{
  "id": "poincare-lemma",
  "title": "Poincaré Lemma",
  "lang": "en",
  "source_lang": "en",
  "summary": "The Poincaré lemma states that every closed differential form of positive degree is locally exact, making topology the source of global obstructions to finding primitives.",
  "category": "mathematics",
  "subcategory": "geometry",
  "type": "concept",
  "aliases": [
    "Poincaré’s lemma"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/poincare-lemma/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/poincare-lemma/",
      "json": "https://aiwiki.page/zh/poincare-lemma.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "The Poincaré lemma is a fundamental theorem of differential geometry stating that every closed differential form of positive degree is locally exact. More concretely, a smooth differential form whose exterior derivative vanishes can, near any point, be written as the exterior derivative of another form. On a star-shaped open subset of Euclidean space, this conclusion holds throughout the entire domain. Named after Henri Poincaré, the lemma explains why local integrability conditions can guarantee local potentials, while the existence of global potentials depends on topology."
    },
    {
      "heading": "Closed and exact forms",
      "text": "Let \\(M\\) be a smooth manifold, and let \\(\\Omega^k(M)\\) denote the space of smooth differential \\(k\\)-forms on \\(M\\). The exterior derivative is an operator\n\n\\[d:\\Omega^k(M)\\longrightarrow\\Omega^{k+1}(M)\\]\n\nsatisfying\n\n\\[d^2=0.\\]\n\nA form \\(\\omega\\in\\Omega^k(M)\\) is:\n\n- Closed if \\(d\\omega=0\\).\n- Exact if there is a form \\(\\eta\\in\\Omega^{k-1}(M)\\) such that \\(\\omega=d\\eta\\).\n\nThe form \\(\\eta\\) is called a primitive of \\(\\omega\\). Because \\(d^2=0\\), every exact form is closed. The Poincaré lemma supplies a local converse.\n\nFor a function \\(f\\), regarded as a \\(0\\)-form, its exterior derivative in coordinates is\n\n\\[df=\\sum_{i=1}^{n}\\frac{\\partial f}{\\partial x^i}\\,dx^i.\\]\n\nThus, finding a primitive of a \\(1\\)-form generalizes finding an antiderivative of a function of one variable."
    },
    {
      "heading": "Statement and scope",
      "text": "Local form\n\nLet \\(\\omega\\) be a smooth \\(k\\)-form on a smooth manifold \\(M\\), with \\(k\\geq1\\), and suppose \\(d\\omega=0\\). For every point \\(p\\in M\\), there is an open neighborhood \\(V\\) of \\(p\\) and a smooth \\((k-1)\\)-form \\(\\eta\\) on \\(V\\) such that\n\n\\[\\omega|_V=d\\eta.\\]\n\nThe assertion is local: different neighborhoods may require different primitives, and those primitives need not combine into one globally defined form.\n\nStar-shaped domains\n\nA subset \\(U\\) of Euclidean space is star-shaped with respect to \\(a\\in U\\) if\n\n\\[a+t(x-a)\\in U \\qquad \\text{for every }x\\in U,\\quad 0\\leq t\\leq1.\\]\n\nIf \\(U\\subseteq\\mathbb R^n\\) is open and star-shaped, then every smooth closed \\(k\\)-form on \\(U\\), for \\(k\\geq1\\), is exact on all of \\(U\\). Every open convex set is star-shaped, so the lemma applies in particular to open balls and to \\(\\mathbb R^n\\).\n\nThe local version follows by choosing coordinates near a point and restricting to a coordinate neighborhood corresponding to an open ball.\n\nDegree zero\n\nThe restriction to positive degree is essential. A closed \\(0\\)-form is a smooth function satisfying \\(df=0\\), which means that it is locally constant. On a connected manifold, it is constant. Under the usual convention that there are no forms of degree \\(-1\\), a nonzero constant function is not exact."
    },
    {
      "heading": "Proof by radial integration",
      "text": "The star-shaped version has an explicit proof that constructs a primitive by integrating along straight lines.\n\nAssume first that \\(U\\) is star-shaped with respect to the origin. For a \\(k\\)-form \\(\\omega\\), \\(k\\geq1\\), define a \\((k-1)\\)-form \\(K\\omega\\) by\n\n\\[(K\\omega)_x(v_1,\\ldots,v_{k-1}) = \\int_0^1 t^{k-1}\\, \\omega_{tx}(x,v_1,\\ldots,v_{k-1})\\,dt.\\]\n\nHere tangent vectors at different points of \\(U\\) are identified with vectors in \\(\\mathbb R^n\\).\n\nThe operator \\(K\\) integrates the form along the radial contraction\n\n\\[H(t,x)=tx.\\]\n\nIt satisfies the homotopy identity\n\n\\[dK\\omega+Kd\\omega=\\omega \\qquad (k\\geq1).\\]\n\nIf \\(\\omega\\) is closed, then \\(d\\omega=0\\), and hence\n\n\\[\\omega=d(K\\omega).\\]\n\nThus \\(K\\omega\\) is a primitive.\n\nIn degree zero, the corresponding identity is\n\n\\[K(df)=f-f(0),\\]\n\nwhich recovers the fundamental theorem of calculus along the segment from \\(0\\) to \\(x\\).\n\nFor a domain star-shaped about a point \\(a\\), the same construction uses \\(H(t,x)=a+t(x-a)\\). This explicit operator makes the lemma more than an existence theorem: it gives a formula for a primitive."
    },
    {
      "heading": "Interpretation in vector calculus",
      "text": "In familiar low-dimensional settings, the Poincaré lemma yields the local existence of scalar and vector potentials.\n\nCurl-free fields and scalar potentials\n\nLet\n\n\\[\\alpha=P\\,dx+Q\\,dy+R\\,dz\\]\n\nbe a smooth \\(1\\)-form on an open subset of \\(\\mathbb R^3\\). The condition \\(d\\alpha=0\\) is equivalent to\n\n\\[\\nabla\\times\\mathbf F=0, \\qquad \\mathbf F=(P,Q,R).\\]\n\nThe lemma implies that locally there is a function \\(f\\) such that\n\n\\[\\alpha=df, \\qquad \\mathbf F=\\nabla f.\\]\n\nThus a curl-free vector field is locally a gradient. On a star-shaped domain, the potential can be chosen globally.\n\nFor a \\(1\\)-form \\(\\alpha=\\sum_i a_i(x)\\,dx^i\\) on a domain star-shaped about the origin, the radial formula gives\n\n\\[f(x)=\\int_0^1\\sum_i a_i(tx)x^i\\,dt.\\]\n\nWhen \\(d\\alpha=0\\), this function satisfies \\(df=\\alpha\\).\n\nDivergence-free fields and vector potentials\n\nAssociate a vector field \\(\\mathbf B=(B_1,B_2,B_3)\\) with the \\(2\\)-form\n\n\\[\\beta = B_1\\,dy\\wedge dz + B_2\\,dz\\wedge dx + B_3\\,dx\\wedge dy.\\]\n\nThen\n\n\\[d\\beta = (\\nabla\\cdot\\mathbf B)\\,dx\\wedge dy\\wedge dz.\\]\n\nIf \\(\\nabla\\cdot\\mathbf B=0\\), the lemma gives locally a \\(1\\)-form\n\n\\[\\alpha=A_1\\,dx+A_2\\,dy+A_3\\,dz\\]\n\nsuch that \\(\\beta=d\\alpha\\). In vector notation,\n\n\\[\\mathbf B=\\nabla\\times\\mathbf A.\\]\n\nThis is the local mathematical basis for describing a divergence-free magnetic field by a vector potential. Global existence requires additional conditions on the domain."
    },
    {
      "heading": "Why global exactness can fail",
      "text": "Consider the punctured plane\n\n\\[U=\\mathbb R^2\\setminus\\{(0,0)\\}\\]\n\nand the \\(1\\)-form\n\n\\[\\omega = \\frac{-y\\,dx+x\\,dy}{x^2+y^2}.\\]\n\nDirect differentiation gives \\(d\\omega=0\\). Locally, \\(\\omega=d\\theta\\), where \\(\\theta\\) is a suitable branch of the polar angle.\n\nHowever, on the counterclockwise unit circle \\(C\\),\n\n\\[\\int_C\\omega=2\\pi.\\]\n\nIf \\(\\omega=df\\) for a globally defined smooth function \\(f\\), its line integral around every closed curve would be zero. Therefore \\(\\omega\\) is closed but not exact on \\(U\\).\n\nThe obstruction is not a failure of smoothness or of the equation \\(d\\omega=0\\). It comes from the hole in the domain, which prevents a single-valued global choice of angle.\n\nSimple connectedness and its limits\n\nOn a simply connected smooth manifold, every closed \\(1\\)-form is exact. For \\(1\\)-forms, simple connectedness therefore provides a useful sufficient condition beyond star-shapedness.\n\nIt does not guarantee exactness in higher degrees. The sphere \\(S^2\\) is simply connected, but its standard area \\(2\\)-form \\(\\sigma\\) is not exact. It is closed because all \\(3\\)-forms on a two-dimensional manifold vanish. Yet\n\n\\[\\int_{S^2}\\sigma=4\\pi.\\]\n\nIf \\(\\sigma=d\\eta\\), Stokes’ theorem would instead imply\n\n\\[\\int_{S^2}\\sigma = \\int_{\\partial S^2}\\eta = 0,\\]\n\nsince the sphere has no boundary.\n\nThus, higher-degree forms can detect global obstructions not captured by simple connectedness."
    },
    {
      "heading": "De Rham cohomology and homotopy",
      "text": "The difference between closed and exact forms is measured by de Rham cohomology:\n\n\\[H_{\\mathrm{dR}}^k(M) = \\frac{ \\ker\\!\\left(d:\\Omega^k(M)\\to\\Omega^{k+1}(M)\\right) }{ \\operatorname{im}\\!\\left(d:\\Omega^{k-1}(M)\\to\\Omega^k(M)\\right) }.\\]\n\nThe Poincaré lemma says that these groups vanish in positive degree on sufficiently small coordinate neighborhoods. Nonzero global groups therefore record obstructions to assembling local primitives into global ones.\n\nA broader result is the homotopy invariance of de Rham cohomology. If two smooth maps \\(f_0,f_1:M\\to N\\) are smoothly homotopic, their pullback operators satisfy an identity of the form\n\n\\[f_1^*-f_0^*=dK+Kd.\\]\n\nConsequently, the two maps induce the same map on de Rham cohomology. In particular, a contractible smooth manifold has\n\n\\[H_{\\mathrm{dR}}^k(M)=0 \\qquad (k>0),\\]\n\nso every closed positive-degree form on it is exact.\n\nThe local lemma is also a key ingredient in de Rham’s theorem, which identifies de Rham cohomology with singular cohomology with real coefficients and connects differential calculus with algebraic topology."
    },
    {
      "heading": "Applications and qualifications",
      "text": "The lemma provides a local solvability result for the equation\n\n\\[d\\eta=\\omega:\\]\n\nthe necessary compatibility condition \\(d\\omega=0\\) is locally sufficient. This viewpoint appears in systems of partial differential equations, potential theory, and differential geometry.\n\nSeveral qualifications matter:\n\n- Primitives are not unique. If \\(d\\eta=\\omega\\), then \\(d(\\eta+\\gamma)=\\omega\\) for any closed form \\(\\gamma\\) of the appropriate degree. For scalar potentials on a connected domain, this ambiguity is an additive constant.\n- Local exactness does not establish global exactness. The topology of the domain must be considered separately.\n- Extra constraints require separate analysis. The lemma does not automatically produce a primitive satisfying prescribed boundary values or having compact support.\n- The standard statement concerns smooth forms. Versions for weakly differentiable forms and distributions require a specified function-space setting and suitable interpretations of the exterior derivative.\n\nThese distinctions separate the local integrability problem solved by the lemma from global, boundary, and regularity problems."
    }
  ],
  "links": [
    "differential-geometry",
    "differential-form",
    "topology",
    "manifold",
    "exterior-derivative",
    "antiderivative",
    "euclidean-space",
    "convex-set",
    "fundamental-theorem-of-calculus",
    "vector-field",
    "gradient",
    "magnetic-field",
    "circle",
    "line-integral",
    "stokes-theorem",
    "de-rham-cohomology",
    "algebraic-topology",
    "partial-differential-equation"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}