Transverse Momentum Disbalance in High-Energy Collisions

• Transverse momentum disbalance is the nonzero vector sum of final-state particle momenta in high-energy collisions, probing deviations in momentum conservation and underlying QCD dynamics.
• Observable definitions such as dijet asymmetry (A_J), ratio variable (x_J), and missing transverse momentum are used to quantify jet quenching effects and soft hadron redistribution.
• Experimental methods including event shape analysis, TMD factorization, and precise background subtraction enable disentangling medium effects and nonperturbative dynamics in both proton and heavy-ion collisions.

Transverse momentum disbalance refers to the nonzero vector sum of final-state particle transverse momenta in hard-scattering events, particularly in high-energy nuclear and hadronic collisions. It is a critical observable for quantifying the degree to which momentum conservation is satisfied among detected particles and for probing underlying quantum chromodynamics (QCD) dynamics, including parton energy loss, multi-parton interactions, collective flow, and initial-state transverse momentum effects. Disbalance is defined and operationalized differently depending on the studied final state—dijet, multi-particle, or two-particle correlations—but universally encodes the interplay of hard scattering, medium response, and hadronization processes.

1. Formal Definitions and Observables

The transverse momentum disbalance is formalized through several complementary observables, tailored to event topology:

• Dijet Asymmetry $A_J$:

$A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$

where $p_{T,1}$ and $p_{T,2}$ are transverse momenta of the leading and subleading jets. $A_J=0$ corresponds to a balanced pair; $A_J>0$ signals an imbalance, typically due to parton energy loss in a medium or hard QCD radiation (Gao et al., 2016, Collaboration, 2015, Collaboration, 2018, Collaboration et al., 2016).

• Projected Disbalance $\langle \slashed{p}_T^{\parallel} \rangle$:

$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$

summing over all charged hadrons, probing where "lost" jet energy reappears among low-$p_T$ fragments and at which angular separations (Gao et al., 2016).

• Ratio Variable $x_J \equiv p_{T,2}/p_{T,1}$: Used for inclusive and $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$0-tagged dijets, with $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$1 (Collaboration, 2018).
• Multiparticle Disbalance $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$2:

$A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$3

where the sum is over the $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$4 highest-energy clusters. Tightly constraining $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$5 selects topologies close to momentum balance and enhances alignment observables (Lokhtin et al., 2024).

• Two-Particle Momentum Balance Function $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$6:

$A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$7

where $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$8 involves $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$9, sensitive to charge conservation and the local or nonlocal character of resultant momentum balancing (Behera et al., 27 Aug 2025).

• TMD-based Dijet Imbalance:

$p_{T,1}$0

with $p_{T,1}$1 a parton TMD, and refined via Bessel weighting to regularize UV divergences (Boer, 2014, Signal et al., 2021).

2. Experimental Contexts and Computational Methodologies

The measurement and characterization of transverse momentum disbalance span a range of experimental setups:

• Dijet and Multiparticle Topologies: ATLAS, CMS, and STAR exploit anti-$p_{T,1}$2 jet clustering (FastJet), with selection cuts on $p_{T,1}$3, $p_{T,1}$4, and $p_{T,1}$5 for defining balanced and imbalanced dijet events. Background subtraction (iterative median pedestal, HF/Voronoi), jet energy corrections, b-tagging, and pileup mitigation are systematically applied (Collaboration, 2015, Collaboration, 2018, Collaboration et al., 2016).
• Charged Hadron Correlations: Momentum projections and annular sampling around jet axes, as in CMS, are implemented by binning in $p_{T,1}$6 and angular intervals ($p_{T,1}$7) to map the spatial redistribution of momentum (Collaboration, 2015, Gao et al., 2016).
• Event Shape Engineering: Event classification via transverse spherocity $p_{T,1}$8, separating jet-like (low $p_{T,1}$9) and isotropic (high $p_{T,2}$0) topologies, allows for isolation of hard-process-driven and collective flow effects in $p_{T,2}$1 collisions (Behera et al., 27 Aug 2025).
• Event Selection by Disbalance Cut: In multiparticle alignment studies, event samples are filtered by imposing a maximal allowed $p_{T,2}$2, thereby biasing the topology towards nearly collinear or back-to-back configurations (Lokhtin et al., 2024).
• Missing Transverse Momentum ($p_{T,2}$3): At the analysis level, modern frameworks reconstruct $p_{T,2}$4 as $p_{T,2}$5 over all identified objects and soft terms, encoding both hard and invisible energy/momentum flow. Overlap resolution and systematic treatments (e.g., METMaker in ATLAS) provide the necessary precision for dark matter and $p_{T,2}$6 boson mass analyses (Balunas et al., 2023).
• TMD Factorization and Evolution: For small $p_{T,2}$7 processes (e.g., Drell-Yan, SIDIS), the cross section differential in transverse momentum is obtained via a convolution of TMDs, often Gaussian, with evolution up to the experimental hard scale using Sudakov kernels and nonperturbative $p_{T,2}$8, directly predicting $p_{T,2}$9 and $A_J=0$0 (Signal et al., 2021).

3. Mechanisms of Momentum Redistribution

Transverse momentum disbalance captures diverse QCD and nuclear physics phenomena:

• Parton Energy Loss in Medium (Jet Quenching): In PbPb collisions at LHC and AuAu at RHIC, jet quenching manifests as enhanced $A_J=0$1 and suppressed $A_J=0$2 in central events: subleading jets lose energy via medium-induced gluon radiation and elastic scatterings. The lost momentum is predominantly transferred to soft ($A_J=0$3–$A_J=0$4 GeV) hadrons at large angles, as seen both in AMPT model calculations and CMS data. The compensation shifts from mid-$A_J=0$5 in peripheral/pp events to low $A_J=0$6 and wide angle in central heavy-ion collisions (Gao et al., 2016, Collaboration, 2015, Collaboration, 2018, Collaboration et al., 2016).
• Angular Sector Decomposition: In-cone ($A_J=0$7) negative imbalance is dominated by high-$A_J=0$8 fragments, while out-of-cone compensation arises from soft hadrons at large $A_J=0$9 in central PbPb, reflecting energy spread by elastic partonic scatterings and medium response (Gao et al., 2016).
• Heavy-Flavor Dependence: $A_J>0$0-quark dijets at LHC exhibit similar centrality-dependent imbalance to inclusive dijets, indicating that the color-charge and mass dependence of energy loss is less pronounced at high jet $A_J>0$1 and in central PbPb (Collaboration, 2018).
• Collective Effects and Multiparticle Alignment: In cosmic-ray analogs and modeled heavy-ion events, selecting configurations with small $A_J>0$2 among leading clusters strongly increases longitudinal alignment metrics, signifying a selection-induced correlation rather than necessarily indicating exotic QCD dynamics (Lokhtin et al., 2024).
• Two-Particle Balance Function Dynamics: In $A_J>0$3 collisions, the width of $A_J>0$4 narrows with increasing multiplicity, especially in isotropic (flow-dominated) events in EPOS-LHC with hydrodynamics enabled. This indicates radial flow localizes momentum-balancing charge pairs, while PYTHIA and non-hydrodynamic models show weaker dependence—demonstrating the sensitivity of disbalance observables to collective behavior even in small systems (Behera et al., 27 Aug 2025).

4. Theoretical Approaches and Intrinsic Transverse Momentum

Transverse momentum disbalance is a precision probe of parton-level QCD and nucleon structure:

• Transverse-Momentum-Dependent (TMD) Factorization:
  • For processes with back-to-back pairs (jets, heavy quarks, Higgs+jet, $A_J>0$5), the imbalance $A_J>0$6 is directly sensitive to the shape and width of parton TMDs.
  • Analytical forms relate observed $A_J>0$7 spectra and azimuthal modulations to convolutions of unpolarized ($A_J>0$8) and linearly polarized ($A_J>0$9) gluon TMDs, with cross sections manifesting characteristic angular harmonics. Bessel-weighted definitions regularize UV divergences, rendering lattice-QCD evaluation tractable and isolating the nonperturbative domain (Pisano, 2015, Boer, 2014, Signal et al., 2021).
• Scale Evolution: Intrinsic widths $\langle \slashed{p}_T^{\parallel} \rangle$0 from models such as the MIT bag approach ($\langle \slashed{p}_T^{\parallel} \rangle$10.2–0.3 GeV$\langle \slashed{p}_T^{\parallel} \rangle$2 at low scale) are evolved under CSS/Sudakov logarithms to $\langle \slashed{p}_T^{\parallel} \rangle$30.4–0.6 GeV$\langle \slashed{p}_T^{\parallel} \rangle$4 at $\langle \slashed{p}_T^{\parallel} \rangle$5–5 GeV, accurately describing Drell-Yan $\langle \slashed{p}_T^{\parallel} \rangle$6 spectra ($\langle \slashed{p}_T^{\parallel} \rangle$7 GeV$\langle \slashed{p}_T^{\parallel} \rangle$8) (Signal et al., 2021).
• Sensitivity to Nonperturbative and Perturbative Dynamics: At large $\langle \slashed{p}_T^{\parallel} \rangle$9, disbalance observables are dominated by perturbative tails ($\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$0), necessitating regularization for accessing the intrinsic transverse momentum structure (Boer, 2014).

5. Quantitative Trends and Key Results

Multiple experimental and theoretical studies have established robust, reproducible features of transverse momentum disbalance:

Observable	Central PbPb (most imbalanced)	pp/peripheral	Reference
$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$1 (dijet)	$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$2–$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$3	$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$4–$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$5	(Collaboration, 2018)
$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$6 (dijet)	$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$7	$\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$8	(Collaboration, 2018)
Compensating $\slashed{p}_T^{\parallel} = \sum_{i} \left[ -p_{T}^{i} \cos(\phi_{i} - \phi_{\rm lead}) \right]$9–$p_T$0 GeV $p_T$1	$p_T$2 GeV	Minimal	(Gao et al., 2016)
In-cone vs out-of-cone	Out-of-cone dominated by low $p_T$3	Weak in pp	(Gao et al., 2016)
$p_T$4 width (EPOS core-on)	Strong narrowing with multiplicity	Flat (PYTHIA/EPOS core-off)	(Behera et al., 27 Aug 2025)
Alignment $p_T$5 vs $p_T$6	$p_T$7 as $p_T$8	$p_T$9 (baseline)	(Lokhtin et al., 2024)

These demonstrate: in central heavy-ion collisions, the jet-induced imbalance is compensated by large-angle, low-$x_J \equiv p_{T,2}/p_{T,1}$0 hadrons; jet quenching is of comparable strength for $x_J \equiv p_{T,2}/p_{T,1}$1-jets and light-flavor jets; tuning event selection for small missing $x_J \equiv p_{T,2}/p_{T,1}$2 amplifies geometric alignment features; and collective flow effects can be isolated even in high-multiplicity $x_J \equiv p_{T,2}/p_{T,1}$3 via two-particle momentum balance analyses.

6. Interpretation, Limitations, and Model Sensitivities

• Elastic vs Inelastic Energy Loss: AMPT results indicate that elastic partonic scatterings account for much of the large-angle redistribution, but omission of medium-induced radiative processes leads to an overestimation of out-of-cone disbalance in the models relative to data. Inclusion of inelastic processes is expected to soften this surplus (Gao et al., 2016).
• Detector Effects and Corrections: Experimental extraction of disbalance requires correct handling of pileup, acceptance, tracking inefficiency, and background subtraction. Overlap-aware reconstructions (e.g., ATLAS MissingETAssociationMap) and systematic tools for object calibration are now standard (Balunas et al., 2023).
• Intrinsic vs Selection-Driven Effects: Studies in HYDJET++ reveal that event selection biases—enforcing nearly zero $x_J \equiv p_{T,2}/p_{T,1}$4 among the top $x_J \equiv p_{T,2}/p_{T,1}$5 clusters—are sufficient to produce strong alignment, suggesting a substantial fraction of observed geometric correlations can emerge from kinematic constraints plus conservation laws rather than exotic QCD effects (Lokhtin et al., 2024).
• Theoretical Limitations: Conventional definitions of mean squared imbalance are UV sensitive, dominated by perturbative hard emissions. Bessel-weighted and finite-$x_J \equiv p_{T,2}/p_{T,1}$6 regularizations are essential for extracting genuine nonperturbative information, and only in the $x_J \equiv p_{T,2}/p_{T,1}$7, $x_J \equiv p_{T,2}/p_{T,1}$8 limits does one recover the divergent traditional result (Boer, 2014).

7. Future Directions and Phenomenological Implications

• Disentangling Medium Effects: Enhanced granularity in event shape engineering (e.g., with $x_J \equiv p_{T,2}/p_{T,1}$9, jet substructure, and two-particle correlations) can cleanly separate hard-scattering from medium-induced disbalance, providing crucial experimental constraints on QGP modeling and parton shower modification.
• TMD and Flow Imaging: High-precision $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$00 and disbalance measurements in Run 3 LHC data with advanced model discrimination (PYTHIA vs EPOS core-on) will further elucidate the role and onset of collective phenomena in small systems (Behera et al., 27 Aug 2025).
• Gluon Imaging and TMD Extraction: Transverse momentum disbalance in processes such as Higgs+jet and $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$01 will increasingly serve as quantitative tools for extracting unpolarized and polarized gluon TMDs at high scales, enabling three-dimensional nucleon imaging (Pisano, 2015).
• Model Integration: Implementations such as the ATLAS dynamic MET framework demonstrate the necessity for flexible, overlap-resolving algorithms for reliable missing $A_J = \frac{p_{T,1}-p_{T,2}}{p_{T,1}+p_{T,2}}$02 and disbalance measurement and propagation of systematic uncertainties, already yielding substantial performance gains in physics analyses (Balunas et al., 2023).

Transverse momentum disbalance observables are now established as indispensable probes in both experimental and theoretical QCD, directly linking measured momentum imbalance to the fundamental mechanisms of energy loss, collective behavior, and parton structure in high-energy collisions.