# Fileset

[s41467-025-65431-6.pdf](https://mdr.nims.go.jp/filesets/0451fbd3-ce90-4f43-9764-6fb829ecc689/download)

## Creator

Sviatoslav Kovalchuk, Kyrylo Greben, Abhijeet M. Kumar, Simon Pessel, Jan Soyka, Qing Cao, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Dominik Christiansen, Malte Selig, Andreas Knorr, Siegfried Eigler, Kirill I. Bolotin

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Revealing hidden interlayer excitons in 2D bilayers via hybrid molecular gating](https://mdr.nims.go.jp/datasets/d449a404-4d73-4c35-85f7-a862600f46d4)

## Fulltext

Revealing hidden interlayer excitons in 2D bilayers via hybrid molecular gatingArticle https://doi.org/10.1038/s41467-025-65431-6Revealing hidden interlayer excitons in 2Dbilayers via hybrid molecular gatingSviatoslav Kovalchuk 1 , KyryloGreben 1, AbhijeetM. Kumar1, SimonPessel1,Jan Soyka2, Qing Cao2, Kenji Watanabe 3, Takashi Taniguchi 3,Dominik Christiansen4, Malte Selig4, Andreas Knorr 4, Siegfried Eigler 2 &Kirill I. Bolotin 1Heterostructures of molecules and two-dimensional materials feature emer-gent properties not seen in their individual components. Here, we studyexcitons in bilayer transition metal dichalcogenides exposed to an intenseelectric field produced by charge transfer from proximal molecules. Ourapproach allows for reaching an electric field strength of 0.35 V nm−1, up to afactor of two higher than previously achieved in purely solid-state gateddevices. Under this field, inter- and intralayer excitons are brought into anenergetic resonance, allowing us to explore a newphysical regime.Wedetect apreviously unseen interlayer exciton that only becomes visible at high electricfield throughhybridizationwith the intralayer A exciton.Moreover, the systemexperiences an ultra-strong Stark splitting of > 350meVwith exciton energiestunable over a large range of the optical spectrum, holding potential foroptoelectronics. Our work paves the way for using strong electric fields tostudy new physical phenomena and control exciton hybridization in 2Dsemiconductors.Interlayer excitons (IX) in bilayer transitionmetal dichalcogenides (2L-TMDs) are Coulomb-boundpairs of electrons and holeswith an out-of-plane dipole moment. Compared to intralayer excitons, the electron-hole separation in IXs is larger and the oscillator strength lower1–4. As aresult, IXs feature lifetimes in the tens of nanoseconds4–6 and diffusionlengths up to microns7,8, much higher than their intralayer counter-parts. These properties led to an explosion of interest in IXs in fieldssuch as excitonic transport8,9, Bose-Einstein condensation10–12, exci-tonic insulators13,14, and quantum simulation15. In homobilayers, thedistinguishing property of IXs is their coupling to intralayer excitonsresulting from interlayer hole tunneling3,16–19. This mechanism leads toa tunable enhancement of the oscillator strength of IXs in some 2L-TMDs, e.g., 2L-MoS2, allowing their observation via optical absorptionspectroscopy16,17,20. Finally, IXs exhibit a Stark splitting in electric fieldsoriented perpendicularly to the plane of the material due to their outof plane static dipolemoment4,16,17,20–22. As a result, the energy position,oscillator strength and coupling strength to other excitonic speciescan be tuned by an electric field.A perpendicular electric field in 2L-TMDs is conventionallyapplied in a dual-gated field effect transistor geometry. Electrostaticgates consisting of a dielectric (e.g., hBNor SiO2) and a conductor (e.g.,gold, Si or graphene) are assembled on both sides of the 2L-TMD. Insuch a configuration, the difference between gate voltages applied tothe top and bottom conductors controls the field across the material,while the sum of gate voltages controls the carrier density and Fermienergy23. Generally, the strength of the perpendicular electric fieldcontrols the energy splitting between IXs with oppositely orienteddipole moments (denoted IX+ and IX-). The maximum reported split-ting in conventional dual-gated devices17,24–26 is in practice limited bythe breakdown of the dielectric material. At the point of dielectricbreakdown of hBN, the electric field inside a 2L-TMD reaches ≈ 0.2 Vnm−1. Assuming an interlayer exciton dipolemoment of 0.6 e ⋅ nm, thisReceived: 9 June 2025Accepted: 14 October 2025Check for updates1Physics Department, Freie Universität Berlin, Berlin, Germany. 2Institute of Chemistry and Biochemistry, Freie Universität Berlin, Berlin, Germany. 3NationalInstitute for Materials Science, Tsukuba, Japan. 4Physics Department, Technische Universität Berlin, Berlin, Germany. e-mail: kovalchook@gmail.com;bolotin@zedat.fu-berlin.deNature Communications |         (2025) 16:9893 11234567890():,;1234567890():,;http://orcid.org/0000-0002-4817-1939http://orcid.org/0000-0002-4817-1939http://orcid.org/0000-0002-4817-1939http://orcid.org/0000-0002-4817-1939http://orcid.org/0000-0002-4817-1939http://orcid.org/0000-0003-2852-7384http://orcid.org/0000-0003-2852-7384http://orcid.org/0000-0003-2852-7384http://orcid.org/0000-0003-2852-7384http://orcid.org/0000-0003-2852-7384http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0009-0001-1712-4590http://orcid.org/0009-0001-1712-4590http://orcid.org/0009-0001-1712-4590http://orcid.org/0009-0001-1712-4590http://orcid.org/0009-0001-1712-4590http://orcid.org/0000-0002-0536-8256http://orcid.org/0000-0002-0536-8256http://orcid.org/0000-0002-0536-8256http://orcid.org/0000-0002-0536-8256http://orcid.org/0000-0002-0536-8256http://orcid.org/0000-0003-1821-3429http://orcid.org/0000-0003-1821-3429http://orcid.org/0000-0003-1821-3429http://orcid.org/0000-0003-1821-3429http://orcid.org/0000-0003-1821-3429http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65431-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65431-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65431-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-65431-6&domain=pdfmailto:kovalchook@gmail.commailto:bolotin@zedat.fu-berlin.dewww.nature.com/naturecommunicationselectric field corresponds to a Stark shift of 120 meV, smaller than theseparation between A and B excitons for most 2L-TMDs, e.g., 240meVinWSe2. As a result, it is challenging to explore the fascinating regimesof hybridization of IXs with both of these intralayer excitons.While an order of magnitude higher electric field has beenrecently generated using ionic liquids23,27, that approach is so far lim-ited to room temperature and incompatible with opticalmeasurements.Here, to study the regime of tunable coupling between IXs andother excitonic species, we overcome the limits of solid-state gatingtechnologies. We develop a hybrid molecular gating approach thatallows the generationof anelectricfieldof >0.35Vnm−1 (displacementfield > 2.2 V nm−1), nearly doubling the previous limit. A Stark splittingof > 350 meV allows us to discover a new high energy interlayerexciton in bilayer MoS2, labeled IX2, by hybridyzing with XA. In bilayerMoSe2, meanwhile, the high electric field reveals a new dark interlayerexciton state.Device concept/Evaporation TechniqueTo overcome the limits of conventional gating, we add layers ofcharges next to the 2L-TMD (Fig. 1a). The top layer consists of acceptormolecules with charge density σt. We use either 2,3,5,6-Tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4TCNQ), a well-known commer-cially available molecular acceptor with electron affinity of 5.3 eV28–34or hexacyano-trimethylene-cyclopropane (CN6-CP), a tailor-mademolecular acceptor with high electron affinity of 5.94 eV35–37 (detailsof synthesis and characterization are given in SI). The bottom layer ofcharges originates from the donor states already present at the inter-faces between theTMDand the SiO2/Si stack. These states,with chargedensity σb, arise due to a combination of photodoping38–40 and inter-face charge trapping41,42.A large electric field inside the 2L-TMD is generated due to a highdensity of charges in the top layer compensated by the combinationofbottom layer and electrostatic gate (voltage VG). The subtle but criticalaspect of this molecular gating approach is that the charges on bothsides of the device are localized rather than free. These charges cannotresult in persistent currents, which is one of themainmechanisms thatleads to dielectric breakdown. Using a three-capacitor electrostaticmodel (Fig. 1a), we determine the electric field in the TMD layer as (SInote S1):FZ � 12ε0εTMDðσt � σb � VGCGÞ, ð1Þwhere CG is the areal capacitance between the bilayer and the Si gate.The formula shows that the electric field inside the 2L-TMD can exceedthe field inside the SiO2, the last term inside the parenthesis. We notethat the gate voltage VG controls the chemical potential alignmentbetween the 2L-TMD and the molecules and hence determines thechargedensity σt in them (Fig. 1a).WhenVG is gradually increased, boththe electric field and the Fermi energy increase with it, up to the pointwhen the Fermi energy reaches the minimum of the conduction bandεTMDεbεtσbσtVGCGMoleculesEvap. chamberObjec�veVEVAPTMDVGb)a)c)ELUMOe-ECBMEFEVBM2L-TMD Moleculese-Si/SiO2d)VGCr, Au Molecules2L-TMDhBNSiO2 trapsSiO2Si++VG (V)Electricfield(Vnm-1)F4TCNQCN6-CP0.20.30.40.10.0Solid-state-100 -50 0 50 100V G0(V)� tmax(1012cm-2)Evap. cycles50-50080431 20 4 5F4TCNQPL(arb. units)Energy (eV)10XTXA1.90 1.951.85e)Fig. 1 | Hybrid molecular gating. a Device schematic of a 2L-TMD with two layersof charge, and an equivalent 3-capacitor circuit. The top inset illustrates the ener-getic alignment between SiO2/Si gate, 2L-MoS2 and F4TCNQ. b Evaporationchamber setup schematic. Organic molecules are deposited onto a TMD in situ bycontrollably heating the coil on a separate chip. c Calculated electric field depen-dence on the gate voltage for the device described in the text based on themolecular dopant (blue and red) versus solid-state only device (black).d Photoluminescence spectra of 2L-TMD before (blue) and after (orange) eva-poration of F4TCNQ, at fixed VG = -20 V. The decrease of the trion peak XT intensityindicates that 2L-MoS2 becomesmoreneutral due to charge transfer.eThe smallestvoltage at which the trion feature is visible in the optical spectra (V0G) vs. thenumber of the evaporation cycle. The maximum charge transfer density in themolecules (σmaxt ) is shown on the right axis. Molecular density at points shown asblue and orange symbols correspond to the curves of the same color in (d).Article https://doi.org/10.1038/s41467-025-65431-6Nature Communications |         (2025) 16:9893 2www.nature.com/naturecommunicationsof one of the layers. In this situation, the screening due to free carriersinduced in the 2L-TMD limits additional field increase.To illustrate the utility of the molecular gating approach, wemodeled the electric field vs. VG inside 2L-TMD in Fig. 1c (see SI note 1formodeling details).We consider three cases: devices based on eitherF4TCNQ or CN6-CP molecules as well as a double-gated device basedon top and bottom dielectric-based gates. The maximum electric fieldin the latter device is limited to 0.2 V/nm by the dielectric breakdownof the dielectric. In contrast, for the caseof theCN6-CPdevice, thefieldstrength exceeds that value almost by a factor of 2.We control the maximum carrier density in the molecular layerσmaxt - and with it the maximum FZ - during the experiment using anewly developed in situ molecular evaporation technique. Theapproach works by applying short pulses of current to a micro-fabricated coil on a separate chip loaded with molecules. That chip isplacedclose to themeasured sample inside the cryostat (Fig. 1b, detailsin Methods). The temperatures of both chips are monitored bymicrofabricated thermometers. Even when the temperature of theevaporator chip reaches 400 K, the sample remains at near Heliumtemperature (Si Fig. 10). This approach addsmore flexibility comparedto traditional deposition techniques33,43–45, enables the evaporation ofoxidizing molecules, and allows precise control of the surfacecoverage, which could be even more crucial when using other organicmolecules, for example dyes46.We confirm molecular deposition by recording photo-luminescence (PL) and reflectivity spectra during cycles of molecularevaporation (Fig. 1d, e). In PL measurement, we observe spectralchanges in the region of the intralayer neutral exciton (XA at 1.94 eV)and intralayer trion (XT at 1.91 eV). The gradual decrease of XTbrightness is consistent with the change of the carrier density in thesample due to the deposition of acceptor molecules onto the 2L-MoS233. The quantitative analysis of an exemplary F4TCNQ sampleindicates that σt can be gradually increased from 0 to 9 × 1012 cm−2 inthat device via molecular deposition (Fig. 1e).ResultsStark splitting in bilayer TMD systems in linear approximationWe now study the effect of the electric field on interlayer excitons in aCN6-CP/2L-MoS2 device (the data for a F4TCNQ/MoS2 sample are inthe SI). Figure 2a shows the map of the second derivative of thereflectivity contrast as a function of VG. We identify the spectral fea-tures corresponding to intralayer XA and XB excitons (1.93 and 2.10 eV,respectively, at VG = − 80 V) as well as interlayer exciton IX1 ( ~ 2 eV atVG = − 80 V) that undergoes Stark splitting in a non-zero electric field.1.9279Electricfield (Vnm- 1)Osc. strength (arb. units)XA- IX2-0.300.260.340.0 1.0e)Electricfield (Vnm-1)CN6-CP2L-MoS2x10-0.4 0.42ω2ΔR/R (arb. units)0.150.050.250.35V G(V)04080-80-40Energy (eV)2.0 2.11.9 2.2a) b)top           botK KTIX1IX2FZdBLtop bottop botIX1- IX1+FZ > 0FZMAXEFEF0c) d)Electricfield (Vnm-1)Energy (eV)XA- XA+ IX1+XB- XB+IX1- IX2+IX2-CN6-CP2L-MoS22.31.9 2.1XA- XA+ IX1+XB-IX2-IX’1-XB+IX010.150.050.250.35Fig. 2 | Measuring excitonic response under a strong electric field. aMap of thesecond derivative of reflectivity contrast (RC =ΔR/R) for a CN6-CP/2L-MoS2 sample.The spectra above 1.95 eV are multiplied by 10 to increase contrast. b Sketchesshowing the composition of interlayer excitons for different electricfield strengths.IX1+ is shown as dark-blue, and IX1- as light-blue. The black line indicates the Fermilevel. c The dependence of the excitonic peak energies on the electric fieldextracted from the data in Fig. 2a (diamonds), along with theoretical predictionsbased on the Bloch equations (lines). Note that the coupling to a new interlayerexciton IX2 leads to the energy splitting of XA at high electric field. d Normalizedoscillator strength of the IX2− and XA- states and corresponding error bars vs. theelectric field extracted from the spectra (black dots), and the oscillator strengthpredicted from the Bloch equations model (red line). e The configurations of var-ious interlayer and intralayer excitons in 2L-TMDs at non-zero electric field. Thecoupling between the excitons sharing electronwavefunctions ismediated by spin-conserving interlayer hole tunneling (arrows). Coupled pairs of inter- and intralayerexcitons are marked with the same color.Article https://doi.org/10.1038/s41467-025-65431-6Nature Communications |         (2025) 16:9893 3www.nature.com/naturecommunicationsThe interlayer exciton hybridizes with the intralayer excitons at higherfield leading to an avoided crossing pattern. We start with a simplemodel that neglects inter-exciton hybridization, while later confirmingthe results of thatmodel by taking hybridization into account. For eachVG, we determine the Fermi energies of top and bottom TMD layersand find the corresponding electric field strengths (see details on theelectrostatic model in SI note S1). We then obtain the expected posi-tion of the Stark-split interlayer excitons EIX1 ± = E0IX1 ± FZdBL, whereE0IX1 = 1:99 eV is the known spectral position of IX1 at zero field(denoted IX01 in Fig. 2a) and dBL= 0.6 e ⋅ nm is its dipole momentcorresponding to the TMD interlayer distance1. The positions of IX1+and IX1- as a function of VG obtained in that way are shown as dashedlines in Fig. 2a. The left axis shows the electric field obtained from thesame model.At VG = − 80 V, where FZ ≈ 0, the two interlayer components IX1+and IX1- are close to degenerate. The Fermi energies of the two TMDlayers are aligned with the lowest unoccupied molecular orbital(ELUMO) of CN6-CP, while σt = 0; the electric fields from σb and the gateelectrode are compensated. In the range of VG from -80 V to -40 V, thelimit of small FZ, spectral shifts of the IX1 ± excitons are linear,matching our simple model. At higher VG, the electric field becomeslarge enough to bring inter- and intralayer excitons into an energeticresonance16,17,20,47. In general, it is known that the interlayer exciton IX1couples to the intralayer exciton XB via hole tunneling1 (arrows inFig. 2e; T denotes the tunneling strength parameter). This means thatthe state IX1 partially acquires the intralayer character of XB and hencedeviates from the linear electric field dependence1,48. Conversely, thestate XB hybridizes with IX1 acquiring an interlayer character and splitsinto two components XB+ and XB-. At VG > 0, a higher-lying Stark-splitinterlayer exciton (IX1+) crosses the position of XB and the hybridiza-tion with it decreases afterwards. The exciton again approaches a lin-ear electric field dependence (dashed line) in the regime of high VG.Finally, above VG = 90 V, the electric field does not increase further asthe Fermi energy of the bottom layer reaches the conduction bandminimum and free carriers begin to screen the field (Fig. 2b, top).A striking feature of the high electric field region is the appear-ance of two new features: an excitonic peak around 1.97 eV and thesplitting of XA into two peaks. We suggest that both features originatefromanother interlayer exciton IX2 schematically shown in Fig. 2e. Thisexciton, which can also be labeled as the “interlayer B exciton”,acquires its oscillator strength via hybridization with the XA excitonmediated by hole tunneling, in exact analogy with the well-knownbrightening of IX1 through hybridization with the XB exciton (Fig. 2e).Coupling between inter- and intralayer excitonsTo confirm the nature of the new excitonic state as well as to ascertainthe electric field magnitude, we analyze the effect of the couplingbetween inter- and intralayer excitons. We extract the energies of allexcitonic peaks (Fig. 2c, diamonds for CN6-CP/2L-MoS2, data forF4TCNQ/2L-MoS2 are in the SI Fig. 6) vs. electric field and fit them to amodel based on the Bloch equations formalism16 (solid lines, details inSI note S4). Thismodel assumes that IX1- couples to XB- only, IX1+ to XB+only, IX2− to XA- only, and IX2+ to XA+ only and neglects other morecomplex types of couplings47,49. The free parameters of the model arethe energies of inter- and intralayer excitons at zero electric field(unperturbed by interexcitonic interactions), dipole moment dBL,bottomcarrier density σb, hole tunneling strength T, assumed same forXA - IX2 and XB - IX1 couplings48 and intralayer excitons polarizability βZ(SI Fig. 8). The modeling result matches the observed positions andamplitudes of all excitonic peaks for both types of samples for all gatevoltages (SI Fig. 6, 7). We extract the tunneling strength T =37.58 ± 1 meV.From the coupling model, we find that under the maximumelectric field the new observed IX2− state is shifted by 152meV from itsposition at zero electric field, E0IX2 = 2:139±0:002 eV. This is 168 meVabove E0IX1, a value close to the spin-orbit splitting of the valenceband50. In the small electric field regime, the state is far fromXA, and itsoscillator strength, acquired by coupling to XA, is low. In the regime ofhigh electric field strength, hybridization with XA brightens IX2−,allowing its direct observation with spectral position and oscillatorstrength of the state thatmatches ourmodel (Fig. 2c, d). The derivativeof the reflectivity contrast with respect to VG51, shown in SI Fig. 13, alsoreveals features consistent with the positions of IX2+ predicted by themodel (solid lines in Fig. 2c).Dark excitons in 2L-MoSe2We now turn to another bilayer material from the TMD family, MoSe2.Using the same analysis as above, we characterize excitons in F4TCNQ/2L-MoSe2, Fig. 3a. We find a dipole moment of dBL = 0.65 ± 0.02 e ⋅ nmfor interlayer excitons and an exciton tunneling strength T = 44 ± 2meV. The parameter T is 10 meV lower compared to calculations48,50.Finally, due to a larger energy difference between XA and IX2 (371 meVin MoSe2, compared to 236 meV in MoS2) the splitting of XA is smallerand less pronounced, reaching 6 meV at the highest electric field.Additionally there is a new feature that was not identified inMoS2samples.We observe theweak avoided crossing between XA and IX1, aswell as the coupling of XA to the previously unobserved state 30 meVbelow IX1 (Fig. 3a, zoomed in region around XA shown in Fig. 3b). Wesimulated the absorption spectrum assuming that XA couples to twointerlayer states at the corresponding energies, with couplingstrengths of 5 and 10 meV, respectively. This simulation matches thedata well (Fig. 3d). We obtain additional information about this statefrom the PL spectra and estimate the dipolemoment of this new ‘dark’interlayer state to be ddark = 0.84 ± 0.14e ⋅ nm (SI Fig. 14).We suggest that the new interlayer ‘dark’ state below IX1 is asso-ciated with spin- or momentum- forbidden interlayer excitons asdepicted in Fig. 3c. It was recently shown that spin-selection rulesprohibiting scattering of such a state can be lifted under a strongelectric field due to a Rashba-like effect52 or under strong translationaldisorder introduced by the presence of molecules46. Alternatively, thestate could be related to recently observed quadrupolar excitons53.The nature of this dark state deserves further study.DiscussionThe electricfield achieved via hybridmolecular gating roughlydoublesthe limit achievable with dielectric gates. In electric fields of up to 0.35V nm−1 (Fig. 2a, e) we observe, in addition to the well-known couplingbetween the excitons IX1 and XB16,17,20: 1) a new interlayer exciton IX2hybridizing with the XA exciton at high fields, and 2) signatures ofcoupling between a dark interlayer exciton and XA.To further examine the capabilities and limitations of the mole-cular gating technique, we applied our simple model to differentcombinations of TMD bilayers, heterostructures, and molecules. ForTMDs, we choose combinations of the four most common materials,MoS2, MoSe2, WS2, and WSe250. We expose these TMDs to four dif-ferent cases of molecular doping (Fig. 4a). The first two cases are thesame as considered above, F4TCNQ or CN6-CP molecule on top of theTMDbilayer and SiO2 below. In addition,we consider a donor n-dopantBenzyl Viologen (BV0)54 at the bottom and an acceptor CN6-CP at thetop. To study the limits of our technique, we analyze the sample withCN6-CP in combination with one of the strongest organic electrondonors (OED) reported – Me-OED, which has a high doping efficiencyand a small surface area55,56.We find that the maximum achievable electric field depends bothon the type of 2D material used as well as on the chosen molecule(Fig. 4b). Interestingly, switching from F4TCNQ to a stronger acceptorCN6-CP does not necessarily increase the electric field strength. This isbecause in that case it is limited by the Fermi level entering the con-duction band of the TMD. A combination of Me-OED and CN6-CPproduces an electric field strength above 0.5 V nm−1, tripling the solid-Article https://doi.org/10.1038/s41467-025-65431-6Nature Communications |         (2025) 16:9893 4www.nature.com/naturecommunicationsa)c) d)2ω2ΔR/R (arb. units)-1.5 1.50.140.070.00.280.21Electricfield (Vnm-1)1.8 1.91.6 1.71.5XA-XA+ IX1+ XB- XB+IX1-F4TCNQ2L-MoSe2Energy (eV)b)top botK KIXdXAElectricfield (Vnm-1)Energy (eV)XA2ω2ΔR/R (arb. units)-0.5 3.0CN6-CP2L-MoSe20.080.010.220.151.681.6 1.64Electricfield (Vnm-1)Energy (eV)XAε20.0 1.0simula�on0.080.010.220.151.681.6 1.64IX1-IXdIX1-IXdFig. 3 | 2L-MoSe2 in a strong electric field. a Map of the second derivative ofreflectivity contrast for 2L-MoSe2. Gray lines show modeled positions for XA, XBand IX1- excitons. b Zoomed in region of the second derivative of reflectivity con-trast around XA exciton for sample 2. Dashed lines show predicted positions for IX1-and dark interlayer exciton IXd. c The proposed schematic of dark interlayer exci-ton. d Simulated absorption where IX1- and IXd states cross XA with couplingstrengths of 5 and 10 meV.b)a)Energy(eV)MoS2MoSe2WS2WSe2F4TCNQBV0CN6-CPMe-OED-5.0-6.0-4.0CBVBHOMOLUMOElectric fieldlimit (Vnm-1)0.20.00.40.6Fig. 4 | Estimationof themaximumelectricfieldproducedbymolecular doublegating for several TMD/molecule combinations. a Energy alignment of con-duction/valence bands for TMDs and LUMO/HOMO levels of molecules relative tothe vacuum energy. bMaximum electric field achievable in various TMD structuresdouble-gated via variousmolecular layers. For heterostructures, the stacking orderis top to bottom. Gray box corresponds to an electric field of 0.2 V nm−1, achievablevia dielectric gates. The star symbols indicates cases experimentally tested inthis work.Article https://doi.org/10.1038/s41467-025-65431-6Nature Communications |         (2025) 16:9893 5www.nature.com/naturecommunicationsstate limit (details in SI note S5). Ultimately, the magnitude of theelectric field is limited by the doping efficiencies of the consideredorganic molecules.We anticipate several potential avenues for the application of ourresults. First, the ability to control the molecular density in situ atcryogenic temperatures in nanofabricated devices may prove usefulfor the emerging field of organic-inorganic 2D heterostructures. Sec-ond, ourwork indicates thatTMDheterostructures functionalizedwithdonor/acceptor species from the top and the bottom (Fig. 4) can beconsidered novel optical materials with the characteristic absorptionpeak tunable over a large part of the visible spectrum. To enablepotential applications in such materials in, e.g., LED devices, it wouldbe interesting to investigate direct chemical synthesis routes for suchheterostructures57. Third, the response of an exciton to the out-of-plane electric field indicates its out-of-plane character and can be used,in principle, for the “fingerprinting”of excitonic species58,59. Aswe haveshown, the field response of some states (e.g., XA) is only resolvable inhigh enough electric fields. Therefore, the application of high electricfields may enable a more detailed identification of various excitonicspecies with debated character. Fourth, the state IX2 reported here isan attractive candidate to transmit information in excitoniccircuits60,61. This state is normally dark and should have a very longlifetime, enabling its propagation over long distances. The informationencoded in it could be “written” or “read out” in the regions of thecircuit exposed to a high electric field. In those areas, the state isbrought into an energetic resonance with an intralayer exciton,thereby increasing its coupling to light.MethodsDevice fabricationSamples were prepared using a PDMS dry stamping method andtransferred onto hBN directly exfoliated onto a 285 nm SiO2/Si chip62.Contacts were made using electron beam lithography (EBL) followedby thermal evaporation of Cr/Au (3 nm/70 nm). All samples werecleaned by AFM “nano-squeegee” (60 nN force) to clean the surfaceand improve the contact with molecules63,64.In situ evaporationIn our technique, we place a small amount of organic acceptorsF4TCNQ (Sigma-Aldrich, amount < 1 mg) or CN6-CP onto an eva-poration coil fabricated on a 285 nm SiO2/Si chip. The coil is madeusing EBLwith the sameparameters as the contacts on the sample. Thecoil resistance is 60Ohm, and the design is similar to Ref. 65. This chipis loaded into our optical cryostat right next to the 2L-TMD. To eva-porate a controlled dose of molecules, we apply a short voltage pulseto the evaporator coil (Fig. 1b, d, VEVAP, duration is selected between 1sand 3s), heating the molecules above their melting temperature.During the heating process, the temperature of the 2L-TMD remainsvirtually unchanged (details in SI Fig. 10). The evaporation chamber issealed inside the inner heatshield of the cryostat to avoid contamina-tion. This in situ evaporation approach has multiple advantages. First,the density of molecules can be adjusted during the experimentwithout heating the device. Second, evaporation at cryogenic tem-peratures solves the problem associated with molecules agglomerat-ing into clusters, which occurs during room-temperature deposition33.Finally, we avoid the exposure of a thinmolecular layer to the ambientenvironment.Optical measurementsWe use a home-built confocal PL/reflectivity setup at cryogenic tem-perature (4K). PLmeasurements were done using a 532 nm continuouswave laser. Reflectivity measurements were carried out using abroadband pulsed supercontinuum laser source (Super K, 400–1000nm). All measurements employed a SLWD Nikon objective (0.4 NA, 22mm WD) and an Andor spectrometer (300 and 600 lines/mmgratings). Toextract the positions of excitonic resonances fromFig. 2a,we used Kramers–Kronig relations to model the dielectric function(see SI note S3 for details).Data availabilityThe data generated in this study have been deposited in the Zenododatabase under accession code (https://doi.org/10.5281/zenodo.16951134).References1. Deilmann, T. & Thygesen, K. S. Interlayer excitons with large opticalamplitudes in layered van der Waals materials. Nano Lett. 18,2984–2989 (2018).2. Gerber, I. C. et al. Interlayer excitons in bilayer MoS2 with strongoscillator strength up to room temperature. Phys. Rev. B 99,035443 (2019).3. Paradisanos, I. et al. Controlling interlayer excitons in MoS2 layersgrown by chemical vapor deposition. Nat. Commun. 11, 2391 (2020).4. Wang, Z., Chiu, Y.-H., Honz, K., Mak, K. F. & Shan, J. Electrical tuningof interlayer exciton gases in WSe2 bilayers. Nano Lett. 18, 137–143(2018).5. Miller, B. et al. Long-lived direct and indirect interlayer excitons invan der waals heterostructures. Nano Lett. 17, 5229–5237 (2017).6. Rivera, P. et al. Observation of long-lived interlayer excitons inmonolayer mose2-wse2 heterostructures. Nat. Commun. 6, 6242(2015).7. Jauregui, L. A. et al. Electrical control of interlayer exciton dynamicsin atomically thin heterostructures. Science 366, 870–875 (2019).8. Unuchek, D. et al. Room-temperature electrical control of excitonflux in a van der waals heterostructure. Nature 560, 340–344(2018).9. Yuan, L. et al. Twist-angle-dependent interlayer exciton diffusion inWS2-WSe2 heterobilayers. Nat. Mater. 19, 617–623 (2020).10. Blatt, J. M., Böer, K. W. & Brandt, W. Bose-einstein condensation ofexcitons. Phys. Rev. 126, 1691–1692 (1962).11. Wang, Z. et al. Evidence of high-temperature exciton condensationin two-dimensional atomic double layers.Nature574, 76–80 (2019).12. Harats, M. G. & Bolotin, K. I. Limits of funneling efficiency in non-uniformly strained 2d semiconductors. 2D Mater. 8, 015010 (2021).13. Gu, J. et al. Dipolar excitonic insulator in a moiré lattice. Nat. Phys.18, 395–400 (2022).14. Xu, Y. et al. A tunable bilayer hubbard model in twisted WSe2. Nat.Nanotechnol. 17, 934–939 (2022).15. Montblanch, A. R.-P. et al. Confinement of long-lived interlayerexcitons in WS2/WSe2 heterostructures. Commun. Phys. 4,119 (2021).16. Lorchat, E. et al. Excitons in bilayer mos2 displaying a colossalelectric field splitting and tunable magnetic response. Phys. Rev.Lett. 126, 037401 (2021).17. Leisgang, N. et al. Giant stark splitting of an exciton in bilayer MoS2.Nat. Nanotechnol. 15, 901–907 (2020).18. Kumar, A. et al. Spin/valley coupled dynamics of electrons andholes at the mos2-mose2 interface. Nano Lett. 21, 7123–7130 (2021).19. Gong, Z. et al. Magnetoelectric effects and valley-controlled spinquantum gates in transition metal dichalcogenide bilayers. Nat.Commun. 4, 2053 (2013).20. Peimyoo, N. et al. Electrical tuning of optically active interlayerexcitons in bilayer mos2. Nat. Nanotechnol. 16, 888–893 (2021).21. Roch, J. G. et al. Quantum-confined stark effect in a mos2 mono-layer van der waals heterostructure. Nano Lett. 18, 1070–1074(2018).22. Zhao, Y. et al. Interlayer exciton complexes in bilayer mos2. Phys.Rev. B 105, L041411 (2022).23. Weintrub, B. I. et al. Generating intense electric fields in 2d mate-rials by dual ionic gating. Nat. Commun. 13, 6601 (2022).Article https://doi.org/10.1038/s41467-025-65431-6Nature Communications |         (2025) 16:9893 6https://doi.org/10.5281/zenodo.16951134https://doi.org/10.5281/zenodo.16951134www.nature.com/naturecommunications24. Hattori, Y., Taniguchi, T., Watanabe, K. & Nagashio, K. Anisotropicdielectric breakdown strength of single crystal hexagonal boronnitride. ACS Appl. Mater. Interfaces 8, 27877–27884 (2016).25. Murrell, M. P. et al. Spatially resolved electrical measurements ofSiO2 gate oxides using atomic force microscopy. Appl. Phys. Lett.62, 786–788 (1993).26. Illarionov, Y. Y. et al. Insulators for 2d nanoelectronics: the gap tobridge. Nat. Commun. 11, 3385 (2020).27. Domaretskiy, D. et al. Quenching the bandgap of two-dimensionalsemiconductors with a perpendicular electric field. Nat. Nano-technol. 17, 1078–1083 (2022).28. Le, O. K., Chihaia, V., Van On, V. & Son, D. N. N-type and p-typemolecular doping on monolayer mos2. RSC Adv. 11, 8033–8041(2021).29. Nourbakhsh, A., Cantoro,M., Heyns, M.M., Sels, B. F. &DeGendt, S.(invited) toward ambient-stable molecular gated graphene-fet: Adonor/acceptor hybrid architecture to achieve bandgap in bilayergraphene. ECS Trans. 53, 121–129 (2013).30. Park, J. et al. Single-gate bandgap opening of bilayer graphene bydual molecular doping. Adv. Mater. 24, 407–411 (2012).31. Tian, X., Xu, J. &Wang, X. Band gap opening of bilayer graphene byf4-tcnq molecular doping and externally applied electric field. J.Phys. Chem. B 114, 11377–11381 (2010).32. Wang, T. H., Zhu, Y. F. & Jiang, Q. Bandgap opening of bilayergraphene by dual doping from organic molecule and substrate. J.Phys. Chem. C. 117, 12873–12881 (2013).33. Wang, J. et al. Charge transferwithin the f4tcnq-mos2 vanderwaalsinterface: Toward electrical properties tuning and gas sensingapplication. Adv. Funct. Mater. 28, 1806244 (2018).34. Zhang,W. et al. Opening an electrical bandgap of bilayer graphenewith molecular doping. ACS Nano 5, 7517–7524 (2011).35. Dai, X., Liu, L., Ji, Z., Meng, Q. & Zou, Y. Surface charge transferdoping of graphene using a strongmolecular dopant cn6-cp.Chin.Chem. Lett. 34, 107239 (2023).36. Fukunaga, T., Gordon, M. D. & Krusic, P. J. Negatively substitutedtrimethylenecyclopropanes and their radical anions. J. Am. Chem.Soc. 98, 611–613 (1976).37. Karpov, Y. et al. High conductivity in molecularly p-doped diketo-pyrrolopyrrole-based polymer: The impact of a high dopantstrength and good structural order. Adv. Mater. 28, 6003–6010(2016).38. Epping, A. et al. Quantum transport through MoS2 constrictionsdefined by photodoping. J. Phys.: Condens. Matter 30, 205001(2018).39. Ju, L. et al. Photoinduced doping in heterostructures of grapheneand boron nitride. Nat. Nanotechnol. 9, 348–352 (2014).40. Roch, J. G. et al. Spin-polarized electrons in monolayer mos2. Nat.Nanotechnol. 14, 432–436 (2019).41. Guo, Y. et al. Charge trapping at the mos2-sio2 interface and itseffects on the characteristics of mos2 metal-oxide-semiconductorfield effect transistors. Appl. Phys. Lett. 106, 103109 (2015).42. Illarionov, Y. Y. et al. The role of charge trapping in mos2/sio2 andmos2/hbn field-effect transistors. 2D Mater. 3, 035004 (2016).43. Mouri, S.,Miyauchi, Y. &Matsuda, K. Tunablephotoluminescenceofmonolayer mos2 via chemical doping. Nano Lett. 13, 5944–5948(2013).44. Park, S. et al. Demonstration of the key substrate-dependent chargetransfer mechanisms between monolayer MoS2 and moleculardopants. Commun. Phys. 2, 109 (2019).45. Park, S. et al. Temperature-dependent electronic ground-statecharge transfer in van der Waals heterostructures. Adv. Mater. 33,2008677 (2021).46. Katzer, M. et al. Impact of dark excitons on förster-type resonantenergy transfer between dye molecules and atomically thin semi-conductors. Phys. Rev. B 107, 035304 (2023).47. Sponfeldner, L. et al. Capacitively and inductively coupled excitonsin bilayer MoS2. Phys. Rev. Lett. 129, 107401 (2022).48. Hagel, J., Brem, S. & Malic, E. Electrical tuning of moiré excitons inmose2 bilayers. 2D Mater. 10, 014013 (2022).49. Katsch, F., Selig, M., Carmele, A. & Knorr, A. Theory of exciton-exciton interactions inmonolayer transitionmetal dichalcogenides.Phys. status solidi (b) 255, 1800185 (2018).50. Hagel, J., Brem, S., Linderälv, C., Erhart, P. & Malic, E. Excitonlandscape in van der waals heterostructures. Phys. Rev. Res. 3,043217 (2021).51. Smoleński, T. et al. Signatures of Wigner crystal of electrons in amonolayer semiconductor. Nature 595, 53–57 (2021).52. Mittenzwey, H. et al. Ultrafast optical control of Rashba interactionsin a TMDC heterostructure. Phys. Rev. Lett. 134, 026901 (2025).53. Jasiński, J. et al. Quadrupolar excitons in mose2 bilayers. Nat.Commun. 16, 1382 (2025).54. Kiriya, D., Tosun, M., Zhao, P., Kang, J. S. & Javey, A. Air-stable sur-face charge transfer doping of MoS2 by benzyl viologen. J. Am.Chem. Soc. 136, 7853–7856 (2014).55. Reed-Lingenfelter, S. N. et al. Compact super electron-donor tomonolayer MoS2. Nano Lett. 22, 4501–4508 (2022).56. Yarali, M. et al. Near-unitymolecular doping efficiency inmonolayerMoS2. Adv. Electron. Mater. 7, 2000873 (2020).57. Bonaccorso, F., Bartolotta, A., Coleman, J. N. & Backes, C. 2d-crystal-based functional inks. Adv. Mater. 28, 6136–6166 (2016).58. Hernández López, P. et al. Strain control of hybridization betweendark and localized excitons in a 2d semiconductor. Nat. Commun.13, 7691 (2022).59. Kumar, A.M. et al. Strainfingerprintingof exciton valley character in2d semiconductors. Nat. Commun. 15, 7546 (2024).60. High, A. A., Novitskaya, E. E., Butov, L. V., Hanson, M. & Gossard, A.C. Control of exciton fluxes in an excitonic integrated circuit. Sci-ence 321, 229–231 (2008).61. Ciarrocchi, A., Tagarelli, F., Avsar, A. & Kis, A. Excitonic devices withvan der Waals heterostructures: valleytronics meets twistronics.Nat. Rev. Mater. 7, 449–464 (2022).62. Castellanos-Gomez, A. et al. Deterministic transfer of two-dimensional materials by all-dry viscoelastic stamping. 2D Mater. 1,011002 (2014).63. Yagodkin, D. et al. Extrinsic localized excitons in patterned 2dsemiconductors. Adv. Funct. Mater. 32, 2203060 (2022).64. Rosenberger, M. R. et al. Nano-"squeegee” for the creation of clean2dmaterial interfaces.ACSAppl.Mater. Interfaces 10, 10379–10387(2018).65. Greben, K., Arora, S., Harats, M. G. & Bolotin, K. I. Intrinsic andextrinsic defect-related excitons in tmdcs.Nano Lett. 20, 2544–2550(2020).AcknowledgementsS.K., K.G., A.M.K., S.P. and K.I.B. acknowledge the Deutsche For-schungsgemeinschaft (DFG) for financial support through the projectGZ: BO 5142/4-1 and Collaborative Research Centers TRR 227 (projectB08), SfB 951 (project B15) and SfB 1772 (project B01), as well as theFederalMinistry of Education and Research (BMBF, 05K2022 – ioARPES).J.S., Q.C. and S.E. acknowledge financial support from the DPG via SfB1772 (project C01). D.C., M.S. and A.K. acknowledge financial supportfrom the DPG via SfB 1772 (project A03). S.K. acknowledges usefulconversations with Nele Stetzuhn, Ben Weintrub, Denis Yagodkin,Georgy Gordeev, and Theresia Knobloch.Author contributionsS.K., K.G. and K.I.B. conceived the project. S.K. and K.G. designed theexperimental setup and developed the in situ evaporation technique.S.K., A.M.K. and S.P. prepared the samples and performed the opticalmeasurements. J.S., Q.C. and S.E. synthesized and characterized theArticle https://doi.org/10.1038/s41467-025-65431-6Nature Communications |         (2025) 16:9893 7www.nature.com/naturecommunicationsCN6-CPmolecules. K.W. and T.T. grew the hBN crystals. D.C., M.S., A.K.,developed a theory for excitons. S.K. performed electrostatic simula-tions. S.K. and S.P. analyzed the data. S.K. and K.I.B. wrote the manu-script with input from all co-authors.FundingOpen Access funding enabled and organized by Projekt DEAL.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-025-65431-6.Correspondence and requests for materials should be addressed toSviatoslav Kovalchuk or Kirill I. Bolotin.Peer review information Nature Communications thanks the anon-ymous reviewers for their contribution to the peer review of this work. Apeer review file is available.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard to jur-isdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,adaptation, distribution and reproduction in any medium or format, aslong as you give appropriate credit to the original author(s) and thesource, provide a link to the Creative Commons licence, and indicate ifchanges were made. The images or other third party material in thisarticle are included in the article’s Creative Commons licence, unlessindicated otherwise in a credit line to the material. If material is notincluded in the article’s Creative Commons licence and your intendeduse is not permitted by statutory regulation or exceeds the permitteduse, you will need to obtain permission directly from the copyrightholder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2025Article https://doi.org/10.1038/s41467-025-65431-6Nature Communications |         (2025) 16:9893 8https://doi.org/10.1038/s41467-025-65431-6http://www.nature.com/reprintshttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/www.nature.com/naturecommunications Revealing hidden interlayer excitons in 2D bilayers via hybrid molecular gating Device concept/Evaporation Technique Results Stark splitting in bilayer TMD systems in linear approximation Coupling between inter- and intralayer excitons Dark excitons in 2L-MoSe2 Discussion Methods Device fabrication In situ evaporation Optical measurements Data availability References Acknowledgements Author contributions Funding Competing interests Additional information