Papers
Topics
Authors
Recent
Search
2000 character limit reached

Vector Boson Fusion: Precision in Collider Studies

Updated 3 December 2025
  • Vector Boson Fusion is an electroweak process where quarks emit space-like bosons that fuse into color-singlet states, crucial for probing Higgs and electroweak properties.
  • It is characterized by two high-energy forward jets with large rapidity separation and suppressed central hadronic activity, which aids in background suppression.
  • Advanced analysis techniques, including high-order QCD corrections and multivariate methods, significantly enhance the sensitivity to both standard and beyond-the-standard-model signals.

Vector Boson Fusion (VBF) is a key electroweak production mechanism at high-energy colliders, characterized by the fusion of two space-like electroweak gauge bosons radiated from incoming quarks. VBF provides a unique laboratory for precision tests of the Standard Model (SM), especially in the Higgs and multi-boson sectors, and offers distinctive experimental signatures that can be exploited for background suppression and new physics searches.

1. Theoretical Foundations of VBF

VBF refers to processes in which two (anti-)quarks radiate space-like weak bosons (W±W^\pm or ZZ), which subsequently fuse to produce a color-singlet final state XX, such as a Higgs boson or another electroweak state. The process is dominated by tt-channel diagrams: q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X) At leading order, the amplitude factorizes into two fermionic currents convoluted with the VVXVVX vertex, and the absence of QCD color exchange between the two quark lines implies suppressed central hadronic activity and two energetic, forward tagging jets. The inclusive VBF cross section can be written in the structure-function approach as a double deep-inelastic scattering (DIS) convolution: σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right] where Fa,bF_{a,b} are DIS structure functions, Ca,bC_{a,b} the relevant Wilson coefficients, and K^\hat{K} encodes electroweak fusion into ZZ0 (Dreyer et al., 2016, Bolzoni et al., 2011).

Numerically, at the LHC with ZZ1 TeV the inclusive SM VBF Higgs cross section is ZZ2 pb, compared to ZZ3 pb for gluon-fusion (ggF) production at NZZ4LO. The characteristic Born-level partonic cross section features a propagator structure ZZ5, reflecting the ZZ6-channel nature of the interaction (Chan et al., 2017).

VBF's sensitivity to the ZZ7 vertex at tree level and dominance by longitudinal weak bosons (ZZ8, ZZ9) make it the "cleanest" collider probe of electroweak symmetry breaking (EWSB), allowing direct access to deviations from the SM structure.

2. Experimental Signatures and Event Selection

The VBF topology is defined by two high-XX0 forward jets with large rapidity separation (XX1) and large invariant mass (XX2), with suppressed hadronic activity in the central region due to color-singlet exchange. Standard experimental selections include:

  • Jet tagging: XX3–XX4 GeV, XX5
  • Forward jets: XX6–XX7
  • Dijet invariant mass: XX8 GeV, often up to 600–700 GeV
  • Central jet veto: no jets above threshold XX9 in the rapidity interval between the tagging jets

These selections yield high VBF purity but are contaminated by "fake" VBF events from ggF+tt0, which can emulate forward jets via QCD radiation (Chan et al., 2017, Buckley et al., 2021).

To quantify contamination, the ratio tt1 is used. In ATLAS VBF selections this contamination is tt2, motivating advanced analysis techniques for background rejection.

3. Higher-Order Corrections and Precision Predictions

VBF is amenable to high-precision theoretical predictions. The structure-function approach enables calculations up to Ntt3LO in QCD, incorporating all factorizable corrections: tt4 Non-factorizable QCD corrections (e.g., double-gluon exchange between quark lines) are suppressed by color and phase-space factors, yielding sub-permille corrections to the total cross section at the LHC (Gates, 2023, Dreyer et al., 2016, Cruz-Martinez et al., 2018). Electroweak corrections at NLO reduce the inclusive cross section by tt5, with larger effects in the high-tt6 tails.

Scale uncertainties are highly suppressed: at Ntt7LO, the scale uncertainty falls below tt8 permille (Dreyer et al., 2016). PDF uncertainties at NNLO are at the few percent level. This theoretical control makes VBF critical for interpreting precision Higgs and EWSB measurements.

4. Background Discrimination: Jet-Shape Variables and Multivariate Techniques

To disentangle VBF from ggF+tt9, which can pass VBF-style cuts, advanced discriminants are exploited:

  • Jet-shape observables: VBF jets initiated by quarks are narrower and harder, while ggF (gluon-initiated) jets are broader and softer. The girth q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)0 and the integrated jet shape q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)1 (with q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)2, q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)3 for central and side regions of the jet cone) provide discrimination:

q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)4

  • Multivariate analysis: A two-step boosted decision tree (BDT) classifier is employed. The first step distinguishes VBF from all non-ggF backgrounds, while the second focuses on separating VBF from ggF, using a powerful subset of variables including girth and jet shapes. This approach reduces the ggF fake fraction by a factor q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)5 at fixed signal efficiency relative to standard single-step BDTs (Chan et al., 2017).

Example results for q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)6: | Selection | VBF Purity | ggF Contamination | |------------------------|------------|-------------------| | Standard BDT | 76.4% | 12.4% | | Two-step BDT | 77.1% | 7.9% |

5. Extensions: Beyond-the-Standard-Model and Higher-Point Couplings

VBF is sensitive to a wide class of BSM modifications:

  • Anomalous q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)7 couplings: Modifications in the tensor structure, including momentum-dependent operators (dimension-6 effective field theory), alter the kinematic distributions of tagging jets, e.g., harder q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)8, narrower rapidity gap, and modified acceptance under standard VBF cuts (Djouadi et al., 2013).
  • Quartic q1q2q1q2+(VVX)q_1 q_2 \rightarrow q_1' q_2' + (V^* V^* \rightarrow X)9 couplings: VBF Higgs pair production (VVXVVX0) provides unique access to the VVXVVX1 vertex. The cross section is highly sensitive to deviations VVXVVX2, exhibiting rapid growth at large di-Higgs invariant mass (Bishara et al., 2016). HL-LHC can achieve VVXVVX3 precision on VVXVVX4; VVXVVX5 at a 100 TeV collider.
  • EFT and form factors: High-VVXVVX6 VBF (removing standard upper jet cuts) is sensitive to new contact operators (VVXVVX7 growing with VVXVVX8) and compositeness form factors. Current sensitivity in VVXVVX9 at HL-LHC translates to σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]0–σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]1 TeV for relevant Wilson coefficients (Han et al., 2023).

6. Experimental Results and Future Prospects

VBF has been experimentally observed in multiple channels:

  • Higgs boson: σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]2, σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]3, σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]4, and invisible decays using VBF selection reach sub-10% accuracy in signal-strength measurements (Pigard, 2017, Chan et al., 2017).
  • Electroweak bosons: W, Z, and diboson production via VBF and fully leptonic decays have been established, including precision measurements in ZσVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]5jj (Pigard, 2017).
  • Dark matter and invisible decays: VBF + σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]6 is a leading LHC search for Higgs-portal and electroweak DM (Duque-Escobar et al., 2021, Brooke et al., 2016).
  • Heavy resonances: For σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]7 (heavy vector), VBF production outpaces Drell–Yan for σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]8–σVBF=dx1dx2 [Fa(x1,Q12)Ca(x1,Q12;αs)]K^(Q1,Q2)[Fb(x2,Q22)Cb(x2,Q22;αs)]\sigma_{\text{VBF}} = \int dx_1\,dx_2\ \left[F_a(x_1,Q_1^2) \otimes C_a(x_1,Q_1^2;\alpha_s)\right]\,\hat{K}(Q_1, Q_2)\,\left[F_b(x_2,Q_2^2) \otimes C_b(x_2,Q_2^2;\alpha_s)\right]9 TeV, highlighting its importance in high-mass searches at HL-LHC (Baker et al., 2022).

VBF remains central to future collider programs, as at multi-TeV lepton colliders VBF processes become dominant, enabling percent-level and sub-percent-level measurements of Higgs, EW, and BSM parameters (Costantini et al., 2020).

7. Outlook and Developments

Current and future directions in VBF phenomenology and experimentation include:

  • Advanced background discrimination via deep learning, jet substructure, and higher-point correlation observables.
  • Incorporation of NLO electroweak and full NFa,bF_{a,b}0LO QCD corrections, and further reduction of theory uncertainties.
  • Model-independent fits of differential distributions to constrain EFT operators and search for anomalous couplings.
  • Precision measurements at HL-LHC and future 100 TeV facilities, enabling sensitivity to tiny deviations in the electroweak sector and robust discrimination among new physics scenarios.

VBF thus constitutes a pillar of the collider EWSB program, providing unparalleled access to the Higgs–gauge sector, quartic and higher gauge couplings, and as a discovery tool for both SM and BSM physics (Rauch, 2016, Dreyer et al., 2016, Chan et al., 2017).

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Vector Boson Fusion.