Mass-Induced Coulomb Drag

• Mass-induced Coulomb drag is a phenomenon where a mass parameter, rather than direct Coulomb coupling, governs the drag response via altered screening, correlations, or dispersion mismatch.
• The paper shows that effective-mass asymmetry in semiconductor bilayers enhances drag, with hole-hole drag observed to be several times larger than electron-hole drag due to lower Fermi temperatures and reduced screening.
• Other systems, including massless–massive bilayers, magnetic topological insulators, and superconducting nanowires, demonstrate diverse drag behavior where mass terms (band masses, Dirac gaps, or plasmon mass gaps) control momentum transfer.

Searching arXiv for recent and foundational work on mass-induced Coulomb drag and closely related drag mechanisms. Mass-induced Coulomb drag denotes a family of drag phenomena in which the drag response is controlled by a mass parameter rather than by interlayer Coulomb coupling alone. In the literature, that mass parameter appears in several distinct forms: the effective-mass asymmetry of electrons and holes in semiconductor bilayers, the mismatch between massless and massive fermions in hybrid double layers, a Dirac mass gap in magnetic topological materials, or a collective-mode gap that acts as a mass term in coupled superconducting nanowires. Across these settings, the common observable remains the same—a voltage or current induced in an electrically isolated passive conductor by driving the active conductor—but the physical route by which “mass” enters the drag problem varies qualitatively (Zheng et al., 2015, Liu et al., 2016, Latyshev et al., 26 Aug 2025).

1. General framework and the meaning of “mass-induced”

In the canonical drag geometry, two nearby conductors are separated by an insulating spacer so that direct tunneling is absent, while long-range Coulomb interaction still couples density fluctuations across the gap. The measurable quantity is the transresistivity or drag resistivity, conventionally written as

$$R_D=-\frac{V_2}{I_1}, \qquad \rho_D\equiv -\rho_{xx}^{(12)}.$$

In linear response, when $\sigma_D\ll \sigma_i$,

$$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$

A simple phenomenological treatment already exposes one route by which mass can appear explicitly: $$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$ where $\tau_D$ is the interlayer momentum-relaxation time. Microscopically, the drag rate is controlled by dynamically screened interlayer interactions and the nonlinear susceptibilities of the two subsystems, so mass dependence may also enter indirectly through $E_F$, $k_F$, screening, density of states, and electron-hole asymmetry (Narozhny et al., 2015).

This immediately implies that “mass-induced” is not a single universal mechanism. In some limits, mass is explicit in transport coefficients; in others, it disappears from the leading asymptote and survives only through screening, correlations, or subleading corrections. A useful organizing principle is therefore to distinguish between effective-mass enhancement, mass-mismatch drag, mass-gap-enabled Hall drag, and gap-lifted collective drag.

Setting	Mass ingredient	Characteristic drag consequence
Ambipolar GaAs/AlGaAs bilayer	$m_h^*\approx 5m_e^*$	hole drag strongly enhanced
Graphene–2DEG or BLG–MLG	massless–massive dispersion mismatch	distinct small-$d$ density scaling
Magnetic TI films	Dirac mass gap $M$	passive-layer anomalous Hall drag
Superconducting nanowires	plasmon mass gap in passive wire	finite drag after cancellation is lifted
Moiré graphene heterostructures	effectively “massive” miniband states near sDP	finite drag with massless Dirac carriers

A complementary point is that mass dependence is often absent in the simplest high-density, weak-coupling limit. That absence is itself diagnostically important: when strong mass sensitivity is observed, it usually indicates departure from the simplest Thomas–Fermi or weak-screening picture.

2. Effective-mass asymmetry in semiconductor bilayers

A particularly clean realization of mass-induced drag enhancement is provided by the ambipolar GaAs/AlGaAs double quantum well studied in “Switching between attractive and repulsive Coulomb-interaction-mediated drag in an ambipolar GaAs/AlGaAs bilayer device” (Zheng et al., 2015). The structure contains two $\sigma_D\ll \sigma_i$0 nm GaAs quantum wells separated by a $\sigma_D\ll \sigma_i$1 nm $\sigma_D\ll \sigma_i$2 barrier, with mean interlayer separation $\sigma_D\ll \sigma_i$3 nm. Front and back gates allow the same device to be configured either as an electron-hole bilayer or as a hole-hole bilayer. This enables a direct comparison between attractive interlayer interactions in the electron-hole case and repulsive interlayer interactions in the hole-hole case.

The drag resistivity is measured from the induced voltage in the passive layer,

$\sigma_D\ll \sigma_i$4

and in linear response it is governed by the standard double integral over momentum and frequency involving $\sigma_D\ll \sigma_i$5, $\sigma_D\ll \sigma_i$6, and $\sigma_D\ll \sigma_i$7. The experiment is performed at low densities, around $\sigma_D\ll \sigma_i$8 to $\sigma_D\ll \sigma_i$9, in a strongly interacting regime where

$$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$0

The resulting hierarchy is

$$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$1

with hole-hole drag about $$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$2–$$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$3 times larger than electron-hole drag, and electron-hole drag about $$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$4 times larger than previously reported electron-electron drag in a similar device.

The central explanation is the large effective-mass asymmetry,

$$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$5

Because $$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$6, the hole layer has a much smaller Fermi temperature. At the lowest density studied,

$$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$7

At the experimental temperatures the holes are therefore much closer to thermal excitation, which weakens screening and increases drag. The larger hole mass also lowers kinetic energy and strengthens intralayer hole-hole correlations, which further reduce screening. The decisive observation is that hole-hole drag remains larger than electron-hole drag even though the former has repulsive interlayer interactions and the latter attractive ones. The mass-driven enhancement of correlations and the associated poor screening dominate over the naive attraction-versus-repulsion expectation.

The density dependence reinforces this interpretation. At $$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$8 K, the paper reports

$$\rho_D=\frac{\sigma_D}{\sigma_1\sigma_2-\sigma_D^2}\approx \frac{\sigma_D}{\sigma_1\sigma_2}.$$9

and

$$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$0

These exponents show that the drag is much more sensitive to the hole density than to the electron density. By contrast, in the high-density Thomas–Fermi limit,

$$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$1

which is independent of particle mass. The measured mass dependence is therefore a signature of being far from that simple limit.

3. Massless–massive fermion bilayers and moiré minibands

A second meaning of mass-induced drag arises when the two layers have different dispersions. “Coulomb drag between massless and massive fermions” analyzes a low-temperature ballistic double layer composed of an $$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$2-doped graphene sheet and a 2DEG, and also a bilayer-graphene–monolayer-graphene system (Scharf et al., 2012). The graphene layer is massless,

$$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$3

whereas the 2DEG or bilayer-graphene layer is massive,

$$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$4

In this regime the drag remains quadratic in temperature,

$$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$5

but the density dependence depends sharply on separation. At small interlayer separation the leading density law distinguishes the three canonical cases: $$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$6 for the special matching $$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$7. At large separation, however, the leading asymptotic behavior becomes universal,

$$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$8

for massive–massive, massless–massless, and massless–massive bilayers alike. In that regime, the memory of mass mismatch survives only in the subleading correction. Mass-induced drag in this sense is therefore most visible at short and intermediate separation, not in the leading large-$$\rho_D=-\rho_{xx}^{(12)}=\frac{m_2}{e^2 n_1\tau_D},$$9 asymptote.

A moiré-based experimental analogue appears in graphene/hBN/graphene heterostructures, where one graphene layer aligned to hBN develops moiré minibands and satellite Dirac points, while the other layer remains intrinsic graphene (Wang et al., 2024). The paper describes carriers near the satellite Dirac point as effectively “massive” or band-engineered relative to the massless fermions near the original Dirac point. Finite drag is observed not only in original-Dirac-point to original-Dirac-point coupling, but also when moiré miniband carriers are coupled to massless Dirac carriers. At high temperature and large density the drag follows

$\tau_D$0

together with

$\tau_D$1

and satisfies layer reciprocity. At low temperature, reciprocity is broken in both original-Dirac-point and satellite-Dirac-point coupled regions, suggesting dominant energy drag. The drag near the satellite Dirac points is smaller than near the original Dirac point and deviates from $\tau_D$2 below $\tau_D$3 K. This suggests that coupling between moiré miniband carriers and original Dirac carriers is not of a simple Fermi-liquid nature.

4. Dirac mass gaps, Berry curvature, and anomalous Hall drag

In magnetic topological insulators, mass-induced drag takes a different form. “Anomalous Hall Coulomb drag of massive Dirac fermions” considers a double-layer topological-insulator film in which both active and passive surfaces are magnetized out of plane (Liu et al., 2016). For each layer,

$\tau_D$4

The mass term $\tau_D$5 opens a gap, makes the Dirac fermions massive, and generates Berry curvature and an anomalous Hall effect.

The central result is negative in a precise sense: the anomalous Hall current in the active layer does not generate drag in the passive layer. The transverse drag channels associated with the active-layer anomalous Hall effect vanish identically, and the active-layer topological Hall response does not directly feed into Coulomb drag. Instead, the mechanism proceeds in two stages. First, the active layer is driven longitudinally, and interlayer scattering transfers a longitudinal drag force to the passive layer. Second, because the passive layer is itself a massive Dirac system, that dragged longitudinal drive is converted by the passive layer’s own anomalous Hall response into a transverse drag current.

This leads to two sharp predictions. The anomalous Hall drag current is independent of the active-layer magnetization $\tau_D$6, because the active-layer Berry-curvature contribution does not couple into the drag channel. By contrast, it depends non-monotonically on the passive-layer magnetization $\tau_D$7: it grows approximately linearly for $\tau_D$8, peaks at intermediate magnetization, and decreases when $\tau_D$9 as the system becomes more 2DEG-like and the anomalous Hall effect fades. The peak is more pronounced at low density. If the active layer is undoped and supports only a quantized anomalous Hall current with no longitudinal current, there is no drag at all.

The broader review “Coulomb drag in topological materials” places this result in a wider TI context and emphasizes that for $E_F$0 the longitudinal drag is essentially unchanged by the mass gap, while the Hall drag channel is the one that carries the distinctive mass dependence (Liu et al., 2017). Mass-induced drag here therefore does not mean a larger longitudinal $E_F$1 by simple effective-mass scaling; it means that a Dirac mass enables a passive-layer Hall response to a drag-induced longitudinal force.

5. Collective-mode mass and drag in superconducting nanowires

A still different usage appears in “Mass-induced Coulomb drag in capacitively coupled superconducting nanowires” (Latyshev et al., 26 Aug 2025). The system consists of two parallel superconducting nanowires coupled capacitively through a mutual capacitance $E_F$2. A current bias $E_F$3 is applied to wire 1, and drag is defined as the stationary voltage induced in wire 2. The relevant low-energy excitations are plasmons, and quantum phase slips in the active wire generate voltage fluctuations that can be transmitted to the passive wire.

When both wires are superconducting, the induced voltage in the passive wire vanishes: $E_F$4 The reason is an exact cancellation of plasmon contributions: a quantum phase slip in wire 1 launches two voltage pulses in wire 2 with opposite sign, and their time-averaged contribution cancels. In this regime there is no drag despite capacitive coupling.

The situation changes when wire 2 is tuned below the superconductor-insulator transition and develops a mass gap,

$E_F$5

That mass term gaps the passive-wire plasmon spectrum, lifts the cancellation, and produces a finite drag coefficient

$E_F$6

For short wires, $E_F$7,

$E_F$8

so drag is weak. For long wires, $E_F$9,

$k_F$0

which is the maximal drag set by the mutual capacitance.

The semiclassical picture is especially transparent. For $k_F$1, the coupled system supports two plasmon modes with different velocities, and the pulses induced in the passive wire have opposite sign. For $k_F$2, the mass term reduces the velocity splitting, synchronizes the plasmon modes, and prevents complete cancellation. Mass-induced drag in this context is therefore a collective-mode effect rather than a single-particle band-structure effect.

6. Conceptual boundaries, related regimes, and common misidentifications

Not every large or unusual drag signal is mass-induced in the same sense. A useful boundary case is exciton condensation in bilayer quantum Hall systems at $k_F$3. In “Exciton Condensation and Perfect Coulomb Drag,” strong interlayer correlations produce a condensate of interlayer excitons and nearly perfect counterflow drag, with the ideal relation

$k_F$4

and measured drag ratios approaching $k_F$5 in the strongest quantum Hall regime (Nandi et al., 2012). This suggests that “perfect drag” is best regarded as a coherence-driven excitonic transport phenomenon rather than a consequence of effective-mass asymmetry or a band mass gap.

The opposite boundary case is double-layer graphene with massless Dirac fermions. Theory for graphene monolayers finds no explicit mass-induced drag mechanism; instead, the characteristic features are chirality, suppressed backscattering, and asymptotic laws

$k_F$6

with little or no $k_F$7 correction because $k_F$8 backscattering is suppressed by chirality (Letelier, 2011). Massless-fermion drag is therefore an instructive contrast: the drag can be substantial, but its controlling variables are not effective mass or a mass gap.

A broader extension appears in “Drag viscosity of metals and its connection to Coulomb drag,” where the interaction-induced part of the stress tensor gives a shear-viscosity contribution closely related to Coulomb drag (Liao et al., 2019). There the relevant object is not a bilayer transresistivity but an intrafluid momentum-transfer channel, sometimes described as a Coulomb-drag-like effect inside a single electronic fluid. This extension preserves the underlying logic of interaction-induced momentum transfer, but it also shows that “mass-induced Coulomb drag” is best treated as a family resemblance among mechanisms rather than a single sharply delimited phenomenon.

Taken together, the literature supports a precise but plural usage. In semiconductor bilayers, mass acts primarily through Fermi temperature, screening, and correlation strength; in massless–massive bilayers, through dispersion mismatch and altered density scaling; in magnetic topological systems, through Berry-curvature-enabled Hall response of the passive layer; and in superconducting nanowires, through a gap that prevents exact cancellation of collective excitations. The unifying theme is that a mass parameter changes how interlayer or nonlocal Coulomb-mediated momentum transfer is converted into measurable drag, while the specific microscopic mechanism depends strongly on the host system and on whether the relevant “mass” is a band mass, a Dirac gap, or a collective-mode gap.