Second Harmonic Injection (SHI)

• Second Harmonic Injection (SHI) is a method that deliberately adds a second harmonic (2ω) component to manipulate system responses across different physical platforms.
• In oscillator Ising machines and accelerator RF systems, SHI facilitates phase binarization and bucket shaping, leading to improved computational mapping and beam dynamics.
• In nonlinear optics, SHI underpins valley current injection and enhanced two-photon transitions, with its effect finely tuned by phase, resonance, and symmetry-breaking mechanisms.

Second Harmonic Injection (SHI) denotes a family of context-dependent techniques in which a drive, forcing term, current, or engineered response at twice a fundamental frequency is deliberately introduced or exploited. In contemporary literature, the term is used most explicitly for oscillator Ising machines, where a second-harmonic forcing term binarizes oscillator phases, and for accelerator RF systems, where a cavity operating at twice the fundamental RF frequency reshapes the longitudinal bucket. Closely related usages appear in nonlinear optics, where one- and two-photon interference produces valley current injection and second-harmonic signals, and in second-order spectroscopy of superconductors under supercurrent injection (Farasat et al., 26 Aug 2025, Madrak et al., 2018, Gong et al., 2019, Huang et al., 2023).

1. Terminology and scope

The meaning of SHI depends on the physical system. In oscillator networks, it is a dynamical term proportional to $\sin(2\phi_i)$; in synchrotrons, it is an RF voltage at exactly twice the fundamental RF frequency; in nonlinear optics, it is associated with current injection or second-harmonic generation mediated by one- and two-photon pathways. This domain dependence is not merely terminological: the injected second harmonic may act as a restoring force, a bucket-shaping voltage, or a route to nonlinear current generation (Farasat et al., 26 Aug 2025, Madrak et al., 2018, Gong et al., 2019).

Context	Meaning of SHI	Immediate purpose
Oscillator Ising machines	Second-harmonic forcing term $-K_s\sin(2\phi_i)$	Phase binarization and spin mapping
Synchrotron RF systems	RF voltage at twice the fundamental frequency	RF bucket shaping and beam-loss reduction
Nonlinear optics in monolayer TMDs	Valley current injection by one-/two-photon interference, connected to SHI	Valley current and second-harmonic response

A recurrent source of confusion is that closely related fields often use neighboring but nonidentical language: SHG for second-harmonic generation, SHI for second harmonic injection, and, in ultrasound, SHI for second harmonic imaging. The underlying commonality is the intentional use of a component at frequency $2\omega$, but the physical object being injected is system-specific.

2. Oscillator Ising machines

In oscillator Ising machines (OIMs), SHI is introduced as the mechanism that converts continuous oscillator phases into effective Ising spins. The target optimization problem is encoded in the Ising Hamiltonian

$H = -\sum_{i,j} J_{ij} s_i s_j,$

with $s_i \in \{+1,-1\}$. Under coupling and SHI, the oscillator phases evolve as

$$\frac{d\phi_i}{dt} = -K \sum_{j \neq i} J_{ij}\sin(\phi_i - \phi_j) - K_s \sin(2\phi_i),$$

and the associated energy function is

$$E(\phi) = -K\sum_{i,j,\,i \neq j} J_{ij} \cos(\phi_i - \phi_j) - K_s \sum_i \cos(2\phi_i).$$

The spin mapping is

$s_i = \mathrm{sgn}(\cos(\phi_i)),$

so $\phi_i \approx 0$ encodes $+1$ and $-K_s\sin(2\phi_i)$0 encodes $-K_s\sin(2\phi_i)$1. The SHI term acts as a restoring force that pulls phases toward $-K_s\sin(2\phi_i)$2 or $-K_s\sin(2\phi_i)$3, and this phase binarization is essential because only at $-K_s\sin(2\phi_i)$4 does the OIM energy match the Ising spin energies (Farasat et al., 26 Aug 2025).

The same paper shows that SHI can also induce spin freezing, a previously unreported phenomenon in which an oscillator cannot switch spin states even when the switch would lower the Ising energy. Near the neutral points $-K_s\sin(2\phi_i)$5, write $-K_s\sin(2\phi_i)$6. The dynamics become

$-K_s\sin(2\phi_i)$7

and freezing occurs when

$-K_s\sin(2\phi_i)$8

or equivalently

$-K_s\sin(2\phi_i)$9

If this condition holds, the SHI force dominates and the oscillator cannot cross the neutral point. The result is reduced degrees of freedom, stagnation in suboptimal solutions, and, when many spins are frozen, effective cessation of computation before steady state is reached (Farasat et al., 26 Aug 2025).

A central empirical finding is that the onset and prevalence of spin freezing are highly sensitive to initialization. Contrary to conventional practice, which favors random initialization, the paper reports that initializing all oscillators at $2\omega$0 or $2\omega$1 delays the onset of spin freezing and consistently yields better optimization results on Erdős–Rényi random graphs with 100–200 nodes. The design recommendation is therefore not simply to maximize $2\omega$2, but to choose it just strong enough to enable binarization without suppressing necessary spin flips; the paper also proposes heterogeneous SHI and SHI annealing as alternative approaches (Farasat et al., 26 Aug 2025).

3. Longitudinal beam dynamics and accelerator RF

In accelerator physics, SHI refers to adding an RF voltage at exactly twice the fundamental RF frequency. The Fermilab Booster 2nd harmonic RF cavity is designed to operate from approximately 76 to 106 MHz, twice the Booster’s fundamental range of approximately 38 to 53 MHz, and is turned on only during injection and transition or extraction. Its main purpose is to reduce beam loss as required by the Proton Improvement Plan (Madrak et al., 2018).

The beam-dynamics role of the second harmonic is bucket shaping. If the total RF voltage is written as

$2\omega$3

then choosing $2\omega$4 at injection makes the second harmonic $2\omega$5 out of phase with the fundamental. The resulting waveform is flatter at the top, the longitudinal focusing is less nonlinear, and the bucket area is increased. The reported consequences are higher capture efficiency, reduced space-charge effects through lower peak line charge density, improved transition matching, and voltage linearization for bunch rotation at extraction (Madrak et al., 2018).

The realized cavity embodies this function in hardware. It is a quarter-wave structure, 844 mm flange-to-flange, with a 76 mm aperture and a design goal of 100 kV peak gap voltage. Frequency tuning is achieved with aluminum-doped garnet rings under perpendicular magnetic bias, using a solenoid inside a magnetic flux return yoke. The shunt impedance varies from 96 k$2\omega$6 at 76 MHz to 180 k$2\omega$7 at 106 MHz, and the power system uses an Eimac Y567B tetrode with demonstrated output power greater than 100 kW at both low and high ends of the frequency range (Madrak et al., 2018).

This usage makes clear that SHI is not inherently an optical concept. Here the injected second harmonic is a deliberately phased RF component used to reshape the accelerator’s longitudinal potential well.

4. Nonlinear optical current injection and valley-selective second-harmonic processes

In monolayer transition-metal dichalcogenides, two-photon transitions in the electron-hole continuum underpin both SHG and valley current injection. The paper identifies two contributions to the two-photon transition: an interband process via remote bands, with a valley selection rule opposite to that of one-photon absorption and nearly independent of photon energy, and an interband-intraband process within the massive Dirac cone, with the same selection rule as the one-photon process and a strength that grows rapidly with photon energy. With strong electron-hole Coulomb interaction, these channels excite orthogonal angular-momentum states: two-photon $2\omega$8 excitation creates p-states with $2\omega$9, while two-photon $H = -\sum_{i,j} J_{ij} s_i s_j,$0 creates s-states with $H = -\sum_{i,j} J_{ij} s_i s_j,$1 (Gong et al., 2019).

The same framework yields cross-circular SHG. For threefold-rotationally-invariant lattices such as monolayer TMDs, two $H = -\sum_{i,j} J_{ij} s_i s_j,$2 photons generate a $H = -\sum_{i,j} J_{ij} s_i s_j,$3 signal and two $H = -\sum_{i,j} J_{ij} s_i s_j,$4 photons generate a $H = -\sum_{i,j} J_{ij} s_i s_j,$5 signal. Strong Coulomb interaction substantially enhances the two-photon transition strength in the p-channel and greatly alters the relative strength of different cross-circular polarized SHG processes. Valley current injection by quantum interference of one-photon and two-photon transition is also substantially enhanced by the Coulomb interaction, and the paper explicitly connects this current-injection picture to SHI (Gong et al., 2019).

A more formal classification appears in the theory of quantum geometry induced second harmonic generation. There the second-order current is written as

$H = -\sum_{i,j} J_{ij} s_i s_j,$6

and the injection contribution is

$H = -\sum_{i,j} J_{ij} s_i s_j,$7

The paper organizes the response into Drude, injection, shift, anomalous, double resonant, and higher-order pole contributions, linking them to the quantum geometric tensor and quantum geometric connection. The symmetry constraints are sharp: the SH injection current vanishes in time-reversal-symmetric systems, survives in parity-time-symmetric systems, and all SH currents vanish in parity-symmetric systems (Bhalla et al., 2021).

Taken together, these results define an optical meaning of SHI in which the “injection” is not an externally applied field at $H = -\sum_{i,j} J_{ij} s_i s_j,$8 but a nonlinear carrier or valley current generated through interference, selection rules, and band geometry.

5. Supercurrent injection, injection locking, and resonant enhancement

A distinct route to second-order response is provided by superconductors under dc supercurrent injection. In clean superconductors, optical transitions between particle-hole pair bands across the superconducting gap become allowed because supercurrent injection breaks inversion symmetry. With vertex correction included to maintain $H = -\sum_{i,j} J_{ij} s_i s_j,$9 gauge symmetry, the theory predicts two pronounced current-dependent peaks in the second-harmonic conductivity $s_i \in \{+1,-1\}$0 at $s_i \in \{+1,-1\}$1 and $s_i \in \{+1,-1\}$2, and one peak in the photocurrent effect $s_i \in \{+1,-1\}$3 at $s_i \in \{+1,-1\}$4; all diverge in the clean limit. In the single-band $s_i \in \{+1,-1\}$5-wave model, the current-induced peak in $s_i \in \{+1,-1\}$6 is proportional to the square of the supercurrent density, with the same order of magnitude as the experimental observation of second-harmonic generation in NbN (Huang et al., 2023).

Integrated photonics reaches the second harmonic through a different control mechanism: self-injection locking of a semiconductor laser to a high-$s_i \in \{+1,-1\}$7 silicon nitride microresonator. In this system, the injection-locking mechanism combined with a high-$s_i \in \{+1,-1\}$8 microresonator produces an ultra-narrow intrinsic linewidth at the fundamental harmonic frequency as small as 57 Hz. Quasi-phase-matched second-order nonlinearity is photoinduced through the coherent photogalvanic effect, efficient SH generation is achieved across the whole C and L telecom bands, and the device outputs SH power exceeding 2 mW with efficiency as high as 280%/W under electrical driving. The backscattered SH light does not participate in self-injection locking; only the fundamental harmonic is involved in the feedback that locks the laser (Clementi et al., 2023).

Metamaterial implementations illustrate yet another engineering pathway. In a doubly resonant metamaterial, a unit cell consists of two coupled resonators, one resonant at the fundamental frequency and the other around the SH frequency. When the resonant condition is satisfied for both frequencies, SH generation is enhanced, and the measured SH generation is 4.6 times as large as in a singly resonant metamaterial (Kanazawa et al., 2011).

These examples do not use the term SHI in an identical way, but they show three recurring control variables for second-harmonic response: symmetry breaking by supercurrent injection, coherence transfer by self-injection locking, and resonant enhancement by deliberate two-frequency design.

6. Misconceptions, distinctions, and engineering implications

A first misconception is that SHI is interchangeable with SHG. The literature does not support that equivalence. In monolayer TMDs, SHG and valley current injection are analyzed together, but the current-injection channel arises from quantum interference of one-photon and two-photon transitions and is treated as conceptually distinct from the emitted second-harmonic field (Gong et al., 2019). In OIMs, by contrast, SHI is a dynamical forcing term rather than a generated optical signal (Farasat et al., 26 Aug 2025).

A second misconception is that stronger second-harmonic forcing is always beneficial. In OIMs, too much SHI causes early and widespread spin freezing, degrading the analog search for lower-energy configurations (Farasat et al., 26 Aug 2025). In doubly resonant metamaterials, enhancement depends on coupling and resonance conditions, and for very large or very small resonator separation the enhancement drops or the weak-coupling approximation breaks down (Kanazawa et al., 2011). These results suggest that second-harmonic control is typically a tuning problem rather than a monotonic one.

A third distinction is terminological. In ultrasound literature, SHI can denote second harmonic imaging rather than injection. In that setting, tissue harmonic imaging is achieved by pulse inversion, and the main problem is low signal-to-noise ratio; the cited paper studies Eigenspace-based minimum variance beamforming and parameter choices $s_i \in \{+1,-1\}$9, $$\frac{d\phi_i}{dt} = -K \sum_{j \neq i} J_{ij}\sin(\phi_i - \phi_j) - K_s \sin(2\phi_i),$$0, and $$\frac{d\phi_i}{dt} = -K \sum_{j \neq i} J_{ij}\sin(\phi_i - \phi_j) - K_s \sin(2\phi_i),$$1 for second harmonic ultrasound imaging (Heidari et al., 2018). This is separate from second harmonic injection in oscillator, accelerator, and optical-response contexts.

Across the cited research, the practical lesson is consistent even though the platforms differ: the second harmonic is most useful when it is engineered to act at the correct strength, symmetry, phase, and resonance. In OIMs this means choosing $$\frac{d\phi_i}{dt} = -K \sum_{j \neq i} J_{ij}\sin(\phi_i - \phi_j) - K_s \sin(2\phi_i),$$2 to enforce binarization without freezing spins; in the Booster it means phasing the second harmonic to flatten the bucket; in nonlinear materials it means exploiting Coulomb interaction, quantum geometry, supercurrent-induced inversion-symmetry breaking, or resonant microcavity enhancement to control the second-order response (Farasat et al., 26 Aug 2025, Madrak et al., 2018, Bhalla et al., 2021, Huang et al., 2023)