Discrete Scale Invariance (DSI)

• Discrete Scale Invariance (DSI) is a symmetry under discrete rescalings that introduces log‐periodic modulations and complex critical exponents, differentiating it from continuous scaling laws.
• DSI underlies phenomena in quantum, statistical, and classical systems, with notable applications in Efimov physics, topological materials, and fractal structures observed via STM imaging and magnetotransport experiments.
• Emerging from renormalization group limit cycles and scale anomalies, DSI provides a unifying framework for understanding hierarchical structures and complex scaling behaviors across diverse experimental and theoretical models.

Discrete Scale Invariance (DSI) is a hierarchical symmetry under rescalings by specific discrete factors, rather than continuous dilations. Unlike conventional scale invariance, DSI produces log-periodic corrections to scaling laws and geometric energy spectra, and is intimately connected to scale anomalies, limit cycles in renormalization group flows, and the emergence of complex critical exponents. DSI appears in a diverse range of quantum, statistical, condensed matter, dynamical, and even cosmological systems, where it serves as a unifying framework for understanding phenomena such as Efimov physics, log-periodic quantum oscillations in topological materials, and fractal structures in both classical and quantum domains.

1. Mathematical Structure and Phenomenology

At the core, continuous scale invariance (CSI) asserts that an observable $f(x)$ obeys $f(\lambda x) = \lambda^{\alpha} f(x)$ for any $\lambda > 0$, leading to power-law scaling and real critical exponents. In contrast, DSI restricts this symmetry to a discrete set: $f(\lambda_0^n x) = \mu^n f(x)$ for some $\lambda_0 > 1$ and integer $n$ (Ovdat et al., 2019). The general solution is no longer a pure power law, but rather

$$f(x) = x^\gamma G\left(\frac{\ln x}{\ln \lambda_0}\right),$$

where $G(u)$ is periodic, encoding log-periodic modulations. Complex critical exponents emerge as a direct consequence (Ovdat et al., 2019). This log-periodic structure is the key visible signature of DSI in experiment and phenomenology.

DSI is intimately related to renormalization-group (RG) limit cycles: rather than flowing to a fixed point under scale transformations, coupling parameters or observables cycle periodically in the logarithm of the scale, and continuous scale invariance is broken to its discrete subgroup (Ovdat et al., 2019, Schröder et al., 2016). This mechanism underpins DSI in many quantum and statistical systems.

2. Quantum and Condensed Matter Realizations

The prototypical physical realization is Efimov physics: three-body bound state spectra in systems with resonant two-body interactions form a geometric sequence $E_n = E_* e^{-2\pi n/s_0}$, with $s_0$ set by the dimensionless coupling. This geometric tower is a direct manifestation of DSI (Ovdat et al., 2019, Hammer et al., 2008). Analogous structures appear in other overcritical quantum systems, such as:

• The Dirac-Coulomb problem in 2D and 3D Dirac/Weyl materials (e.g., graphene, ZrTe$_5$, HfTe$_5$). Above a critical coupling $Z \alpha > |\kappa|$, a geometric sequence of quasi-bound (atomic collapse) states forms, with energy scaling $E_{n+1}/E_n = \lambda$ where $$\lambda = \exp\left[-\pi/\sqrt{(Z\alpha)^2 - \kappa^2}\right]$$ (Shao et al., 2022, Wang et al., 2018, Liu et al., 2018).
• In topological Dirac semimetals, discrete scale invariance emerges in magnetotransport as log-periodic oscillations in $\ln B$, traced to resonant impurity scattering between the lowest Landau level and DSI quasi-bound states (Liu et al., 2020, Wang et al., 2018, Liu et al., 2018).
• Atomic vacancies in HfTe$_5$ and ZrTe$_5$ act as artificial nuclei: spatial imaging (STM/STS) directly observes DSI through ring-like quasi-bound "orbitals" at radii $R_n\sim1/E_n$ scaling geometrically, and spectral peaks in the density of states fall on a log-linear ladder (Shao et al., 2022). Magnetic fields introduce a competing length, breaking DSI beyond a field-dependent cutoff.

DSI scaling ratios, $\lambda$, can be tuned experimentally by controlling material thickness, carrier density, or screening, which modify the effective couplings and thus the RG structure (Liu et al., 2020).

3. Theoretical Foundations: RG Limit Cycles and Scale Anomalies

In theoretical analysis, DSI emerges at quantum phase transitions where continuous scale symmetry is lost due to scale anomalies. Model Hamiltonians, such as the Schrödinger $1/r^2$ potential or massless Dirac Coulomb systems, support continuous scale invariance for subcritical coupling but pass through a phase transition at critical coupling, where the scaling exponent becomes complex and DSI appears (Ovdat et al., 2019, Hammer et al., 2008). The RG flow of boundary parameters, rather than approaching a fixed point, enters a limit cycle: the RG beta function $\beta(g)$ has complex roots, and observables acquire log-periodic dependence.

Berezinskii-Kosterlitz-Thouless (BKT)–type scaling behavior arises near this criticality, with energy gaps vanishing as $\Delta\sim\exp(-\pi/\sqrt{\lambda-\lambda_c})$. Observables, including the local density of states and susceptibilities, acquire log-periodic corrections. This scenario is seen in both many-body contexts (Efimov trimers) and in two-dimensional Dirac materials with supercritical impurities (Ovdat et al., 2019).

4. DSI in Statistical Mechanics, Dynamics, and Random Structures

DSI extends to stochastic and classical settings:

• In supercritical percolation, order parameters after the transition grow in staircase patterns with discrete jumps. The locations and heights of jumps obey exact geometric scaling, characterized by parameters $\mu$ and $\lambda$, encoding DSI (see Table 1) (Schröder et al., 2016).

Model	Fractional Jump $\mu$	Scaling Ratio $\lambda$
Global homophilic percolation	2	$\approx$1.79
Local homophilic percolation	1.5	$\approx$1.32
Modified ER (analytic $\beta$=1)	2	2

• In processes and fields, DSI appears as invariance of joint distributions under dilations by integer powers of the scale, with explicit box-counting and spectral techniques developed for estimation, covariance, and Hurst parameter extraction (Modarresi et al., 2017, Modarresi et al., 2010, Rezakhah et al., 2016, Moharil et al., 18 Mar 2025).
• For random graph models with infinite-mean fitness, the outlier spectrum of adjacency matrices forms a geometric log-periodic sequence in both magnitude and eigenvector scaling, manifesting DSI at the spectral level (Catanzaro et al., 15 Sep 2025).
• In rough surfaces and Loewner-evolved fractal curves, DSI is encoded in Weierstrass-Mandelbrot constructions or the use of DSI driving functions, with scaling exponents and log-periodic features observed in contour statistics (Nezhadhaghighi et al., 2010, Nezhadhaghighi et al., 2010).
• In dynamical systems, DSI can be reframed via geometric inversion symmetries; the scaling relations for Lyapunov exponents and fractal dimensions can be deduced efficiently from inversion laws (Santos et al., 2 Feb 2026).

5. Experimental and Computational Realizations

DSI is observable in several experimental settings:

• Quantum: STM/STS imaging of Dirac semimetals and graphene with vacancies directly visualizes DSI ladders in the local density of states (Shao et al., 2022). Magnetotransport experiments detect log-periodic oscillations in MR and Hall effect, with scaling ratios matching theoretical predictions (Liu et al., 2020, Wang et al., 2018, Liu et al., 2018).
• Cold atoms: Efimov states in ultracold gases allow direct measurement of geometric scaling in three-body loss resonances, with the scaling factor in agreement with RG analysis (Ovdat et al., 2019).
• Trapped ions: Tunable DSI is realized in chains of trapped ions, where discrete scaling factors in bound state spectra and time-fractal return amplitudes are engineered and observed by adjusting inter-ion interactions (Lee et al., 2019).
• Deep Neural Networks: Coarse-geometric analysis of deep weight spaces reveals fractal dimensions and DSI under recursive dilation, with implications for network architecture and complexity (Moharil et al., 18 Mar 2025).
• Classical and financial systems: Piecewise DSI and Hurst exponent estimation techniques have been applied to real financial market time series, leveraging DSI structure to improve parameter extraction and modeling (Modarresi et al., 2017, Rezakhah et al., 2016).

6. Holography, Cosmology, and Extensions

DSI has significant implications in holography and cosmological models:

• Holography: Toy models attempting to engineer DSI or scale-without-conformal-invariance in the AdS/CFT context have revealed that periodic warping in the bulk does not generically break the full conformal group—hidden isometries can restore continuous scale invariance unless extra ingredients disrupt the global symmetric structure (Flory, 2017). Future constructions require explicit breaking of the relevant Killing symmetries or more intricate matter content.
• Cosmology: Discrete Scale Relativity proposes DSI as a universal organizing symmetry of natural hierarchies, linking atomic and stellar scales through fixed transformation factors for fundamental physical parameters (Oldershaw, 2009). This framework predicts, and finds empirically, self-similar scaling in stellar and atomic periods and structures across scales.

7. Outlook and Open Directions

The detailed understanding of DSI has led to advanced tools for extracting and quantifying discrete scaling phenomena in a wide variety of systems. Challenges remain in engineering robust DSI in holographic and top-down models, in clarifying the impact of long-range interactions and screening, and in exploring DSI in non-equilibrium, driven, or many-body quantum systems. The universality of DSI as a signal of RG limit cycles, scale anomalies, and complex exponents ensures its continuing centrality in mathematical physics, quantum materials, and the theory of fractals (Ovdat et al., 2019, Hammer et al., 2008, Nezhadhaghighi et al., 2010, Moharil et al., 18 Mar 2025, Shao et al., 2022).