Beyond Intrinsic Limits: Polarization‐Sensitive Carrier Dynamics Amplify Subtle Intrinsic Optical Anisotropy into Giant Linear Dichroism

ABSTRACT

Filterless polarization‐resolved photodetection is attractive for compact optoelectronic systems, but its polarization contrast is fundamentally constrained by the low intrinsic absorption anisotropy of semiconductors. Here, we introduce polarization‐sensitive carrier dynamics (PSCD), a design principle that shifts polarization discrimination from static absorption anisotropy to the time evolution of photocarrier populations. We implement PSCD in a nonvolatile floating‐gate phototransistor by integrating a molecularly thin p‐type organic single crystal with a two‐dimensional covalent organic framework floating gate. Orthogonal linearly polarized light generates different photogeneration rates in the anisotropic crystal, which induce distinct optical‐erasure kinetics and temporally divergent recovery of the channel hole population. Consequently, the absorption linear dichroism ratio (LDR) of 3.4 is dynamically amplified by over four orders of magnitude, reaching a transient photocurrent LDR of 3.6 × 10⁴ under optimized bias conditions. The device further maintains strong polarization discrimination under partially polarized illumination, achieving a transient photocurrent LDR of 9.8 × 10² at a degree of linear polarization of 0.3. These results establish time‐domain amplification of weak optical anisotropy through carrier‐dynamic modulation as a strategy for filterless, high‐contrast polarimetric sensing in compact optoelectronic platforms.

Weak intrinsic optical anisotropy is amplified into giant transient photocurrent dichroism through polarization‐sensitive carrier dynamics in a floating‐gate organic phototransistor. Polarization‐dependent optical erasure produces temporally separated carrier‐recovery pathways, converting an absorption LDR of 3.4 into a photocurrent LDR of 3.6 × 10⁴ and enabling robust polarimetric sensing.

1. Introduction

Polarization‐resolved light detection is essential for remote sensing, biomedical diagnostics, secure optical communication, and intelligent vision [1, 2, 3, 4, 5, 6]. For highly integrated optoelectronic systems, miniaturized photodetectors that distinguish polarization states directly on chip, without external polarizers, offer an ideal route to polarimetric sensing [7, 8, 9, 10]. A central figure of merit for these devices is the photocurrent linear dichroism ratio (LDR), which quantifies the electrical contrast between orthogonal linear polarization states and therefore determines the attainable polarization sensitivity [11, 12].

Intrinsic optical anisotropy provides the most direct material‐level basis for filterless polarization detection [13]. Low‐symmetry inorganic crystals, including black phosphorus [14]. ReS₂ [15], and GeSe [16], as well as organic crystals such as 2,6‐diphenylanthracene [17] and thieno[3,2‐b]thiophene derivatives [18], can generate polarization‐dependent photocurrent. However, their absorption anisotropy is usually modest, with absorption LDR values generally limited to 1–5 (Table S1) [12, 18]. Because conventional photodetectors largely convert polarization‐dependent absorption into proportional photocurrent contrast, this weak optical anisotropy directly limits the device‐level LDR. As a result, most intrinsic‐anisotropy‐based photodetectors remain far below the extinction ratios of commercial polarizers, which can reach 10³–10⁴ in the visible range.

To enhance polarization discrimination beyond this absorption‐limited response, previous studies have explored heterojunctions [19, 20, 21, 22, 23], aligned nanowire crystals [24, 25, 26, 27], photovoltage engineering [28, 29, 30, 31, 32], circuit‐level designs [33, 34, 35], and metasurfaces [36, 37, 38], to enhance polarization‐dependent absorption or carrier collection. These strategies can increase photocurrent LDR values into the 10¹–10³ range (Table S2), but they often rely on increased structural complexity, more demanding fabrication, or architecture‐specific optimization. More fundamentally, most of these strategies rely on static amplification of optical anisotropy. By contrast, a design principle that converts weak intrinsic anisotropy into giant photocurrent dichroism through temporal carrier dynamics remains largely unexplored.

Here, we introduce polarization‐sensitive carrier dynamics (PSCD), a design principle that shifts polarization discrimination from static optical absorption to the time evolution of photocarrier populations. In this framework, polarization‐dependent photogeneration modulates optical‐erasure kinetics in a nonvolatile floating‐gate device, allowing weak intrinsic optical anisotropy to be dynamically amplified into large transient photocurrent contrast. We implement PSCD by integrating a molecularly thin p‐type organic single crystal with a two‐dimensional covalent organic framework (COF) floating gate. Orthogonal polarization states induce distinct optical‐erasure dynamics, leading to temporally divergent recovery of the channel hole population and photocurrent. Consequently, a crystal with an absorption LDR of only 3.4 produces a transient photocurrent LDR of 3.6 × 10⁴ under optimized bias conditions. This dynamic amplification mechanism further preserves polarization discrimination under partially polarized illumination and enables polarization‐resolved image reconstruction in optically complex scenes.

2. Results and Discussions

2.1. Polarization‐Sensitive Carrier Dynamics

The principle of PSCD is to shift polarization discrimination from static absorption contrast to the time‐dependent evolution of carrier populations. To implement this concept, we designed a polarization‐sensitive nonvolatile floating‐gate phototransistor in which a molecularly thin p‐type organic single crystal serves simultaneously as the charge‐transport channel and the light‐absorbing layer, while an insulating floating gate modulates the channel carrier concentration under electrical and optical stimuli (Figure 1a). Because the channel is molecularly thin, its carrier population is highly sensitive to charges stored in the floating gate.

FIGURE 1.

Structural design and physical model of the polarization‐sensitive carrier dynamics (PSCD). (a) Device architecture of the nonvolatile organic floating‐gate phototransistor used to implement PSCD. A molecularly thin p‐type organic single crystal functions as both the light‐absorbing layer and charge‐transport channel, while the floating gate stores charges that modulate the channel carrier population. (b) Operating scheme showing electrical programming from the low‐resistance state (LRS) to the high‐resistance state (HRS), followed by polarization‐dependent optical erasure. (c) Interfacial energy diagrams of the LRS and HRS, illustrating carrier‐population modulation during electrical programming and optical erasure. (d) Carrier‐dynamics model describing photogeneration, intraband recombination, and photoexcited electron transfer from the crystal to the floating gate under illumination. (e) Calculated time‐dependent evolution of the hole increment Δp in the HRS under orthogonal polarization states at 800 µW cm⁻² and 365 nm. (f) Calculated photocurrent LDR derived from panel (e) based on ideal hole drift under external electrical field, showing a transient amplification window generated by polarization‐dependent divergence of carrier dynamics.

The operating principle is illustrated in Figure 1b,c. A negative gate pulse injects electrons from the floating gate into the single crystal through tunneling, suppressing hole‐dominated channel conduction and programming the device from the low‐resistance state (LRS) to the high‐resistance state (HRS). Upon illumination, photoexcited electrons transfer back to the floating gate, optically erasing the programmed state and driving the channel toward the LRS. The key point is that the optical‐erasure rate depends on the incident polarization because the photogeneration rate in the anisotropic crystal is polarization dependent. Thus, a modest absorption difference can be converted into a large temporal difference in the recovery of the channel hole population.

To capture this process, we modeled the carrier dynamics under illumination (Figure 1d). Here, G and R represent the carrier photogeneration rate and the intraband recombination rate, respectively. R _e denotes the photoexcited electron‐transfer rate from the crystal to the floating gate. Given the robust memory retention of the device, the recombination rate R _h between trapped electrons in the floating gate and holes in the crystal is virtually negligible compared to R _e. The hole and electron concentration increments in the crystal, Δp and Δn, evolve with time according to:

$$\frac{\mathrm{d}\mathrm{\Delta }p}{\mathrm{d}t}=G-R=G-r\left(p_0+\mathrm{\Delta }p\right)\left(n_0+\mathrm{\Delta }n\right)$$ (1)

$$\frac{\mathrm{d}\mathrm{\Delta }n}{\mathrm{d}t}=G-r\left(p_0+\mathrm{\Delta }p\right)\left(n_0+\mathrm{\Delta }n\right)-R_{\mathrm{e}}$$ (2)

where t is the illumination time, r is the bimolecular recombination constant, p ₀ and n ₀ are the hole and electron concentrations before illumination, respectively. The photogeneration rate is given by:

$$G=\frac{{\eta }_{\mathrm{i}}AP}{T_{\mathrm{h}}{E}_{\mathrm{p}}}$$ (3)

where η_i is the internal quantum efficiency, A is the absorptance, P is the incident power density, T _h is the crystal thickness, and E _p is the photon energy. The cumulative transferred electron density is expressed as:

$$N_{\mathrm{t}}=N\left[1-\exp \left(\frac{-t}{\tau }\right)\right]$$ (4)

where N is the storage capacity of the floating gate and τ is the effective acceptor lifetime under continuous illumination, and N  =  βGτ, where β represents the probability that a photogenerated electron in the crystal is successfully captured by an acceptor site in the floating gate. In this framework, Δp − Δn  =  N _t and R _e =  dN _t/dt.

The single crystal exhibits intrinsic absorption anisotropy, thus the incident polarization angle directly modulates G and thereby drives polarization‐dependent carrier dynamics. The initial resistance state of the device before illumination, determined by p ₀ and n ₀, strongly affects this temporal evolution. To examine this effect, we calculated the time‐dependent Δp under polarized illumination for both the LRS and HRS (Figures 1e and S1a). In the calculation, the single‐crystal absorbance at 0° and 90° polarization was set to 0.6 and 0.2, respectively, consistent with an experimental absorption LDR of about 3 (see details in Supporting Information Section S1.4).

Because electron mobility in typical p‐type organic semiconductors is negligible, the photocurrent LDR under bias is primarily governed by the LDR of Δp in the ideal limit of hole‐dominated drift. For the LRS device, the LDR of Δp remains low and do not exceed the intrinsic absorption contrast (Figure S1b). By contrast, the HRS device exhibits a much larger time‐dependent LDR, with the model yielding a peak above 10⁸ (Figure 1f). This large dynamic range arises from the nonlinear temporal evolution of Δp. At early illumination times, the abundant electrons in the crystal promotes rapid intraband recombination and suppresses the increase in Δp. As illumination proceeds, photoexcited electron transfer to the floating gate depletes electrons in the crystal, reduces recombination, and enables rapid accumulation of holes. Subsequently, Δp saturates as the floating‐gate acceptor sites become filled. Under 0° polarized light, which generates carriers more efficiently, Δp enters this rapid‐accumulation regime earlier than under 90° polarized light. This polarization‐dependent temporal separation provides the basis for amplifying modest absorption anisotropy into a large transient device‐level LDR.

2.2. Materials Synthesis and Interfacial Energetics

To experimentally implement PSCD, we constructed a van der Waals heterostructure consisting of an AB‐DB COF floating gate and a p‐type 7‐decyl‐2‐phenyl[1]benzothieno[3,2‐b][1]benzothiophene (Ph‐BTBT‐C10) single‐crystal channel. The AB‐DB COF film was synthesized at the air–water interface through an acid‐catalyzed Schiff‐base reaction between 1,3,5‐tri(4‐aminophenyl)benzene (AB) and 2,5‐dihydroxybenzaldehyde (DB) (Figure 2a,b). This interfacial synthesis route produced large‐area, highly crystalline, and atomically flat films, which are suitable for forming a uniform tunneling interface with a low density of interfacial defects. FTIR spectroscopy confirms the formation of imine linkages characteristic of the AB‐DB COF (Figure 2c), while optical microscopy, AFM, and XRD further verify the film morphology and crystallinity (Figure S2).

FIGURE 2.

Heterostructure characterization, interfacial energetics, and nonvolatile memory behavior. (a) Chemical structures of AB, DB, and the resulting AB‐DB COF. (b) Schematic illustration of interfacial COF synthesis at the air–water interface. (c) FTIR spectra confirming imine‐bond formation in the AB‐DB COF. (d) Polarized UV–vis spectra of the molecularly thin Ph‐BTBT‐C10 single crystal, showing an absorption dichroism ratio of 3.4 at 378.5 nm. (e) In situ KPFM surface‐potential maps before programming, after a gate pulse, and after optical erasure, showing reversible charge modulation at the heterointerface. (f) Transfer curves of the initial device and the programmed HRS device. (g) Threshold‐voltage recovery under continuous 365 nm illumination. (h) Nonvolatile resistive switching at V _DS = −1 V. The HRS device is obtained by a gate pulse (V _GS = −60 V, 1 s, V _DS = 0 V), and the LRS device is obtained from the illuminated HRS device (365 nm, 520 µW cm⁻², 2 s).

For the transport and light‐absorbing layer, we selected molecularly thin Ph‐BTBT‐C10 single crystals because they combine efficient hole transport with intrinsic optical anisotropy. Their morphological and structural characterizations are summarized in Figure S3. Polarized UV–vis spectra show an absorption dichroic ratio of 3.4 at 378.5 nm (Figure 2d), providing the polarization‐dependent photogeneration input required for PSCD. This moderate absorption anisotropy also represents the intrinsic contrast bottleneck that typically limits conventional organic polarization photodetectors.

The energy‐level alignment between Ph‐BTBT‐C10 and the AB‐DB COF is consistent with reversible photo‐assisted charge transfer across the heterointerface. Cyclic voltammetry and UV–vis analysis place the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of Ph‐BTBT‐C10 at −5.65 and −2.51 eV, respectively, while the AB‐DB COF exhibits a HOMO of −4.92 eV and a LUMO of −2.57 eV (Figure S4). Kelvin probe force microscopy (KPFM) gives Fermi levels of −4.88 eV for Ph‐BTBT‐C10 and −4.62 eV for the COF (Figure S5). Together, these values indicate a staggered energy‐level configuration that can support the bidirectional tunneling processes required for electrical programming and optical erasure (Figure S6) [39, 40, 41, 42].

In situ KPFM mapping provides spatial evidence for reversible interfacial charge transfer [43, 44, 45]. Devices without the COF show only a negligible surface‐potential change after a gate pulse of V _GS = −60 V for 1 s at V _DS = 0 V (Figures 2e and S7). In contrast, devices containing the COF exhibit a pronounced shift in the Ph‐BTBT‐C10 surface potential from 0.137 V to above 10 V under the same programming condition. Subsequent 365 nm illumination at 520 µW cm⁻² for 2 s reduces the potential to −0.071 V, consistent with optical erasure through reverse electron transfer. Control devices containing only the AB‐DB layer do not show intrinsic charge migration after pulsing (Figure S8), indicating that the observed charge‐storage behavior originates from the Ph‐BTBT‐C10/COF heterointerface rather than from the isolated COF.

Macroscopic electrical measurements further support reversible charge modulation under electrical pulsing and illumination. After electrical programming, the transfer curve shifts from an initial threshold voltage of 1 V to an HRS threshold voltage of −25 V (Figure 2f). Under continuous illumination, the threshold voltage gradually recovers toward positive values (Figure 2g), reflecting time‐dependent optical erasure of the programmed state. The device also exhibits stable nonvolatile retention (Figure 2h), whereas COF‐free control devices do not show corresponding memory behavior (Figure S9). These results establish the AB‐DB COF as an effective charge‐storage layer that enables the floating‐gate modulation required for PSCD.

2.3. Transient Photocurrent Dynamics

The transient photoresponse of the programmed HRS device differs markedly from that of a conventional phototransistor. Immediately after illumination, the current enters a plateau‐like regime rather than increasing monotonically (Figure 3a). This behavior is reproducible over repeated optical‐erasure and electrical‐reset cycles (Figure 3b), indicating that it is a stable feature of the programmed device state rather than an accidental instability. Measurements on COF‐only architectures further show that the observed photocurrent originates from the Ph‐BTBT‐C10 channel, because the in‐plane conductivity of the COF layer remains negligible (Figure 3c).

FIGURE 3.

Transient photocurrent dynamics. (a) Time‐resolved photoresponse of the initial device and the programmed HRS device under 365 nm illumination at 705 µW cm⁻², showing a plateau‐like transient response in the HRS. (b) Cycling stability of the HRS photoresponse under repeated optical‐erasure and electrical‐reset processes. (c) Control photoresponse measurements of COF‐only devices, confirming the negligible in‐plane contribution of the COF layer. (d) Bias dependence of the current showing weak sensitivity to drain bias during the early plateau regime. (e) Interfacial energy diagram during the plateau phase, illustrating a contact‐associated bidirectional capacitive‐discharge process. (f) Energy diagram during the rise phase, illustrating the transition to drift‐dominated hole transport under negative bias. The HRS device is obtained by a gate pulse (V _GS = −60 V, 1 s, V _DS = 0 V), and the same reset gate pulse is applied after each illumination cycle.

A key feature of the plateau regime is its weak dependence on the applied drain bias (Figure 3d). This behavior is inconsistent with a conventional through‐channel drift current dominated by hole transport, which should vary strongly with both bias magnitude and polarity. Instead, it indicates that the early‐stage photocurrent contains a dominant transient component associated with charge redistribution near the metal–semiconductor contacts. A physically consistent mechanism is illustrated in Figure 3e. In the HRS, photoexcited electrons in the crystal transfer into the floating gate. This transfer locally modifies the channel potential near the contact region. To maintain electrostatic equilibrium, holes are expelled from the space‐charge region into the external circuit, giving rise to a bidirectional discharge current. As the floating‐gate acceptor states are progressively filled, the electron‐transfer rate decreases and the discharge current decays with time. During this interval, the bulk hole population in the channel remains too low to sustain substantial field‐driven drift current; therefore, the measured photoresponse is dominated by the contact‐associated discharge process.

After longer illumination, the bulk hole concentration recovers sufficiently for conventional channel transport to re‐emerge. At this stage, the current becomes strongly voltage dependent, and the device transitions from the discharge‐dominated plateau to a drift‐dominated rise regime (Figure 3f). To test this interpretation, we used a dual‐ammeter configuration to record the source and drain currents simultaneously (Figure S10a). During the plateau phase, both terminals show negative currents regardless of bias polarity (Figure S10b,c), indicating that holes leave the device from both ends simultaneously. This behavior supports a synchronized capacitive‐discharge picture rather than a conventional channel‐current response. The resulting transition from a bias‐insensitive discharge regime to a bias‐dependent drift regime provides the transient transport basis for the dynamic amplification of polarization contrast discussed below.

2.4. Giant Linear Dichroism Enabled by Dynamic Amplification

Having clarified the transient transport process, we next examined how polarization‐dependent carrier dynamics are converted into electrical dichroism. Under 90° polarized light, corresponding to the weaker absorption axis, the lower photogeneration rate prolongs the discharge‐dominated plateau over a broad range of drain biases (Figure 4a). In contrast, under 0° polarized light, the higher photogeneration rate accelerates optical erasure, causing the current to leave the plateau earlier and enter the drift‐dominated regime more rapidly (Figure 4b). Thus, the two orthogonal polarization states do not merely generate different photocurrent amplitudes; they drive the device through distinct transient transport regimes at different rates. The calculated responsivity (R) and specific detectivity (D*) is about 8 mA/W and 3.2×10¹³ Jones according to photoresponse curve at V _DS = −10 V in Figure 4b.

FIGURE 4.

Dynamic amplification of intrinsic anisotropy into giant linear dichroism in the HRS device. (a and b) Time‐resolved photoresponse under (a) 90° and (b) 0° polarized illumination at different V _DS values, showing polarization‐dependent timing of the transition from the discharge‐dominated plateau to the drift‐dominated rise regime. (c) Time‐dependent LDR extracted from panels (a) and (b), showing a pronounced transient amplification peak. (d) Representative photoresponse under 0° and 90° illumination and the corresponding LDR evolution at V _DS = −10 V. (e) Reversible switching during repeated alternating 0°/90° illumination cycles. (f) Benchmark comparison between the PSCD device and previously reported linear‐polarization photodetectors that amplify the LDR of anisotropic materials [21, 22, 23, 33, 34, 46, 47]. The HRS was programmed by a gate pulse of V _GS = −60 V for 1 s at V _DS = 0 V before each photoresponse measurement. The illumination condition was 365 nm and 800 µW cm⁻².

The time‐dependent LDR is defined as $\mathrm{LDR}(t)=I_{0^{\circ }}/I_{{90}^{\circ }}$. As shown in Figure 4c, LDR(t) exhibits a pronounced transient peak that increases with |V _DS|. At V _DS = −10 V, the device reaches a maximum LDR of 3.6 × 10⁴. This peak occurs when the 0° photocurrent has largely transitioned into the voltage‐dependent drift regime, whereas the 90° photocurrent is only beginning this transition and remains mainly within the voltage‐independent discharge plateau (Figures 4d and S11). Therefore, the giant LDR does not originate from a dramatic increase in intrinsic absorption anisotropy. Instead, it emerges from the polarization‐dependent timing of the transition between discharge‐dominated and drift‐dominated transport regimes.

The polarization response is reversible over repeated alternating 0°/90° illumination cycles (Figure 4e). In comparison, memory‐free control devices without the floating gate show an LDR below 3 (Figure S12), close to the intrinsic optical anisotropy of the crystal. This contrast indicates that the giant LDR arises from floating‐gate‐enabled temporal amplification rather than from crystal anisotropy alone. We further benchmarked the PSCD architecture against recently reported linear‐polarization photodetectors that amplify the LDR of anisotropic materials (Figure 4f and Table S3). Most structure‐engineered or heterojunction‐based devices operate in the LDR range of 10¹−10³, whereas the PSCD device reaches 3.6 × 10⁴ under optimized transient conditions. This comparison highlights dynamic carrier modulation as an effective strategy for achieving high polarization contrast while retaining a comparatively simple device architecture.

2.5. Robust Polarization Imaging Under Partially Polarized Light

In practical polarimetric imaging, incident light is often only partially polarized because scattering, reflection, and propagation through complex media reduce the degree of polarization. Therefore, effective polarization discrimination requires maintaining contrast not only under fully polarized illumination but also at reduced degrees of linear polarization (DoLP). To evaluate this condition, we introduced a scattering medium into the optical path and reduced the DoLP from 1 to 0.3 (Figure 5a). Under partially polarized illumination, the memory‐free control device exhibits an LDR below 1.6 (Figure 5b), indicating that intrinsic anisotropy alone provides insufficient contrast at low DoLP. In contrast, the PSCD‐enabled device preserves strong polarization discrimination in the same regime. Even at DoLP ≈ 0.3, it maintains a LDR of 9.8 × 10² (Figure 5c), indicating that the dynamic amplification mechanism remains effective despite substantial polarization degradation. This retained contrast enables the device to resolve a “T” pattern under partially polarized illumination, whereas the control device fails to produce distinguishable polarization‐dependent image contrast (Figure S13).

FIGURE 5.

Low‐DoLP polarization discrimination and PSCD‐enabled image reconstruction. (a) Optical transmission calibration for fully polarized light and partially polarized light with DoLP ≈ 0.3. (b and c) Photoresponse of (b) the memory‐free control device and (c) the PSCD‐enabled device under partially polarized illumination at 365 nm and 703 µW cm⁻², showing the preservation of high LDR in the PSCD architecture. (d) High‐dynamic‐range scene used for imaging analysis, containing both glare and deep‐shadow regions. Conventional imaging outputs of the scene, including (i) original, (ii) intensity, (iii) DoLP, and (iv) AoLP maps. (e) Mapping from pixel‐resolved optical input to time‐dependent photocurrent based on experimentally measured device dynamics. (f) Time evolution of the reconstructed image obtained from the transient response. (g) Reconstructed image at t = 15 s. (h) CLAHE‐processed comparison image generated from the same scene. (i) Optical photograph of the scene under uniform indoor lighting.

To examine the sensing implications of the PSCD response, we simulated on‐chip imaging of a high‐dynamic‐range scene (Figure 5di) based on experimentally measured device dynamics. The scene contains both a deep‐shadow region (including the text “Tianjin”) and a glare region (including the text “University”). Conventional intensity imaging loses the text information in both regions (Figure 5dii), while polarization maps based on DoLP and angle of linear polarization (AoLP) provide only partial recovery (Figure 5diii, iv). We then mapped the pixel‐resolved optical input onto the measured transient photocurrent response of the device (Figure 5e; Supporting Information Section S2). By taking the time derivative of the simulated current, we reconstructed a dynamic image sequence (Figure 5f). At t = 15 s, the reconstructed image resolves both the texts “Tianjin” and “University,” as well as the feather texture in both the deep‐shadow and glare regions (Figure 5g). The reconstructed image quality is comparable to that obtained using conventional contrast‐limited adaptive histogram equalization (CLAHE) algorithm applied to the same scene (Figure 5h), while also recovering fine structural details under uniform indoor lighting (Figure 5i). Notably, this enhancement arises from the intrinsic transient response of the device rather than solely from post‐processing algorithms. These results indicate that PSCD can serve as a front‐end sensing mechanism for enhancing polarization contrast and dynamic range in polarimetric imaging under optically complex conditions.

3. Conclusion

We have introduced polarization‐sensitive carrier dynamics (PSCD) as a design principle for filterless polarization photodetection, in which polarization discrimination is shifted from an absorption‐limited regime to the time domain of carrier‐population evolution. By integrating a molecularly thin Ph‐BTBT‐C10 single crystal with an AB‐DB COF floating gate, we realize a device architecture in which polarization‐dependent photogeneration modulates optical‐erasure kinetics, leading to temporally divergent recovery of the channel hole population. As a result, an absorption LDR of 3.4 is dynamically amplified into a transient photocurrent LDR of 3.6 × 10⁴ under optimized bias conditions. This dynamic amplification mechanism remains effective under partially polarized illumination, with the device maintaining a high LDR of 9.8 × 10² at a low DoLP of 0.3. These results indicate that PSCD opens a route toward compact polarimetric sensors that combine architectural simplicity, strong contrast, and physically embedded front‐end image processing.

4. Methods

4.1. Growth of Ph‐BTBT‐C10 Single Crystals

Ph‐BTBT‐C10 two‐dimensional molecular crystals were grown by interfacial crystallization on liquid glycerol. Briefly, 30 mL of glycerol was added to a cleaned glass vessel, followed by gentle deposition of 80 µL of a Ph‐BTBT‐C10 toluene solution (1 mg mL⁻¹) onto the liquid surface. The vessel was sealed and held at 16°C and 30% relative humidity to allow slow solvent evaporation and formation of large‐area floating single crystals.

4.2. Synthesis of AB‐DB COF Films

AB‐DB COF films were prepared by Schiff‐base condensation at the air‐water interface. AB and DB were dissolved in a mixed solvent of 1,4‐dioxane, mesitylene, and chloroform (1:1:2 by volume) at concentrations of 1.76 and 1.27 mg mL⁻¹, respectively. After addition of acetic acid catalyst (20 µL mL⁻¹), 50 µL of the precursor solution was spread onto the surface of deionized water. The reaction proceeded under static conditions at 16°C for 24 h to produce continuous COF films.

4.3. Device Fabrication

Highly doped n‐Si wafers with a 300 nm thermally grown SiO₂ layer were used as global back‐gate substrates. The oxide surface was modified with OTS to form a hydrophobic self‐assembled layer. Device assembly followed a layer‐by‐layer transfer sequence. First, the OTS‐treated substrate was brought into contact with the floating AB‐DB COF film, after which the substrate was rinsed with deionized water, acetone, and isopropanol and dried with N₂. Second, the COF‐coated substrate was used to adhere the floating Ph‐BTBT‐C10 crystal. Third, Au source and drain electrodes (180 nm) were thermally evaporated through a shadow mask onto a sacrificial OTS‐treated wafer and then mechanically transferred onto the crystal with a tungsten‐probe micromanipulator to avoid thermal damage.

4.4. Characterization and Measurements

Electrical characteristics were measured in the dark under ambient conditions using a Keithley 2636B SourceMeter. Dual‐channel transient currents were recorded using an Agilent B1500A semiconductor parameter analyzer. Polarization‐dependent optoelectronic measurements were carried out with linearly polarized light generated by passing a continuous‐wave laser beam (365 nm unless otherwise specified) through a Glan–Taylor prism and a half‐wave plate. For low‐DoLP measurements, a polyethylene terephthalate (PET) scattering diffuser was inserted into the optical path. KPFM measurements were performed on a Bruker Dimension Icon system. Structural and morphological characterization was carried out by XRD using a Rigaku SmartLab instrument and by polarized optical microscopy using a Nikon ECLIPSE Ci‐POL microscope.

Author Contributions

D. Zeng and S. Zhang fabricated the devices, performed the optoelectronic measurements, and conducted the simulations. P. Gao synthesized the COF films. X. Shen carried out part of the imaging experiments. R. Li designed the experiments, validated the results, supervised the project, and contributed to writing and revising the manuscript. W. Hu provided overall supervision and project administration.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma74211‐sup‐0001‐SuppMat.docx.

Data Availability Statement

All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supporting Information. Further data related to this work are available from the authors upon reasonable request.