms, vanishing superpotential, $g>0$ and $q\xi >0$, the D-term scalar potential $V=\frac{1}{2}(\xi +gq|\varphi {|}^2{)}^2$ has its minimum at $\varphi =0$. Both real scalar modes have squared mass $gq\xi$, while the chiral Weyl fermion, photon and gaugino are massless; the gaugino is the goldstino. The scalar-to-fermion squared-mass splitting is $gq\xi$. On the other branch $q\xi <0$, $v^2=-\xi /(gq)$ cancels $D$ and preserves supersymmetry; the Higgsed vector, remaining real scalar and Dirac fermion have common squared mass $2g^2{q}^2{v}^2$. A single nonzero-charge chiral fermion has a gauge anomaly, so a quantum realization requires an anomaly-cancelling completion; these are the stated classical vacuum spectra.

For canonical Kähler potential, set $a=|{\varphi }_0{|}^2$, $b=|{\varphi }_{+}{|}^2$, $c=|{\varphi }_{-}{|}^2$. The supplied supergravity F-term potential is $|\lambda {|}^2{e}^{{\kappa }^2(a+b+c)}[ab+ac+bc+3{\kappa }^{2}abc+{\kappa }^{4}abc(a+b+c)]$. Adding $g(b-c-\zeta {)}^2$ with $g>0$, $\lambda \ne 0$ and $\zeta >0$ gives a supersymmetric vacuum at $a=c=0$, $b=\zeta$. The branch $b=c=0$ has positive energy $g{\zeta }^2$, D-term breaking and a flat direction of a scalar potential in ${\varphi }_0$. Its charged squared masses are $|\lambda {|}^{2}a{e}^{{\kappa }^{2}a}\mp 2g\zeta$, so it is a local minimum valley only when both are positive. For negative $\zeta$, exchange the charged fields. These are branches rather than two isolated minima; a conventional constant Fayet–Iliopoulos term additionally needs a consistent supergravity gauge completion.

A Fayet–Iliopoulos term is a gauge-invariant term linear in the auxiliary $D$ field of an Abelian vector multiplet. Supersymmetry is broken only if no charged-scalar configuration can make the shifted D-term vanish simultaneously with every F-term.

In the conventional two-derivative supergravity action described by a Kähler potential, superpotential and gauge kinetic function, a constant Fayet–Iliopoulos term cannot be added independently of the gauging. It makes the Abelian gauging an R-symmetry gauging, and the superpotential must have the corresponding gauge transformation. Thus the rigid Abelian constant is subject to extra local consistency conditions. This statement concerns the conventional matter-coupled action, not distinct nonstandard FI constructions.

In an exactly supersymmetric Abelian theory a one-loop auxiliary-field tadpole can generate a cutoff-dependent Fayet–Iliopoulos term proportional to the trace of the matter charges. Gauge invariance with arbitrary background spurions forbids a general coupling-dependent full-superspace coefficient multiplying the vector superfield. In the supersymmetry-preserving Wilsonian setting the additive perturbative correction is exhausted by the one-loop term; it vanishes for a traceless charge generator. Canonical vector normalization and soft breaking require separate treatment. For a simple non-Abelian group the term itself is forbidden.

A constant linear auxiliary-field term ${\xi }_a{D}^a$ is gauge-invariant only when $\xi$ is an invariant vector in the dual adjoint representation. Equivalently, it annihilates all Lie brackets. A simple non-Abelian Lie algebra equals its commutator algebra and has no such nonzero vector. Abelian factors do admit the term. Identifying a Fayet–Iliopoulos parameter in a simple non-Abelian theory therefore requires setting it to zero, rather than treating it as an arbitrary coupling.