# Fileset [2025A01280G_Paper-CGT-PRA-3rdRevision-Accepted.pdf](https://mdr.nims.go.jp/filesets/13f79b02-291f-4a31-ac32-09dac5f100c4/download) ## Creator Annika Bergmann-Iwe, Swarup Deb, Klaus Zollner, Veronika Schneidt, Mustafa Hemaid, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Rico Schwartz, Jaroslav Fabian, Tobias Korn ## Rights ©2026 American Physical Society[In Copyright](http://rightsstatements.org/vocab/InC/1.0/) ## Other metadata [Effect of spin-dependent tunneling in a Mo Se 2 / Cr 2 Ge 2 Te 6 van der Waals heterostructure on exciton and trion emission](https://mdr.nims.go.jp/datasets/2088c377-75f0-402c-ba52-083b9ee136ad) ## Fulltext Effect of spin-dependent tunneling in a MoSe2/Cr2Ge2Te6 van der Waalsheterostructure on exciton and trion emissionAnnika Bergmann-Iwe,1 Swarup Deb,1, 2, 3 Klaus Zollner,4 Veronika Schneidt,1 Mustafa Hemaid,1Kenji Watanabe,5 Takashi Taniguchi,6 Rico Schwartz,1 Jaroslav Fabian,4 and Tobias Korn1, ∗1Institute of Physics, Rostock University, 18059 Rostock, Germany2Saha Institute of Nuclear Physics, Kolkata, India3Homi Bhabha National Institute, Mumbai, India4Institute for Theoretical Physics, Regensburg University, 93040 Regensburg, Germany5Research Center for Electronic and Optical Materials, NIMS, 1-1 Namiki, Tsukuba 305-0044, Japan6Research Center for Materials Nanoarchitectonics, NIMS, 1-1 Namiki, Tsukuba 305-0044, JapanWe study van der Waals heterostructures consisting of monolayer MoSe2 and few-layer Cr2Ge2Te6fully encapsulated in hexagonal Boron Nitride using low-temperature photoluminescence and polarmagneto-optic Kerr effect measurements. Photoluminescence characterization reveals a partialquenching and a change of the exciton-trion emission ratio in the heterostructure as compared tothe isolated MoSe2 monolayer. Under circularly polarized excitation, we find that the exciton-trionemission ratio depends on the relative orientation of excitation helicity and Cr2Ge2Te6 magnetization,even though the photoluminescence emission itself is unpolarized. This observation hints at anultrafast, spin-dependent interlayer charge transfer that competes with exciton and trion formationand recombination.I. INTRODUCTIONIn recent years, two-dimensional (2D) crystals and vander Waals (vdW) heterostructures [1] consisting of differ-ent 2D crystals have been one of the most active fieldsin solid-state research. Besides graphene, the semicon-ducting transition metal dichalcogenides (TMDCs) suchas MoSe2 have garnered a lot of research attention. Thisis due to their exciting electronic and optical properties.In the monolayer (ML) limit, they become direct-gapsemiconductors [2] with a peculiar band structure leadingto spin-valley coupling [3]. These properties make thempotentially interesting for spintronics [4] and valleytron-ics [5], where information is encoded in the spin or valleydegree of freedom of carriers, instead of their charge. InML TMDCs, the optical selection rules allow generationof a coupled spin-valley polarization of excitons usingcircularly polarized excitation, and the excitonic valleypolarization degree can be read out directly in helicity-resolved photoluminescence (PL) measurements [6, 7].Depending on the specific TMDC material, this mech-anism for generating a valley polarization can be veryeffective and robust, even for highly nonresonant excita-tion.For the specific case of MoSe2, however, the valleyrelaxation rate is extremely fast compared to that ofexciton recombination, so that a significant circular po-larization of the PL emission can only be observed fornear-resonant excitation [8–10]. Despite the unfavour-able relaxation rates, an excitonic valley polarization canbe achieved by lifting the energy degeneracy of oppositevalleys. A usual strategy has been to exploit externalmagnetic fields perpendicular to the plane of a TMDC ML∗ tobias.korn@uni-rostock.deto break time-reversal symmetry. This introduces a valleyZeeman splitting and leads to a preferential occupation ofthe energetically favorable valley, even for unpolarized andnonresonant excitation [11, 12]. However, the effective gfactors for TMDC monolayers correspond to a splittingof only about 0.2 meV per Tesla, so that magnetic fieldsof several Tesla are required to achieve a significant valleypolarization even at liquid helium temperature.As such field strengths are impractical for device applic-ations, the use of magnetic proximity effects [13], wherethe proximity of a magnetic material induces a magnet-ization through exchange interaction, has been exploredin recent years. This has made it possible to tailor thespin-valley properties in TMDCs without the need for anexternal magnetic field [14]. Alternatively, spin injectionfrom ferromagnetic materials was used to generate valleypolarization in TMDCs [15]. A variety of bulk ferromag-netic materials ranging from metals like Nickel [16] toferromagnetic semiconductors such as Ga(Mn)As [15] andEuS [17] were employed in experimental studies. However,the presence of surface states, dangling bonds, interfacereconstruction, etc. greatly complicates their integrationinto vdW heterostructures.For this, layered ferromagnetic materials [18, 19], madeof chemically stable atomically thin sheets, have recentlyemerged as a viable alternative for integration into vdWheterostructures. While there have been a number ofexperimental studies [14, 20–22], among the plethora ofavailable TMDC and vdW ferromagnets, it remains elu-sive which material combinations, thicknesses, and otherparameters offer the best material platform for study-ing phenomena such as spin-dependent tunneling andmagnetic proximity effects.Here, we report on an optical spectroscopy study ofvdW heterostructures consisting of monolayer MoSe2 andthe ferromagnetic semiconductor Cr2Ge2Te6 (CGT) [23,24] encapsulated in hexagonal Boron Nitride (hBN). Low-mailto:tobias.korn@uni-rostock.de2temperature photoluminescence measurements reveal apartial quenching of the PL and a significantly reducedtrion-to-exciton emission ratio in the heterostructure ascompared to an isolated MoSe2 ML, indicating an electrontransfer from the MoSe2 into the CGT. Under nonres-onant, circularly polarized excitation, we find that theexciton-trion ratio depends on the relative orientation ofexcitation helicity and CGT magnetization. Remarkably,this effect is observable even though the PL emissionitself is unpolarized, indicating a vanishing valley polar-ization of excitons and trions. DFT calculations of theheterostructure predict a type-I band alignment and apronounced spin splitting of the CGT conduction band,yielding a large spin-polarized density of states for inter-layer electron tunneling. Based on this, we interpret ourobservations as an interplay of spin-dependent tunnelingof the initially valley-polarized electrons from MoSe2 intoCGT with exciton and trion formation, valley relaxationand recombination.Our findings offer a glimpse into material combinationsthat can be utilized for spin generation and hosting, andmore importantly, provide a potential means to read outspin polarization and light helicity through spin-selectivetunneling - an essential component in the process andmeasurement chain of spin-valleytronic devices.II. RESULTS AND DISCUSSIONTo begin with, we discuss the structure and character-ization of our samples. As an example, Figure 1a showsa fully hBN-encapsulated MoSe2/CGT heterostructure.Attached to bulk CGT, two thin CGT regions I and IIof different thickness cover an underlying MoSe2 mono-layer and are thus of particular interest. After exfoliation,optical images of the CGT flake were taken in transmis-sion mode inside an inert gas glovebox used for samplepreparation (see methods). From the absorbance in con-junction with AFM measurements, we estimate the CGTflake to consist of 14 layers in region I and of 9 layersin region II (see supplementary materials [25] Figure S1).Low-temperature PL measurements at 5K were used tocharacterize the resulting heterostructure. The samplewas scanned and the PL map depicted in Figure 1b wascreated using an automated routine that fitted the mostintense peak in each spectrum with a Gaussian. Numer-ical integration of the energetic range containing bothtrion and exciton emission reveals a quenching by a factorof about five of the MoSe2 monolayer emission in theheterostructure region compared to the isolated MoSe2monolayer, indicating a high-quality interfacial contactfor interlayer charge transfer between both layers [26].Maps to further characterize the exciton and trion emis-sion as well as the exciton and trion peak energies canbe found in Figure S2. Due to the CGT flake on topof the MoSe2 monolayer, which acts as an additionalabsorbing layer, the quenching is slightly overestimated.Exemplary spectra (Fig.1c) illustrate that both trion andexciton emission are suppressed in the heterostructureregion, with trion emission being more strongly affected.Whereas trion emission dominates the isolated monolayerspectrum, the exciton-trion ratio is significantly largerin the heterostructure region. Whether either trion orexciton emission is more pronounced in the spectrumdepends on the position on the heterostructure, whichwe attribute to locally different interface qualities due toinclusions between the two layers. A color map depictingthe spatial distribution of exciton-trion ratios is shown inFigure S3.(c)(a)PL Intensity (arb. units)10 µm lowhigh(b)Energy (eV)10203040PL Intensity (kCounts), shifted1.5 1.6 1.70x5x5x5TXMoSe2MonolayerCGT/MoSe2structureHetero-(c)Kerr-Signal(µV), shiftedApplied Field (mT)III05100510-25-50 0 25 50(d)IIIhBNCGTMoSe2hBN(a)10 µmFigure 1. (a) Optical microscope image of an hBN-encapsulated MoSe2/CGT heterostructure. A CGT flake withregions of various layer thickness (I) and (II) covers an under-lying MoSe2 monolayer. The yellow box corresponds to thefalse color map of the PL scan area (b), showing quenchingof the MoSe2 monolayer emission in the heterostructure re-gion. (c) Compared to the MoSe2 monolayer the integratedPL emission is reduced by a factor of about five. In addi-tion, the ratio of exciton (X) and trion (T) is increased in theheterostructure, with different exciton-trion ratios observed,depending on the measurement position. All spectra weretaken from the PL scan. Their positions are marked by thecolored dots corresponding to the color of the spectrum. (d)MOKE measurements prove ferromagnetic behavior of theCGT flake. Magnetic saturation is reached at 32mT and18mT for I and II, respectively. The arrows indicate the ap-plied field of ± 50mT for the helicity-resolved measurementsdescribed below.For further polarization-resolved PL measurements, afully saturated magnetization along the out-of-plane axisof the CGT flake is crucial. This makes the magnetiza-tion (anti-)parallel to the spin orientation in the TMDCK valleys. Polar magneto-optical Kerr effect (MOKE)measurements at a nominal sample temperature of 5K31.58 1.60 1.62 1.64 1.66PL Intensity (arb. units) Energy (eV)1.58 1.60 1.62 1.64 1.66Energy (eV)Excitation Applied Field-σExcitation+σλ/4 Rotation Angle (°)0 600.000.020.04rel.IntensityDifference-20-60 -40 4020-0.04-0.02TrionExciton(b)(a)Figure 2. (a) Low-temperature PL spectra taken in the heterostructure region II in presence of an external magnetic field(± 50mT) for left (σ−) and right (σ+) circularly polarized excitation. For a fixed excitation helicity, trion and exciton intensitieschange in dependence of the magnetic field direction (arrows). (b) The relative intensity difference for exciton and trion uponflip of magnetic field depends on the degree of circularly polarized excitation, vanishing for linear polarization at a λ/4 rotationangle close to 0 ◦. The solid lines represents the sliding average of five data points and serve as guide to the eye. The data werecollected in heterostructure region I. The maximum relative intensity differences observed are comparable to those in region II.(illustrated in Figure 1d) show clear hysteresis loops forboth heterostructure regions. As expected, the thinnerregion II shows a smaller Kerr signal range than thethicker region I. We observe a small remanence and lowcoercive field for region II. By contrast, we find a morecomplex, bow-tie-like shape of the hysteresis loop for thethicker region I, which resembles previous reports on hys-teresis loops for CGT flakes thicker than 13 layers [27] or10 nm [28], closely matching our thickness estimates basedon absorbance. As a consequence of the small magneticremanence observed in MOKE, all subsequent measure-ments were conducted at an external magnetic field of± 50mT, ensuring magnetization saturation parallel tothe direction of the external magnetic field.Valley polarization effects were investigated by nonres-onant PL measurements using an excitation wavelength of633 nm. Circularly polarized excitation was achieved usingan achromatic quarter-wave-plate (λ/4-plate). A secondλ/4-plate in conjunction with a linear polarizer was usedfor helicity-resolved detection. With a constant excitationhelicity, multiple PL spectra were measured with altern-ating magnetic field direction. Spectra recorded at thesame magnetic field direction were then summed up to im-prove the signal-to-noise ratio. This was repeated for theopposite excitation helicity. We emphasize that in thesemeasurements the detection helicity was fixed. Remark-ably, the population dynamics described below is observedirrespective of the detection helicity and can also be seenin linearly polarized detection (Fig. S4). Figure 2a showsevidence for magnetic-field-dependent changes of excitonand trion emission. The spectra measured at a spot inheterostructure region II show that for left-polarized (σ−)excitation populating the K− valley, the trion intensityis larger when a positive (red) magnetic field is applied,compared to a negative (black) magnetic field of the samemagnitude. The opposite trend is observed for excitonemission where a larger intensity is seen at a negativemagnetic field. In contrast, under right-polarized (σ+)excitation the intensity distributions for both trion and ex-citon are reversed. However, the magnetic-field-dependentintensity variation has a different magnitude for right andleft circularly polarized excitation. We attribute this toa minor deviation from the optimum excitation helicity.In addition, a small beam offset induced by rotating theλ/4-plate can result in a slightly different excitation spoton the sample.Qualitatively, the same effect occurs in heterostructureregion I and is reproduced in another sample (see Fig.S5).To quantify the normalized magnetic-field-induced in-tensity differences, we define them as:∆I =I(B ↓)− I(B ↑)I(B ↓) + I(B ↑)(1)where ↓↑ denotes the external magnetic field direction. Ex-citon and trion intensities were determined by numericalintegration around the respective peak positions. Chan-ging the degree of circular polarization by rotating theλ/4-plate in the excitation (Fig. 2b), demonstrates avanishing intensity difference for linear excitation corres-ponding to an angle of about 0 ◦, whereas an increasingdifference occurs for excitation deviating from linear po-larization, reaching maximum values of about 4%.To confirm that the intensity differences result froman interaction between the MoSe2 monolayer and themagnetized CGT, the same measurement was performedon an isolated part of the monolayer. Here, the sametrion (exciton) intensity was detected irrespective of themagnetic field direction for both left and right circularlypolarized excitation (Fig. S5).We note that we do not observe any discernible ValleyZeeman splitting in our helicity-resolved PL measure-ments, which were performed using applied magneticfields of 50 mT, neither for the heterostructure nor for4the isolated MoSe2. Our experimental accuracy shouldallow us to resolve Valley Zeeman splitting on the orderof 0.5 meV or above. This gives an upper limit for anyproximity-effect-induced Valley Zeeman splitting effectpresent in our samples. Such values are well below thosereported in previous studies [14, 22]. Our calculationsyield a value of 3.7 meV in the single-particle picturebased on the splitting of MoSe2 conduction and valencebands (see Fig. S11). However, we note that the theorycalculation is likely to overestimate the splitting as itrequires crystallographic alignment of the TMDC andthe CGT layers at a specific twist angle of about 16 de-grees in order to obtain a reasonable supercell size forDFT calculations. By contrast, in the experiment, thealignment angle between the constituents of the hetero-structure is optimized to yield large overlap of flake areasand alignment of straight edges, which is likely to result intwist angles close to zero degrees. Previous theory calcula-tions on twist angle dependence have shown the magneticproximity effect to be fragile and strongly dependent oninterlayer twist [29, 30].Based on work functions for both MoSe2 monolayer andCGT flakes of the determined thickness [21], the hetero-structure should possess a type I band alignment with theMoSe2 conduction band being energetically higher thanthat of CGT, which is verified by DFT calculations. Theseare discussed in detail in the supplementary materials(See [25] and references therein [31–43]). The exciton andtrion emission quenching in the heterostructure regioncompared to the isolated monolayer as well as the largerexciton-trion ratio in the heterostructure indicate electrontransfer from MoSe2 into CGT. This is visible in FiguresS2a and S2c, as well as Figure 1c. In the latter, sincethe quenching is more pronounced for the trion, a shift ofrelative spectral weight towards the exciton is observed.Predominantly, the exciton-trion ratio in the hetero-structure is modulated by the following mechanisms: Elec-tron transfer reduces the background carrier density inMoSe2, lowering the probability of trion formation, whichleads to an increase in the exciton-trion emission ratiotypical for TMDCs [44, 45]. Besides, trions in MoSe2 havesignificantly longer photoluminescence lifetimes (15 ps)than excitons (2 ps) [8, 46], so that they are more sus-ceptible to nonradiative decay channels, such as electrontunneling into the CGT, during their lifetime.In addition to the common carrier density-dependentchanges in PL intensity of exciton and trion, we observe asubtle helicity- and magnetic field-dependent modulationin the exciton-trion emission ratio, as shown in Figure 2.Interestingly, for a fixed excitation helicity, trion and ex-citon show opposite intensity changes when the magneticfield is flipped. We identify two main factors that explainthis observation: (i) spin-dependent tunneling of electronsfrom MoSe2, which determines the quasi-equilibrium car-rier concentration under photo-excitation and (ii) theexciton-trion conversion, which depends on excess carrierconcentration, as well as on competing exciton and trionformation and recombination times.K-K+Ϭ-tunnelingscatteringTCGTMoSe2recombinationexcitationϬ+XEkEDOSϬ-Figure 3. Illustration of processes causing the population dy-namics for σ− excitation and B ↓. Spin-down electrons in theK− valley tunnel into the CGT due to a high density of avail-able spin-down states in the conduction band. The decreasedcharge carrier density in MoSe2 reduces the trion formationrate and a large fraction of excitons recombines radiatively. Achange of magnetic field direction would result in a smallerdensity of spin-down states in the CGT, leading to increasedtrion formation at the expense of exciton recombination. Theinitial valley polarization is lost due to scattering between thevalleys, causing the collected PL to stem from both valleys.The collected PL is thus irrespective of the detection helicity.To explain the intricate interplay of these factors, weconsider the behavior under left circularly polarized ex-citation (σ−), which selectively populates the K− valleywith spin-down electrons, as illustrated in Figure 3.Due to the external magnetic field, the CGT flake ismagnetically saturated with the spin states aligned withthe external magnetic field direction. As evident fromDFT calculations, a spin-dependent splitting of the CGTconduction band exists. For a negative magnetic field, theelectrons residing in the K− valley therefore encounter ahigh density of available spin-down states in the CGT con-duction band. This increases the tunneling rate, whichreduces the number of free spin-down electrons in theMoSe2 conduction band. Interlayer charge transfer inTMDC heterostructures has been reported, dependingon the material combination, to usually occur within50 fs [47, 48]. Similar transfer rates can be expected forTMDC-CGT heterostructures. Whereas a considerableportion of excitons already forms within 0.4 ps after ex-citation [49], in parallel with interlayer charge transfer,trion formation in MoSe2 monolayers takes place on amuch slower timescale of 2 ps [50] and depends on thebackground carrier density. In the presence of a negativemagnetic field, trion formation and subsequent recom-bination is thus suppressed as a result of the reducedavailability of excess electrons in MoSe2. In parallel, weobserve a higher exciton intensity at a negative magnetic5field. As trion formation and exciton recombination occuron similar timescales of about 2 ps [8, 34], a larger frac-tion of excitons recombines radiatively. In contrast, for apositive magnetic field, the tunneling rate for spin-downelectrons is reduced, so that trion formation is enhancedand consequently a larger fraction of excitons is convertedto trions before recombination.Changing the excitation helicity to σ+ flips the spincompatibilities with the spin-split density of states inthe CGT and thus reverses the behavior explained above.Therefore, the magnetic-field dependent differences of ex-citon and trion emission invert their sign, as observedin Figure 2b. Notably, the effect described above wasreproduced in a second sample (Fig. S5), although there,measurement positions were also found where both trionand exciton emission were suppressed under conditions ofhigh tunneling probabilities (Fig. S6). We attribute thesedistinct manifestations of the spin-selective tunneling pro-cess to differences in the relative rates of the processesinvolved, highlighting the sensitivity of the sample systemto local changes.At this point, it shall again be highlighted that theoverall detected PL emission is unpolarized. We ascribethis to intervalley scattering mediated by long-range Cou-lomb interaction [51] which was identified to be the mainmechanism for exciton valley depolarization after initialcircular excitation in MoSe2 [9]. We assume that a consid-erable fraction of excitons and trions is scattered betweenthe K+ and K− valleys before exciton and trion recom-bination take place. With fast depolarization rates largerthan 1 ps−1 for MoSe2 monolayers under weak excitationconditions [52], the scattering process competes with orundercuts MoSe2 exciton and trion lifetimes of about 2 psand 15 ps, respectively [8, 46]. The dependence of theintensity difference on the degree of excitation polariz-ation can directly be explained by the different initialpopulation of the valleys. The imbalance between thetwo valleys rises as the degree of excitation polarizationincreases, which is reflected in larger intensity differencesupon reversal of the magnetic field. Remarkably, thechanges in exciton and trion populations observed in PLare thus a consequence of a spin-dependent tunneling pro-cess, even though any initial spin and valley polarizationis lost well before exciton and trion recombination.In summary, we have fabricated MoSe2/CGT hetero-structures that reveal charge transfer from the MoSe2monolayer into the CGT, as predicted from the bandalignment resulting from DFT calculations. The magnet-ization of the CGT flake and resultant spin-split densityof states enables spin-dependent tunneling after nonreson-ant, circularly polarized excitation which manifests itselfin altered exciton and trion intensities upon change of theexternal magnetic field direction. Intervalley scatteringcauses a loss of initial spin and valley polarization priorto exciton and trion recombination, resulting in an unpo-larized collected PL. Our results underline the complexinterplay of these competing processes on sub-picosecondtimescales.The observed effect holds significant promise for deviceapplications. Circularly polarized excitation of our struc-tures yields a helicity-dependent exciton/trion occupationimbalance which arises from spin-dependent tunneling.Consequently, the tunnel current itself depends on ex-citation helicity, potentially allowing for direct electricalreadout of light helicity in suitably contacted heterostruc-tures without the use of wave plates and polarizers toanalyze the light polarization. A related concept wasdemonstrated recently using CrI3 [53]. While the use ofCGT with its low Curie temperature of about 60 K [23]limits applications, ferromagnetic behavior well aboveroom temperature has been reported, e.g., in the layeredferromagnetic material Fe3GaTe2 [54, 55]. Given the highremanence observed in this material, it may support theobserved effect without the applied external magneticfield needed to align the magnetization in the case ofCGT. Other relevant properties like its band alignmentwith MoSe2 will require further investigation. With therapid evolution of layered magnetic materials in recentyears, other suitable alternatives to CGT are likely toemerge [56].The observed effect relies neither on valley polarizationnor on proximity-effect-induced valley Zeeman splittingand can thus be considered as rather robust. As ourresults demonstrate, nonresonant excitation yields a siz-able occupation imbalance, indicating a spectrally broadrange for helicity detection. Additionally, the picosecondtimescales associated with the observed effect potentiallyenable ultrafast detector response times.III. METHODSA. Sample fabricationUnder ambient conditions, hBN flakes and MoSe2 mono-layers were isolated from bulk crystals by mechanical ex-foliation. The hBN flake was stamped with PDMS onto aSi/SiO2 substrate via deterministic transfer [57]. A MoSe2monolayer was deposited on the hBN accordingly and af-terwards annealed at about 180 ◦C in mild vacuum. CGTflakes were exfoliated from a bulk crystal (HQ graphene)under nitrogen atmosphere in a glovebox [58], thin flakeswere thereby identified under an optical microscope basedon their optical contrast. Immediately after exfoliation,the CGT flake was placed on top of the beforehand pre-pared hBN/MoSe2 structure using a second deterministicstamping setup and PDMS transfer inside the glovebox.To protect the CGT from oxidation, the heterostructurewas then fully encapsulated inside the glovebox by addingan hBN top layer.6B. Optical measurements1. PhotoluminescenceFor the PL measurements the sample was excited witha 1.96 eV continuous-wave diode laser focused to a spotsize of about 1 µm using an 80x microscope objective. Thesample was mounted in a He-flow cryostat and cooled toa nominal temperature of about 5K. To prevent sampleheating, the excitation density was kept below 4 kW/cm2.The PL light emitted by the sample was collected usingthe same objective, filtered by long pass and analyzed witha combination of a spectrometer and a charge-coupled-device. To obtain PL maps of the sample the cryostat,with the sample inside, was moved in relation to thefixed laser spot through a computer-controlled xy stage.For helicity-resolved excitation, a linear polarizer and anachromatic quarter-wave plate (λ/4) were placed in theexcitation beam path. The wave plate was mounted in amotorized rotation stage, so that its angle could be variedautomatically. Similarily, for helicity-resolved detection,an achromatic λ/4-plate and a linear polarizer (actingas an analyzer) were placed in the detection beam pathin front of the spectrometer. For application of externalmagnetic fields, an air coil was placed around the cryostat,so that magnetic fields of up to 200 mT could be appliedperpendicular to the sample plane. The current for thecoil was supplied by a bipolar current source.2. Polar magneto-optical Kerr effectThe polar MOKE measurements were performed inthe setup described above using the 1.96 eV diode laserapplying an excitation density of about 1 kW/cm2. Allmeasurements shown in the manuscript were performed ata nominal temperature of about 5K. Laser intensity wasmodulated using a flywheel chopper. A beamsplitter cubewas introduced into the detection beam path, so that thereflected laser light could be guided to an optical bridgedetector with a pair of balanced photodiodes (see [59] fordetails). The difference signal from the photodiodes wasdetected using a lock-in amplifier. MOKE loops weremeasured by tracking the difference signal as a functionof the applied magnetic field controlled by the bipolarcurrent source.IV. ACKNOWLEDGEMENTSThe authors gratefully acknowledge technical assist-ance by E. Moldt and M. Kronseder, as well as fruitfuldiscussions with R. Edhib. S. D. acknowledges financialsupport by the Humboldt foundation and a startup fund-ing grant provided by the Deutsche Forschungsgemeinsch-aft (DFG, German Research Foundation) via SPP2244.T.K. acknowledges financial support by the DFG viathe following grants: SFB1477 (project No. 441234705),SPP2244 (project No. 443361515), KO3612/7-1 (projectNo. 467549803) and KO3612/8-1 (project No. 549364913).K. Z. and J. F. acknowledge funding by the DFG via SFB1277 (Project No. 314695032) and SPP 2244 (ProjectNo. 443416183), as well as the European Union Horizon2020 Research and Innovation Program under contractnumber 881603 (Graphene Flagship) and FLAGERA pro-ject 2DSOTECH. K.W. and T.T. acknowledge supportfrom the JSPS KAKENHI (grant numbers 21H05233 and23H02052) and World Premier International ResearchCenter Initiative (WPI), MEXT, Japan.[1] A. K. Geim and I. V. Grigorieva, Van der Waals hetero-structures., Nature 499, 419 (2013).[2] K. F. Mak, C. Lee, J. Hone, J. Shan, and T. F. Heinz,Atomically thin MoS2: A new direct-gap semiconductor,Phys. Rev. Lett. 105, 136805 (2010).[3] D. Xiao, G. B. Liu, W. Feng, X. Xu, and W. Yao, Coupledspin and valley physics in monolayers of MoS2 and othergroup-VI dichalcogenides, Phys. Rev. Lett. 108, 196802(2012).[4] J. F. Sierra, J. Fabian, R. K. Kawakami, S. Roche, andS. O. Valenzuela, Van der waals heterostructures for spin-tronics and opto-spintronics, Nature Nanotechnology 16,856 (2021).[5] J. R. Schaibley, H. Yu, G. Clark, P. Rivera, J. S. Ross,K. L. Seyler, W. Yao, and X. Xu, Valleytronics in 2Dmaterials, Nature Reviews Materials 1, 16055 (2016).[6] K. F. Mak, K. He, J. Shan, and T. F. Heinz, Control ofvalley polarization in monolayer MoS2 by optical helicity,Nature Nanotechnology 7, 494 (2012).[7] H. Zeng, J. Dai, W. Yao, D. Xiao, and X. Cui, Valleypolarization in MoS2 monolayers by optical pumping,Nature Nanotechnology 7, 490 (2012).[8] G. Wang, E. Palleau, T. Amand, S. Tongay, X. Marie,and B. Urbaszek, Polarization and time-resolved photolu-minescence spectroscopy of excitons in MoSe2 monolayers,Applied Physics Letters 106, 112101 (2015).[9] M. Baranowski, A. Surrente, D. K. Maude, M. Ballot-tin, A. A. Mitioglu, P. C. M. Christianen, Y. C. Kung,D. Dumcenco, A. Kis, and P. Plochocka, Dark excitonsand the elusive valley polarization in transition metaldichalcogenides, 2D Materials 4, 025016 (2017).[10] H. Tornatzky, A.-M. Kaulitz, and J. Maultzsch, Resonanceprofiles of valley polarization in single-layer MoS2 andMoSe2, Phys. Rev. Lett. 121, 167401 (2018).[11] D. MacNeill, C. Heikes, K. F. Mak, Z. Anderson,A. Kormányos, V. Zólyomi, J. Park, and D. C. Ralph,Breaking of valley degeneracy by magnetic field in mono-layer MoSe2, Phys. Rev. Lett. 114, 037401 (2015).[12] A. Mitioglu, P. P. Plochocka-Maude, A. Granados delAguila, P. C. M. Christianen, G. Deligeorgis, S. Anghel,L. Kulyuk, and D. K. Maude, Optical investigation ofmonolayer and bulk tungsten diselenide (WSe2) in highhttps://doi.org/10.1038/nature12385https://doi.org/10.1103/PhysRevLett.105.136805https://doi.org/10.1103/PhysRevLett.108.196802https://doi.org/10.1103/PhysRevLett.108.196802https://doi.org/10.1038/s41565-021-00936-xhttps://doi.org/10.1038/s41565-021-00936-xhttps://doi.org/10.1038/natrevmats.2016.55https://doi.org/10.1038/nnano.2012.96https://doi.org/10.1038/nnano.2012.95https://doi.org/10.1063/1.4916089https://doi.org/10.1088/2053-1583/aa58a0https://doi.org/10.1103/PhysRevLett.121.167401https://doi.org/10.1103/PhysRevLett.114.0374017magnetic fields, Nano Letters , 150611125011005 (2015).[13] B. Scharf, G. Xu, A. Matos-Abiague, and I. Žutić, Mag-netic proximity effects in transition-metal dichalcogenides:Converting excitons, Phys. Rev. Lett. 119, 127403 (2017).[14] D. Zhong, K. L. Seyler, X. Linpeng, R. Cheng, N. Sivadas,B. Huang, E. Schmidgall, T. Taniguchi, K. Watanabe,M. A. McGuire, W. Yao, D. Xiao, K.-M. C. Fu, and X. Xu,Van der waals engineering of ferromagnetic semiconductorheterostructures for spin and valleytronics, Science Ad-vances 3, e1603113 (2017).[15] Y. Ye, J. Xiao, H. Wang, Z. Ye, H. Zhu, M. Zhao, Y. Wang,J. Zhao, X. Yin, and X. Zhang, Electrical generation andcontrol of the valley carriers in a monolayer transitionmetal dichalcogenide, Nature Nanotechnology 11, 598(2016).[16] H. Tornatzky, C. Robert, P. Renucci, B. Han, T. Blon,B. Lassagne, G. Ballon, Y. Lu, K. Watanabe, T. Tanigu-chi, B. Urbaszek, J. M. J. Lopes, and X. Marie, Spindependent charge transfer in MoSe2/hBN/Ni hybrid struc-tures, Applied Physics Letters 119, 263103 (2021).[17] C. Zhao, T. Norden, P. Zhang, P. Zhao, Y. Cheng, F. Sun,J. P. Parry, P. Taheri, J. Wang, Y. Yang, T. Scrace,K. Kang, S. Yang, G.-x. Miao, R. Sabirianov, G. Kioseo-glou, W. Huang, A. Petrou, and H. Zeng, Enhanced valleysplitting in monolayer WSe2 due to magnetic exchangefield, Nature Nanotechnology 12, 757 (2017).[18] K. F. Mak, J. Shan, and D. C. Ralph, Probing and con-trolling magnetic states in 2D layered magnetic materials,Nature Reviews Physics 1, 646 (2019).[19] S. Zhang, H. Wu, L. Yang, G. Zhang, Y. Xie, L. Zhang,W. Zhang, and H. Chang, Two-dimensional magneticatomic crystals, Mater. Horiz. 9, 559 (2022).[20] T. P. Lyons, D. Gillard, A. Molina-Sánchez, A. Misra,F. Withers, P. S. Keatley, A. Kozikov, T. Taniguchi,K. Watanabe, K. S. Novoselov, J. Fernández-Rossier, andA. I. Tartakovskii, Interplay between spin proximity effectand charge-dependent exciton dynamics in MoSe2/CrBr3van der waals heterostructures, Nature Communications11, 6021 (2020).[21] S. Rahman, B. Liu, B. Wang, Y. Tang, and Y. Lu,Giant photoluminescence enhancement and resonantcharge transfer in atomically thin two-dimensionalCr2Ge2Te6/WS2 heterostructures, ACS Applied Materi-als & Interfaces 13, 7423 (2021).[22] T. Zhang, S. Zhao, A. Wang, Z. Xiong, Y. Liu, M. Xi,S. Li, H. Lei, Z. V. Han, and F. Wang, Electrically andmagnetically tunable valley polarization in monolayerMoSe2 proximitized by a 2D ferromagnetic semiconductor,Advanced Functional Materials 32, 2204779 (2022).[23] C. Gong, L. Li, Z. Li, H. Ji, A. Stern, Y. Xia, T. Cao,W. Bao, C. Wang, Y. Wang, Z. Q. Qiu, R. J. Cava,S. G. Louie, J. Xia, and X. Zhang, Discovery of intrinsicferromagnetism in two-dimensional van der waals crystals,Nature 546, 265 (2017).[24] Y. F. Li, W. Wang, W. Guo, C. Y. Gu, H. Y. Sun, L. He,J. Zhou, Z. B. Gu, Y. F. Nie, and X. Q. Pan, Electronicstructure of ferromagnetic semiconductor CrGeTe3 byangle-resolved photoemission spectroscopy, Phys. Rev. B98, 125127 (2018).[25] See Supplemental Material at for additional experimentalresults and details of DFT calculations.[26] H. Fang, C. Battaglia, C. Carraro, S. Nemsak, B. Ozdol,J. S. Kang, H. A. Bechtel, S. B. Desai, F. Kronast, A. A.Unal, G. Conti, C. Conlon, G. K. Palsson, M. C. Martin,A. M. Minor, C. S. Fadley, E. Yablonovitch, R. Maboud-ian, and A. Javey, Strong interlayer coupling in van derwaals heterostructures built from single-layer chalcogen-ides, Proceedings of the National Academy of Sciences111, 6198 (2014).[27] A. Vervelaki, K. Bagani, D. Jetter, M.-H. Doan, T. K.Chau, B. Gross, D. V. Christensen, P. Bøggild, andM. Poggio, Visualizing thickness-dependent magnetic tex-tures in few-layer Cr2Ge2Te6, Communications Materials5, 1 (2024).[28] A. Noah, H. Alpern, S. Singh, A. Gutfreund, G. Zisman,T. D. Feld, A. Vakahi, S. Remennik, Y. Paltiel, M. E.Huber, V. Barrena, H. Suderow, H. Steinberg, O. Millo,and Y. Anahory, Interior and edge magnetization in thinexfoliated CrGeTe3 films, Nano Letters 22, 3165 (2022).[29] K. Zollner, P. E. Faria Junior, and J. Fabian, Proximityexchange effects in MoSe2 andWSe2 heterostructures withCrI3: Twist angle, layer, and gate dependence, PhysicalReview B 100, 085128 (2019).[30] K. Zollner, P. E. Faria Junior, and J. Fabian, Strongmanipulation of the valley splitting upon twisting andgating in MoSe2/CrI3 and WSe2/CrI3 van der waals het-erostructures, Physical Review B 107, 035112 (2023).[31] S. R. Bahn and K. W. Jacobsen, An object-oriented script-ing interface to a legacy electronic structure code, Comput.Sci. Eng. 4, 56 (2002).[32] P. Lazic, Cellmatch: Combining two unit cells into acommon supercell with minimal strain, Computer PhysicsCommunications 197, 324 (2015).[33] D. S. Koda, F. Bechstedt, M. Marques, and L. K. Teles,Coincidence lattices of 2D crystals: heterostructure predic-tions and applications, The Journal of Physical ChemistryC 120, 10895 (2016).[34] S. Carr, S. Fang, and E. Kaxiras, Electronic-structuremethods for twisted moiré layers, Nature Reviews Mater-ials 5, 748 (2020).[35] V. Carteaux, D. Brunet, G. Ouvrard, and G. Andre,Crystallographic, magnetic and electronic structures ofa new layered ferromagnetic compound Cr2Ge2Te6, J.Phys.: Condens. Mat. 7, 69 (1995).[36] W. J. Schutte, J. L. De Boer, and F. Jellinek, Crystalstructures of tungsten disulfide and diselenide, Journal ofSolid State Chemistry 70, 207 (1987).[37] P. Hohenberg and W. Kohn, Inhomogeneous electron gas,Phys. Rev. 136, B864 (1964).[38] P. Giannozzi and et al., Quantum espresso: a modularand open-source software project for quantum simulationsof materials, J. Phys.: Cond. Mat. 21, 395502 (2009).[39] G. Kresse and D. Joubert, From ultrasoft pseudopotentialsto the projector augmented-wave method, Phys. Rev. B59, 1758 (1999).[40] J. P. Perdew, K. Burke, and M. Ernzerhof, Generalizedgradient approximation made simple, Phys. Rev. Lett. 77,3865 (1996).[41] S. Grimme, Semiempirical gga-type density functionalconstructed with a long-range dispersion correction, J.Comput. Chem. 27, 1787 (2006).[42] S. Grimme, J. Antony, S. Ehrlich, and H. Krieg, A con-sistent and accurate ab initio parametrization of densityfunctional dispersion correction (DFT-D) for the 94 ele-ments H-Pu, J. Chem. Phys. 132, 154104 (2010).[43] V. Barone, M. Casarin, D. Forrer, M. Pavone, M. Sambi,and A. Vittadini, Role and effective treatment of dispers-ive forces in materials: Polyethylene and graphite crystalshttps://doi.org/10.1021/acs.nanolett.5b00626https://doi.org/10.1103/PhysRevLett.119.127403https://doi.org/10.1126/sciadv.1603113https://doi.org/10.1126/sciadv.1603113https://doi.org/10.1038/nnano.2016.49https://doi.org/10.1038/nnano.2016.49https://doi.org/10.1063/5.0075554https://doi.org/10.1038/nnano.2017.68https://doi.org/10.1038/s42254-019-0110-yhttps://doi.org/10.1039/D1MH01155Chttps://doi.org/10.1038/s41467-020-19816-4https://doi.org/10.1038/s41467-020-19816-4https://doi.org/10.1021/acsami.0c20110https://doi.org/10.1021/acsami.0c20110https://doi.org/https://doi.org/10.1002/adfm.202204779https://doi.org/10.1038/nature22060https://doi.org/10.1103/PhysRevB.98.125127https://doi.org/10.1103/PhysRevB.98.125127https://doi.org/10.1073/pnas.1405435111https://doi.org/10.1073/pnas.1405435111https://doi.org/10.1038/s43246-024-00477-5https://doi.org/10.1038/s43246-024-00477-5https://doi.org/10.1021/acs.nanolett.1c04665https://doi.org/10.1103/physrevb.100.085128https://doi.org/10.1103/physrevb.100.085128https://doi.org/10.1103/physrevb.107.035112https://doi.org/10.1109/5992.998641https://doi.org/10.1109/5992.998641https://doi.org/https://doi.org/10.1016/j.cpc.2015.08.038https://doi.org/https://doi.org/10.1016/j.cpc.2015.08.038https://doi.org/10.1021/acs.jpcc.6b01496https://doi.org/10.1021/acs.jpcc.6b01496https://doi.org/10.1038/s41578-020-0214-0https://doi.org/10.1038/s41578-020-0214-0https://doi.org/10.1088/0953-8984/7/1/008https://doi.org/10.1088/0953-8984/7/1/008https://doi.org/10.1016/0022-4596(87)90057-0https://doi.org/10.1016/0022-4596(87)90057-0https://doi.org/10.1103/PhysRev.136.B864https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1002/jcc.20495https://doi.org/10.1002/jcc.20495https://doi.org/10.1063/1.33823448as test cases, J. Comput. Chem. 30, 934 (2009).[44] K. F. Mak, K. He, C. Lee, G. H. Lee, J. Hone, T. F.Heinz, and J. Shan, Tightly bound trions in monolayerMoS2, Nature Materials 12, 207 (2012).[45] Y. Liu, T. Shen, S. Linghu, R. Zhu, and F. Gu, Electro-static control of photoluminescence from A and B excitonsin monolayer molybdenum disulfide, Nanoscale Advances4, 2484 (2022).[46] C. Robert, D. Lagarde, F. Cadiz, G. Wang, B. Las-sagne, T. Amand, A. Balocchi, P. Renucci, S. Tongay,B. Urbaszek, and X. Marie, Exciton radiative lifetime intransition metal dichalcogenide monolayers, Phys. Rev. B93, 205423 (2016).[47] X. Hong, J. Kim, S.-F. Shi, Y. Zhang, C. Jin, Y. Sun,S. Tongay, J. Wu, Y. Zhang, and F. Wang, Ultrafastcharge transfer in atomically thin MoS2/WS2 heterostruc-tures, Nature Nanotechnology 9, 682 (2014).[48] C. Jin, E. Y. Ma, O. Karni, E. C. Regan, F. Wang,and T. F. Heinz, Ultrafast dynamics in van der waalsheterostructures, Nature Nanotechnology 13, 994 (2018).[49] P. Steinleitner, P. Merkl, P. Nagler, J. Mornhinweg,C. Schüller, T. Korn, A. Chernikov, and R. Huber, Directobservation of ultrafast exciton formation in a monolayerof WSe2, Nano Letters 17, 1455 (2017).[50] A. Singh, G. Moody, K. Tran, M. E. Scott, V. Overbeck,G. Berghäuser, J. Schaibley, E. J. Seifert, D. Pleskot, N. M.Gabor, J. Yan, D. G. Mandrus, M. Richter, E. Malic,X. Xu, and X. Li, Trion formation dynamics in monolayertransition metal dichalcogenides, Physical Review B 93,041401 (2016).[51] M. M. Glazov, T. Amand, X. Marie, D. Lagarde, L. Bouet,and B. Urbaszek, Exciton fine structure and spin decoher-ence in monolayers of transition metal dichalcogenides,Physical Review B 89, 201302 (2014).[52] F. Mahmood, Z. Alpichshev, Y.-H. Lee, J. Kong, andN. Gedik, Observation of exciton–exciton interaction me-diated valley depolarization in monolayer MosS2, NanoLetters 18, 223 (2017).[53] X. Cheng, Z. Cheng, C. Wang, M. Li, P. Gu, S. Yang,Y. Li, K. Watanabe, T. Taniguchi, W. Ji, and L. Dai, Lighthelicity detector based on 2D magnetic semiconductorCrI3, Nature Communications 12, 1 (2021).[54] G. Zhang, F. Guo, H. Wu, X. Wen, L. Yang, W. Jin,W. Zhang, and H. Chang, Above-room-temperaturestrong intrinsic ferromagnetism in 2D van der waalsFe3GaTe2 with large perpendicular magnetic anisotropy,Nature Communications 13, 5067 (2022).[55] S. Wu, Z. He, M. Gu, L. Ren, J. Li, B. Deng, D. Wang,X. Guo, W. Li, M. Chen, Y. Chen, M. Meng, Q. Ye,B. Shen, X. Chen, J. Guo, G. Xing, I. K. Sou, and S. Li,Robust ferromagnetism in wafer-scale Fe3GaTe2 aboveroom-temperature, Nature Communications 15, 10765(2024).[56] B. Zhang, P. Lu, R. Tabrizian, P. X.-L. Feng, and Y. Wu,2D magnetic heterostructures: spintronics and quantumfuture, npj Spintronics 2, 6 (2024).[57] A. Castellanos-Gomez, M. Buscema, R. Molenaar,V. Singh, L. Janssen, H. S. J. van der Zant, and G. A.Steele, Deterministic transfer of two-dimensional materialsby all-dry viscoelastic stamping, 2D Materials 1, 011002(2014).[58] P. Gant, F. Carrascoso, Q. Zhao, Y. K. Ryu, M. Seitz,F. Prins, R. Frisenda, and A. Castellanos-Gomez, A sys-tem for the deterministic transfer of 2D materials underinert environmental conditions, 2D Materials 7, 025034(2020).[59] M. Kempf, A. Schubert, R. Schwartz, and T. Korn, Two-color Kerr microscopy of two-dimensional materials withsub-picosecond time resolution, Review of Scientific In-struments 92, 113904 (2021).https://doi.org/10.1002/jcc.21112https://doi.org/10.1038/nmat3505https://doi.org/10.1039/d2na00071ghttps://doi.org/10.1039/d2na00071ghttps://doi.org/10.1103/PhysRevB.93.205423https://doi.org/10.1103/PhysRevB.93.205423https://doi.org/10.1038/nnano.2014.167https://doi.org/10.1038/s41565-018-0298-5https://doi.org/10.1021/acs.nanolett.6b04422https://doi.org/10.1103/physrevb.93.041401https://doi.org/10.1103/physrevb.93.041401https://doi.org/10.1103/physrevb.89.201302https://doi.org/10.1021/acs.nanolett.7b03953https://doi.org/10.1021/acs.nanolett.7b03953https://doi.org/10.1038/s41467-021-27218-3https://doi.org/10.1038/s41467-022-32605-5https://doi.org/10.1038/s41467-024-54936-1https://doi.org/10.1038/s41467-024-54936-1https://doi.org/10.1038/s44306-024-00011-whttps://doi.org/10.1088/2053-1583/1/1/011002https://doi.org/10.1088/2053-1583/1/1/011002https://doi.org/10.1088/2053-1583/ab72d6https://doi.org/10.1088/2053-1583/ab72d6https://doi.org/10.1063/5.0058110https://doi.org/10.1063/5.0058110 Effect of spin-dependent tunneling in a MoSe2/Cr2Ge2Te6 van der Waals heterostructure on exciton and trion emission Abstract Introduction Results and discussion Methods Sample fabrication Optical measurements Photoluminescence Polar magneto-optical Kerr effect Acknowledgements References