Quantized Hall Array Resistance Standards

• Quantized Hall Array Resistance Standards (QHARS) are resistance systems that interconnect multiple quantum Hall effect elements to synthesize traceable resistance values based on the relation R_H = h/(νe²).
• They employ various interconnection schemes—such as series, parallel, bridge, star–mesh, and programmable topologies—to minimize parasitic effects and achieve sub-ppb precision in metrological applications.
• Graphene-based QHARS expand the operational range and enhance robustness, enabling practical calibration of resistance values from ~12.9 kΩ to the GΩ regime using wafer-scale uniformity and advanced network synthesis.

Quantized Hall Array Resistance Standards (QHARS) are resistance standards that synthesize practical values from the quantum Hall resistance by interconnecting multiple quantum Hall effect elements. Their metrological basis is the exact relation $R_H = h/(\nu e^2) = R_K/\nu$, with most implementations centered on the $\nu=2$ plateau because it combines the best plateau robustness with the best demonstrated precision in both GaAs/AlGaAs and graphene systems. In contemporary graphene-based realizations, QHARS encompass direct series and parallel arrays, mixed topologies, bridge networks, ambipolar p–n junction arrays, programmable edge-channel circuits, and star–mesh-derived recursive networks, extending accessible values from the single-device $R_K/2$ scale to the G$\Omega$ regime while preserving traceability to $R_K=h/e^2$ (Rigosi, 2022, Mhatre et al., 2022, Jarrett et al., 2023, Scaletta et al., 31 Jul 2025).

1. Metrological definition and quantization framework

The defining constant for QHARS is the von Klitzing constant,

$R_K = \frac{h}{e^2},$

and the quantized Hall resistance on an integer plateau is

$R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$

For the $\nu=2$ plateau used most widely in resistance metrology, the elemental value is $R_H = R_K/2 \approx 12{,}906.4037~\Omega$. In an array of identical elements operated on the same plateau, simple scaling follows standard network rules: $N$ elements in series give $\nu=2$0, $\nu=2$1 identical elements in parallel give $\nu=2$2, and more generally $\nu=2$3 (Rigosi, 2022, Ribeiro-Palau et al., 2015).

QHARS are needed because a single quantum Hall device normally provides only one canonical value on its most robust plateau, whereas practical metrology requires resistance values near $\nu=2$4, $\nu=2$5, $\nu=2$6, $\nu=2$7, and $\nu=2$8. The general synthesis problem is therefore to approximate a target ratio $\nu=2$9 by a rational network value $R_K/2$0, while keeping the number of Hall elements small and retaining insensitivity to contacts and wiring. In the network-synthesis literature this is treated using Stern–Brocot and continued-fraction approximations, followed by minimal planar-network constructions based on rectangle tilings (1311.0756).

A central point in graphene-based QHARS is that higher Landau-level plateaus such as $R_K/2$1 and $R_K/2$2 have been studied but “simply do not offer the same level of precision as $R_K/2$3.” The practical consequence is that most array strategies increase or decrease resistance by topological interconnection of many $R_K/2$4 elements rather than by relying on higher-$R_K/2$5 plateaus of a single device (Rigosi, 2022).

2. Interconnection schemes and suppression of parasitic errors

The precision of a QHARS is limited primarily by contact resistance $R_K/2$6, finite longitudinal resistance $R_K/2$7, and residual wiring, leakage, and thermoelectric terms. A qualitative metrological error model used for arrays writes the relative deviation as

$R_K/2$8

with the coefficients determined by connection topology. QHARS architectures are therefore designed so that $R_K/2$9 is strongly suppressed by multiple interconnections, $\Omega$0 is minimized by operating where $\Omega$1, and $\Omega$2 is reduced through guarded wiring, current reversal, and CCC-based measurement practice (Rigosi, 2022).

The canonical topology is the Delahaye-style multiple-series or multiple-parallel connection. In these schemes, voltage terminals are cross-connected so that parasitic lead and contact resistances appear as common-mode contributions and are rejected to higher order. The multiple-bridge extension generalizes the same idea to bridge networks. In the notation of the bridge-connection analysis, the canonical series and parallel results are

$\Omega$3

and for the five-element bridge example the effective resistance becomes

$\Omega$4

This bridge formalism enlarges the accessible set of rational multiples $\Omega$5 without giving up the parasitic-rejection properties of multiple-series and multiple-parallel arrays (Callegaro et al., 2013).

Graphene adds a distinct interconnection mechanism through ambipolar p–n junctions. Because p-type and n-type regions can be defined in one continuous graphene sheet, series addition can be realized without crossed metallic interconnects. In the Landauer–Büttiker description, the Hall resistance across adjacent regions at filling factors $\Omega$6 and $\Omega$7 is

$\Omega$8

For a bipolar junction with $\Omega$9, the terms add and yield $R_K=h/e^2$0, while the longitudinal resistance across the interface vanishes when edge-channel equilibration is complete and $R_K=h/e^2$1 is negligible (Woszczyna et al., 2011).

A conceptually different architecture is the programmable quantum Hall bisector. Rather than wiring many Hall bars in fixed series or parallel combinations, it uses gate-defined mixers and floating equilibration contacts to generate any binary fraction of $R_K=h/e^2$2,

$R_K=h/e^2$3

Its notable metrological claim is structural: it contains no internal current-carrying ohmic contact, so the dominant parasitic mechanism of conventional hard-wired arrays is avoided at the circuit level (Momtaz et al., 2020).

3. Materials platforms and device engineering

Historically, quantum Hall metrology began with GaAs/AlGaAs devices, but graphene broadened the operational envelope that makes arrays practical. A large-area graphene quantum Hall device grown on SiC was shown to sustain quantization within $R_K=h/e^2$4 from $R_K=h/e^2$5 to $R_K=h/e^2$6, up to $R_K=h/e^2$7 at $R_K=h/e^2$8 and $R_K=h/e^2$9, and up to $R_K = \frac{h}{e^2},$0 at $R_K = \frac{h}{e^2},$1 and $R_K = \frac{h}{e^2},$2, with a recommended simultaneous operating point of $R_K = \frac{h}{e^2},$3, $R_K = \frac{h}{e^2},$4, and $R_K = \frac{h}{e^2},$5. The same work reported agreement between graphene and GaAs/AlGaAs quantized Hall resistance with a relative combined standard uncertainty of $R_K = \frac{h}{e^2},$6 (Ribeiro-Palau et al., 2015).

These margins are tied to graphene-specific electronic structure. The Landau levels follow

$R_K = \frac{h}{e^2},$7

with a first-level spacing $R_K = \frac{h}{e^2},$8, compared with $R_K = \frac{h}{e^2},$9. Combined with the robust $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$0 plateau in epitaxial graphene on SiC, this yields wider plateau windows, lower onset fields, higher current margins, and relaxed cryogenic requirements relative to traditional GaAs standards (Ribeiro-Palau et al., 2015).

Large-scale graphene QHARS depend on wafer-level material uniformity and low-loss interconnects. Epitaxial graphene on semi-insulating SiC, using either Si-face or C-face growth, supports wafer-scale arrays; superconducting NbN or NbTiN pads and interconnects, often in split-contact geometries inspired by Delahaye, keep contact and lead resistances effectively negligible under quantum Hall conditions. Functionalization with $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$1 or polymer-assisted doping is used to homogenize carrier density across many elements so that all Hall bars enter the $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$2 plateau together (Rigosi, 2022).

A direct fabrication demonstration of this approach used PASG-grown epitaxial graphene on $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$3 Si-face 4H-SiC, $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$4 NbTiN superconducting interconnects, and gate-less density control by $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$5 functionalization followed by vacuum annealing. The selected chips showed $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$6 monolayer coverage and uniform Raman 2D-mode characteristics over widely separated array elements. Large contact pads were placed at the end of every row of 10 elements, enabling multiple direct outputs from $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$7 to $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$8, that is, from approximately $R_H = \frac{R_K}{\nu} = \frac{h}{\nu e^2}.$9 to $\nu=2$0 (Mhatre et al., 2022).

At smaller scale, the four-bar “mini array” used epitaxial graphene grown on the Si-face of 4H-SiC by annealing in argon at approximately $\nu=2$1 for 5 minutes, patterned into four $\nu=2$2 Hall bars on a $\nu=2$3 chip with Ti/Au double metal–graphene contacts and polymer-assisted photochemical-gating capability. Its plateau onset between $\nu=2$4 and $\nu=2$5 already illustrated why epitaxial graphene is attractive for small QHARS modules (Novikov et al., 2015).

4. Direct-array implementations and demonstrated performance

Direct arrays provided the first metrological validation of QHARS in graphene. They established that multiple Hall bars could be made to quantize simultaneously on the same plateau and that multiple-series or multiple-parallel connections could preserve the expected quantum value within bridge-limited or artifact-limited uncertainty (Novikov et al., 2015, Rigosi, 2022).

Configuration	Conditions and nominal value	Reported result
4 series graphene bars	$\nu=2$6, $\nu=2$7, $\nu=2$8; $\nu=2$9	$R_H = R_K/2 \approx 12{,}906.4037~\Omega$0; $R_H = R_K/2 \approx 12{,}906.4037~\Omega$1 vs GaAs-traceable route
10 series array	$R_H = R_K/2 \approx 12{,}906.4037~\Omega$2, $R_H = R_K/2 \approx 12{,}906.4037~\Omega$3; $R_H = R_K/2 \approx 12{,}906.4037~\Omega$4	CCC uncertainty $R_H = R_K/2 \approx 12{,}906.4037~\Omega$5
13 parallel array	$R_H = R_K/2 \approx 12{,}906.4037~\Omega$6, $R_H = R_K/2 \approx 12{,}906.4037~\Omega$7; $R_H = R_K/2 \approx 12{,}906.4037~\Omega$8–$R_H = R_K/2 \approx 12{,}906.4037~\Omega$9	Deviations $N$0 and $N$1 n$N$2/$N$3
Two $N$4 (118 parallel) in series	$N$5; $N$6	No significant deviation within $N$7 n$N$8/$N$9; precision maintained up to $\nu=2$00

The four-device series array is especially important because it demonstrates the complete direct-QHARS measurement chain. The devices were linked in a triple-series configuration with external aluminum bonding wires of about $\nu=2$01 each. The low-field anomalies associated with bilayer inclusions and inhomogeneities disappeared on the quantum Hall plateau, and a differential proxy $\nu=2$02 approached zero above $\nu=2$03, consistent with $\nu=2$04. A conservative operating current of $\nu=2$05 was chosen because one bar pair showed breakdown onset above approximately $\nu=2$06, illustrating the standard array constraint that the element with the lowest breakdown current sets the usable series current (Novikov et al., 2015).

Larger direct arrays extended both range and precision. A $\nu=2$07-element series array delivered approximately $\nu=2$08 with CCC uncertainty near $\nu=2$09, while parallelization to 13 elements lowered the nominal value to about $\nu=2$10 and simultaneously raised the usable current into the $\nu=2$11 regime. The mixed topology consisting of two subarrays of 118 elements in parallel and then connected in series realized $\nu=2$12, with direct array-to-array comparisons showing no significant deviation within $\nu=2$13 n$\nu=2$14/$\nu=2$15 (Rigosi, 2022).

The multi-output $\nu=2$16-element graphene array shifted emphasis from single nominal standards to simultaneous dissemination. Its available taps included $\nu=2$17, $\nu=2$18, a diagnostic $\nu=2$19, and $\nu=2$20. The work emphasized architectural feasibility and simultaneous access rather than a full primary-standard uncertainty budget, but it showed that direct multi-output QHARS can replace separate single-value standards on one chip (Mhatre et al., 2022).

5. Network synthesis beyond simple series and parallel

QHARS design has progressively moved from direct arrays toward network synthesis. One branch of this development is purely combinatorial: find a rational approximation $\nu=2$21 to the desired ratio $\nu=2$22, then realize that rational value with a minimal planar network. Using Stern–Brocot and continued-fraction methods together with square-tiling constructions, compact decadic designs were derived for $\nu=2$23, $\nu=2$24, $\nu=2$25, $\nu=2$26, and $\nu=2$27. Representative outcomes include a 12-element network for $\nu=2$28 using $\nu=2$29, an 18-element network for $\nu=2$30 using $\nu=2$31, and an 88-element network for $\nu=2$32 using $\nu=2$33, all with relative errors of order $\nu=2$34 in the stated examples (1311.0756).

A second branch uses star–mesh transformations to generate much larger effective resistances from modest numbers of $\nu=2$35 elements. In the NIST $\nu=2$36 device, two $\nu=2$37-element series arms and a single-element ground arm form a Y network with

$\nu=2$38

The equivalent $\nu=2$39 resistance between the Hi and Lo terminals is

$\nu=2$40

This proof-of-concept established that star–mesh transformations can lift QHARS into the $\nu=2$41 range while using far fewer elements than a simple series chain. The same study gave explicit examples for higher values, including $\nu=2$42-$\nu=2$43-$\nu=2$44 elements yielding $\nu=2$45 and $\nu=2$46-$\nu=2$47-$\nu=2$48 yielding $\nu=2$49 (Jarrett et al., 2023).

Related work on artifact networks, intended as proof-of-concept for future graphene-based QHARS, showed how a single Y–$\nu=2$50 step can produce approximately $\nu=2$51, and how a recursively optimized 37-element network can reach approximately $\nu=2$52. These demonstrations were used to argue that $\nu=2$53 quantum Hall elements can suffice for $\nu=2$54, and only 37 for $\nu=2$55, provided the star–mesh topology is exploited instead of naive series scaling (Jarrett et al., 2024).

The recursive design program has since been formalized further. In a measured near-$\nu=2$56 graphene QHARS, the parameter choice $\nu=2$57, $\nu=2$58, $\nu=2$59 produced an approximate design value of $\nu=2$60, but exact correction for virtual branches created by star–mesh transforms moved the effective resistance to $\nu=2$61, in agreement with LTspice and DSB measurements at approximately $\nu=2$62. The same work introduced Minkowski–Bouligand dimension as a descriptor of design topology and reported a Pearson correlation coefficient of $\nu=2$63 between MBD and the exponential coefficients extracted from device-count scaling. In that framework, partial recursions yield linear growth in $\nu=2$64, whereas full and hybrid recursions yield exponential growth (Scaletta et al., 31 Jul 2025).

A still more aggressive architecture is the cross-square recursion. With $\nu=2$65 epitaxial-graphene Hall elements, this topology was reported to produce $\nu=2$66 for $\nu=2$67, $\nu=2$68 for $\nu=2$69, and a projected $\nu=2$70 for $\nu=2$71. The $\nu=2$72 value was measured with a teraohmmeter, whereas the $\nu=2$73 attempt was limited by leakage in a conventional wet cryogenic system rather than by an identified failure of the quantized network itself (Tran et al., 5 Aug 2025).

6. Applications, misconceptions, limitations, and future directions

The metrological motivation for QHARS is practical dissemination. Series arrays of 4–8 $\nu=2$74 devices naturally generate the $\nu=2$75–$\nu=2$76 range used in quadrature bridges for capacitance metrology, direct dc calibration can be done without resistive scaling, and parallelized low-value arrays enable higher currents for calibrating $\nu=2$77 and $\nu=2$78 standards. Multi-output chips additionally allow simultaneous access to many quantum-traceable values in one cryostat run, from the single-device $\nu=2$79 scale to $\nu=2$80 on one chip (Novikov et al., 2015, Rigosi, 2022, Mhatre et al., 2022).

Several common misconceptions are corrected by the published record. One is that QHARS are synonymous with simple hard-wired series strings; in fact, the literature includes multiple-series, multiple-parallel, multiple-bridge, p–n junction, programmable binary-fraction, star–mesh, and cross-square topologies (Momtaz et al., 2020). Another is that higher graphene plateaus such as $\nu=2$81 or $\nu=2$82 can replace arrays; the explicit metrological conclusion is the opposite, namely that these plateaus do not attain the same precision and robustness as $\nu=2$83 (Rigosi, 2022). A third is that very high effective resistance automatically implies metrological readiness. The teraohm-scale work shows that leakage through cryostat wiring, feedthroughs, guarding, and adsorbate films can dominate before the device physics itself fails (Tran et al., 5 Aug 2025).

The major technical constraints are recurrent across architectures. Arrays require simultaneous plateau overlap across all elements, so carrier-density uniformity and mobility homogeneity remain critical. Breakdown current is an array-wide bottleneck in series topologies because the lowest-breakdown element sets the source current. In large arrays, parasitic capacitances, pad integrity under thermal cycling, and cryostat wiring burden become nontrivial. In direct-array measurements, the uncertainty floor can be set by the bridge rather than by the quantum Hall devices themselves, as in the four-bar graphene mini-array where the reported $\nu=2$84 was limited by the DCC bridge specification (Novikov et al., 2015, Rigosi, 2022).

This suggests that current QHARS research is less a question of the universality of the quantum value than of architecture, fabrication yield, and measurement infrastructure. The experimental comparison of graphene and GaAs at $\nu=2$85, the n$\nu=2$86/$\nu=2$87-level array comparisons near $\nu=2$88, and the bridge-limited results of smaller direct arrays all point in that direction. The main future directions already identified in the literature are robust on-chip interconnects, tighter uniformity control, higher breakdown-current margins, local or programmable matching of plateau conditions, AC characterization for impedance metrology, automated switching, and integration with Josephson programmable standards for broader quantum-electrical architectures (Ribeiro-Palau et al., 2015, Rigosi, 2022, Momtaz et al., 2020).