Copper thiophosphate ferroelectrics

CuInP₂S₆; relatives CuInP₂Se₆, CuCrP₂S₆, AgCrP₂S₆

Also called CuInP₂S₆, CIPS, copper indium thiophosphate, CuInP₂Se₆, CuCrP₂S₆

van der Waals crystal insulator

The layered crystal that made room-temperature ferroelectricity a van der Waals property. Below about 315 K the copper ions in CuInP2S6 settle off-centre and give each layer an electric polarisation pointing out of the plane, which an electric field can flip – so a flake a few nanometres thick can store a bit, steepen a transistor’s switching or act as a memristor. The same copper ions can also wander through the crystal, making it both a ferroelectric and an ionic conductor: a combination devices can exploit, and one that complicates every measurement.

Key properties

  • Ferrielectric below a Curie temperature of ~315 K: the copper and indium sublattices displace in opposite directions by different amounts, so their dipoles only partly cancel
  • An order–disorder transition: above the Curie temperature copper hops between sites, below it the ions settle off-centre
  • Switchable out-of-plane polarisation at room temperature in flakes down to ~4 nm
  • Giant negative longitudinal piezoelectricity – the opposite sign to almost every other piezoelectric, the best-known other example being the polymer PVDF
  • A quadruple potential well: copper can also sit in the van der Waals gap, giving a second, larger polarisation state
  • Mobile Cu+ ions: a biased probe tip can pull copper out of the lattice and let it return reversibly
  • CuCrP2S6, the magnetic relative, orders antiferromagnetically below ≈ 30 K with ferromagnetic coupling inside each layer; a field of 6–8 T suppresses that order, and broad anomalies between 140 and 200 K mark the antiferroelectric ordering of its copper, which at room temperature already sits off-centre, disordered between positions

How it is made

  • Bulk crystals by chemical vapour transport or solid-state reaction of the elements in sealed quartz ampoules
  • CuCrP2S6 and AgCrP2S6 as large crystals by optimised chemical vapour transport; CuCrP2S6 exfoliates down to four layers (2.6 nm)
  • Mechanical exfoliation to flakes a few nanometres thick; thin-flake ferroelectricity is usually probed by piezoresponse force microscopy
  • Van der Waals stacking with graphene, MoS2 or silicon to build diodes, transistors and memristors

Uses, and how close they are

  • Non-volatile ferroelectric memory: ferroelectric diodes and transistorslab
  • Negative-capacitance transistors switching below the 60 mV per decade limitlab
  • Self-rectifying memristors for neuromorphic computinglab
  • Bulk photovoltaic photodetectors and flexible piezoelectric deviceslab

Readiness runs lab → prototype → pilot → deployed.

Open problems

  1. How thin can CuInP2S6 be and stay ferroelectric? Early probe measurements lost polarisation below ~50 nm, later devices switched at ~4 nm – how much do electrodes, screening and mobile copper decide?
  2. In a given switching loop, how much of the signal is ferroelectric polarisation and how much is copper-ion motion?
  3. Can the high-polarisation state, with copper in the van der Waals gap, be stabilised and used at room temperature?
  4. Can the transition be raised far enough above room temperature for memory that must survive chip operating temperatures?

Going deeper

Short notes for specialists. Choose a lens in the header and yours comes first.

For theoreticians · your lens

A clean example of a second-order Jahn–Teller ferroelectric: d10 Cu+ gains energy by moving off-centre in its sulfur cage, and the layered lattice lets that instability order cooperatively. Along the stacking axis the copper potential has four minima – two within the layer and two reaching into the van der Waals gap – so interlayer coupling through copper in the gap matters as much as the single-layer picture. The large displacive instability of copper in a layered lattice also explains its negative piezoelectricity. CuCrP2S6 adds magnetism: calculations predict ferromagnetic layers with an antipolar copper arrangement that an electric field can drive into a polar state.

For experimentalists · your lens

A piezoresponse hysteresis loop alone does not prove ferroelectricity here: electrostatic tip effects and copper migration give similar loops. Check that written domains persist and can be imaged without bias, compare with second-harmonic generation or macroscopic polarisation loops, and keep tip voltages and dwell times low, because a biased tip can extract copper and damage the surface. Always report the temperature: with the Curie temperature at ~315 K, a warm laboratory is already close enough to the transition to reduce the polarisation.

For engineers · your lens

One of the few ferroelectrics that stacks onto 2D semiconductors without dangling bonds. It has already made lab-scale negative-capacitance MoS2 transistors with a minimum subthreshold swing of 28 mV per decade, as well as ferroelectric diodes and memristors. The obstacles are a Curie temperature only ~20 K above room temperature, copper migration that threatens retention and endurance, and the absence of wafer-scale growth. Hafnium-oxide ferroelectrics, already compatible with silicon manufacturing, set the bar it has to clear.

In the research tracks

Recent news

The newest items tagged CuInP2S6, CuCrP2S6, from the news feed updated 5 Oct 2026.

All 16 items tagged CuInP₂S₆, CuCrP₂S₆ in the news feed  ·  RSS feed for CuInP₂S₆, CuCrP₂S₆

Key references

  1. Paraelectric–ferroelectric transition in the lamellar thiophosphate CuInP2S6Simon et al. · Chemistry of Materials 6, 1575 (1994)cited by 159doi:10.1021/cm00045a016
  2. Ferrielectric ordering in lamellar CuInP2S6Maisonneuve et al. · Physical Review B 56, 10860 (1997)cited by 307doi:10.1103/PhysRevB.56.10860
  3. CuInP2S6 room temperature layered ferroelectricBelianinov et al. · Nano Letters 15, 3808 (2015)cited by 536doi:10.1021/acs.nanolett.5b00491
  4. Room-temperature ferroelectricity in CuInP2S6 ultrathin flakesLiu et al. · Nature Communications 7, 12357 (2016)cited by 1,256doi:10.1038/ncomms12357
  5. Two-dimensional ferromagnetism and driven ferroelectricity in van der Waals CuCrP2S6Lai et al. · Nanoscale 11, 5163 (2019)cited by 189doi:10.1039/C9NR00738E
  6. Crystal growth, exfoliation, and magnetic properties of quaternary quasi-two-dimensional CuCrP2S6Selter et al. · Physical Review Materials 7, 033402 (2023)cited by 35doi:10.1103/PhysRevMaterials.7.033402
  7. Crystal growth of the quasi-2D quarternary compound AgCrP2S6 by chemical vapor transportSelter et al. · Crystals 11, 500 (2021)cited by 18doi:10.3390/cryst11050500
  8. Origin of giant negative piezoelectricity in a layered van der Waals ferroelectricYou et al. · Science Advances 5, eaav3780 (2019)cited by 302doi:10.1126/sciadv.aav3780
  9. Van der Waals negative capacitance transistorsWang et al. · Nature Communications 10, 3037 (2019)cited by 271doi:10.1038/s41467-019-10738-4
  10. Tunable quadruple-well ferroelectric van der Waals crystalsBrehm et al. · Nature Materials 19, 43 (2019)cited by 302doi:10.1038/s41563-019-0532-z
  11. Enhanced bulk photovoltaic effect in two-dimensional ferroelectric CuInP2S6Li et al. · Nature Communications 12, 5896 (2021)cited by 301doi:10.1038/s41467-021-26200-3
  12. Manipulation of current rectification in van der Waals ferroionic CuInP2S6Jiang et al. · Nature Communications 13, 574 (2022)cited by 168doi:10.1038/s41467-022-28235-6