APTMS in Perovskite Surface Passivation

• APTMS is a bifunctional organosilane featuring a primary amine and trimethoxysilane group used to passivate perovskite surfaces.
• It forms robust siloxane networks that reduce non-radiative recombination, suppress ion migration, and enhance photovoltaic performance.
• Precise deposition control is critical with APTMS to balance interfacial stabilization against overexposure and insulation effects.

Searching arXiv for papers on APTMS to ground the encyclopedia entry. (3-aminopyl)trimethysilane (APTMS) is a bifunction organosilane of formula $\mathrm{(CH_3O)_3Si(CH_2)_3NH_2}$, also written as $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$, that is widely used as a surface passivator and surface-modification reagent in halide perovskite optoelectronics and, more broadly, in silane-based interfacial chemistry (Shi et al., 2022, Huang et al., 3 Sep 2025). In the perovskite literature, its importance derives from the combination of a Lewis-basic primary amine, which can interact with undercoordinated $\mathrm{Pb^{2+}}$ and halide-vacancy-associated surface defects, and a trimethoxysilane headgroup, which can hydrolyze and condense into siloxane networks that stabilize an ultrathin interfacial layer (Jariwala et al., 2020, Pothoof et al., 2023). Across mixed-cation, mixed-halide perovskites, APTMS has been associated with reduced surface recombination velocity, increased photoluminescence lifetime and quantum yield, suppression of electrically and optically driven halide migration, and improved photovoltaic figures of merit when deposition conditions are appropriately controlled (Shi et al., 2022, Pothoof et al., 2023, Akrami et al., 2023, Huang et al., 3 Sep 2025).

1. Molecular identity and interfacial chemistry

APTMS is a small aminosilane whose defining functional motifs are a primary amine (–$\mathrm{NH_2}$) and a trimethoxysilane (–$\mathrm{Si(OCH_3)_3}$) headgroup (Akrami et al., 2023, Huang et al., 3 Sep 2025). In perovskite surface passivation, the amine is treated as a Lewis base that can coordinate to undercoordinated $\mathrm{Pb^{2+}}$ centers and mitigate halide-vacancy-related surface states, while the trimethoxysilane can hydrolyze to silanols and condense into $\mathrm{Si\!-\!O\!-\!Si}$ linkages, producing a polymeric or crosslinked siloxane layer (Shi et al., 2022, Jariwala et al., 2020).

The hydrolysis and condensation chemistry is central to APTMS behavior. One formulation given for trialkoxysilanes is

$$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$

followed by condensation with a hydroxylated surface:

$$\mathrm{(HO)_3SiR + Surface{-}OH \rightarrow Surface{-}O{-}SiR + H_2O}$$

(Huang et al., 3 Sep 2025). For APTMS specifically, the same chemistry is written in the context of silicon oxide functionalization as hydrolysis to trisilanols, condensation to surface silanols, and lateral $\mathrm{Si\!-\!O\!-\!Si}$ network formation (Kosovari et al., 2024). This chemistry explains both the robustness of APTMS-derived interlayers and the risk of uncontrolled oligomerization or multilayer growth if hydrolysis and condensation are not tightly controlled (Kosovari et al., 2024).

In mixed-halide perovskites, mechanistic assignments are supported indirectly and, in some cases, spectroscopically. High-resolution XPS in $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$0 shows that the N 1s bonding environment changes after APTMS treatment, with an increased fraction of high-binding-energy N 1s species, assigned to amine species engaged in Lewis base interactions, H-bond accepting, or protonated amine states (Jariwala et al., 2020). Angle-resolved XPS further indicates an upright APTMS orientation with terminal amines pointing toward the perovskite surface and methoxy-derived groups orienting away from it (Jariwala et al., 2020). This supports an amine-dominated surface interaction rather than a bulk structural modification.

2. Passivation of non-radiative recombination

APTMS is most extensively documented in perovskite photovoltaics as a passivator that reduces surface recombination velocity (SRV) and suppresses non-radiative recombination (Jariwala et al., 2020, Shi et al., 2022). In $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$1, APTMS treatment transforms the photophysics of thin films: average minority-carrier lifetime increases from 117 ns to 1.75 $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$2s, champion lifetime increases from 0.6 $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$3s to 4.3 $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$4s, and average external PLQE rises from 0.23% to 13.64%, with a champion value of 20.1% at 70 mW/cm$\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$5 532 nm CW excitation (Jariwala et al., 2020). Using the Ross relation and the Shockley–Queisser quasi-Fermi-level-splitting limit for $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$6 eV, the measured $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$7 corresponds to $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$8 eV, or about 97% of the theoretical limit (Jariwala et al., 2020).

The SRV reduction reported for the same composition is particularly large: extracted average SRV decreases from $\mathrm{NH_2\!-\!(CH_2)_3\!-\!Si(OCH_3)_3}$9 cm/s to $\mathrm{Pb^{2+}}$0 cm/s after APTMS passivation, with a champion of $\mathrm{Pb^{2+}}$1 cm/s (Jariwala et al., 2020). In that work, the interpretation is that surface-mediated recombination is the primary non-radiative loss pathway in MA-free mixed-cation mixed-halide films, and that APTMS suppresses it by passivating surface traps associated with undercoordinated $\mathrm{Pb^{2+}}$2 and related defects (Jariwala et al., 2020). The carrier-recombination framework is expressed using

$\mathrm{Pb^{2+}}$3

with APTMS primarily reducing the Shockley–Read–Hall term $\mathrm{Pb^{2+}}$4 through surface-trap passivation (Shi et al., 2022).

In device-relevant interfaces, APTMS also acts as a decoupling layer between the perovskite and $\mathrm{Pb^{2+}}$5 electron-transport layer. In a p–i–n stack based on $\mathrm{Pb^{2+}}$6, APTMS at the perovskite/$\mathrm{Pb^{2+}}$7 interface yields $\mathrm{Pb^{2+}}$8 cm s$\mathrm{Pb^{2+}}$9 in partial stacks, increases quasi-Fermi level splitting by about 100 meV, raises average $\mathrm{NH_2}$0 from 1.03 V to 1.09 V, improves average fill factor from 72.83% to 77.78%, and increases average PCE from 15.90% to 18.03% (Shi et al., 2022). In that study, the passivation layer remains effective after chlorobenzene washing, 100 $\mathrm{NH_2}$1C annealing, and high vacuum, consistent with the formation of a durable siloxane network rather than a labile adsorbate (Shi et al., 2022).

A related observation is that APTMS does not induce low-dimensional perovskite phases in the reported systems. XRD, UV–vis, and steady-state PL show no formation of low-dimensional phases in the cited device study, indicating that the effect is interfacial passivation rather than surface phase reconstruction (Shi et al., 2022).

3. Suppression of ion migration and phase segregation

A second major function of APTMS in wide-bandgap mixed-halide perovskites is suppression of vacancy-mediated halide migration under electrical bias and illumination (Pothoof et al., 2023, Akrami et al., 2023). In $\mathrm{NH_2}$2, scanning Kelvin probe microscopy (SKPM) after local voltage poling shows that control films exhibit a CPD shift near $\mathrm{NH_2}$3 mV at poling extremes of $\mathrm{NH_2}$4 V after a dwell time of only a few seconds, whereas APTMS-treated films show a five-fold reduction to $\mathrm{NH_2}$5 mV (Pothoof et al., 2023). The same study reports that the full-width at half-maximum of the CPD-shift distribution broadens strongly in unpassivated films under negative bias, while APTMS-treated films maintain a narrow FWHM of $\mathrm{NH_2}$6 mV across biases (Pothoof et al., 2023).

The SKPM observable is tied to the usual contact-potential relation

$\mathrm{NH_2}$7

which is used as a proxy for field-driven ionic charge accumulation at the surface (Pothoof et al., 2023, Akrami et al., 2023). In the same context, vacancy-mediated ionic drift is described by

$\mathrm{NH_2}$8

with APTMS passivation interpreted to reduce the mobile defect density $\mathrm{NH_2}$9 and possibly the effective mobility $\mathrm{Si(OCH_3)_3}$0 at surfaces and interfaces (Pothoof et al., 2023).

Under purely optical driving, APTMS also slows photoinduced halide segregation. Hyperspectral PL microscopy of $\mathrm{Si(OCH_3)_3}$1 under $\mathrm{Si(OCH_3)_3}$2 mW/cm$\mathrm{Si(OCH_3)_3}$3 532 nm illumination in dry $\mathrm{Si(OCH_3)_3}$4 shows that control films develop a growing iodide-rich emission feature near $\mathrm{Si(OCH_3)_3}$5 nm alongside the mixed-phase peak near $\mathrm{Si(OCH_3)_3}$6 nm, whereas APTMS-treated films display hindered growth of iodide-rich domains and higher overall PL intensity (Pothoof et al., 2023). Pixel-wise spectral binning using a 765 nm threshold shows that the fraction of iodide-rich pixels increases by $\mathrm{Si(OCH_3)_3}$7 relative to the initial state in control films under prolonged light soaking, but only $\mathrm{Si(OCH_3)_3}$8 in APTMS-passivated films (Pothoof et al., 2023).

A complementary kinetic study on $\mathrm{Si(OCH_3)_3}$9 reports that APTMS-passivated films exhibit a smaller CPD shift under 405 nm illumination and a slower surface-potential relaxation in the dark (Akrami et al., 2023). Fitting the dark relaxation with the stretched exponential

$\mathrm{Pb^{2+}}$0

gives $\mathrm{Pb^{2+}}$1 s for unpassivated films and $\mathrm{Pb^{2+}}$2 s for APTMS-passivated films, a $\mathrm{Pb^{2+}}$3 slower relaxation consistent with hindered halide migration (Akrami et al., 2023). In PL measurements on the same system, continuous 532 nm illumination causes a progressive redshift in unpassivated films that APTMS slows by more than $\mathrm{Pb^{2+}}$4 at matched intensities (Akrami et al., 2023).

The fluence dependence is treated explicitly in that work. When PL shift is plotted against light fluence $\mathrm{Pb^{2+}}$5, with

$\mathrm{Pb^{2+}}$6

the same fluence yields a similar extent of redshift regardless of whether it arises from higher intensity or longer time, which the authors interpret as first-order kinetics with respect to fluence (Akrami et al., 2023). To exclude A-site cation segregation as the dominant mechanism, the same study examined $\mathrm{Pb^{2+}}$7 and observed no significant PL peak shift under illumination, supporting halide migration as the operative process in the mixed-halide case (Akrami et al., 2023).

4. Spatial heterogeneity and interface-specific effects

APTMS modifies not only the magnitude but also the spatial distribution of ionic and electronic responses. In the SKPM study of $\mathrm{Pb^{2+}}$8, the authors distinguish “domains” and “interfaces” operationally, while cautioning that topographic features are not necessarily grain boundaries (Pothoof et al., 2023). Control films show larger CPD shifts near domain interfaces than in domain centers at $\mathrm{Pb^{2+}}$9 V, whereas APTMS-treated films show a more uniform CPD as a function of distance from interfaces (Pothoof et al., 2023). The difference between average CPD at domain centers and domain interfaces exhibits a negative linear trend with poling bias in control films but little change in APTMS-passivated films (Pothoof et al., 2023).

This spatial homogenization is consistent with a model in which interface-related defects are preferential sites for ionic motion and recombination, and APTMS preferentially passivates those sites (Pothoof et al., 2023). A plausible implication is that the beneficial effect of APTMS is especially strong in the surface/interface network that dominates early-stage bias-induced redistribution.

Morphological and conductive-probe measurements in device stacks reinforce this picture. In the interfacial-device study, conductive AFM shows that APTMS coverage is nonuniform: thicker APTMS regions suppress local conductance, while thinner regions retain contact (Shi et al., 2022). After $\mathrm{Si\!-\!O\!-\!Si}$0 deposition, conductivity becomes spatially uniform, suggesting that the ETL homogenizes extraction despite reduced local contact fraction (Shi et al., 2022). AFM mechanical mapping identifies low-modulus regions attributed to polymerized APTMS, with modulus around 10 MPa for hazy regions compared with GPa-range values for exposed perovskite (Shi et al., 2022). This supports a heterogeneous but functionally effective passivation landscape in which a soft siloxane network coexists with nanoscale extraction pathways.

5. Deposition strategies, process windows, and overexposure

APTMS has been implemented by both vapor-phase and solution-based routes, and its performance depends strongly on dose, residence time, and the extent of hydrolysis/condensation (Jariwala et al., 2020, Shi et al., 2022, Huang et al., 3 Sep 2025). In one widely cited vapor-passivation procedure, perovskite films are exposed to APTMS vapor at room temperature in a vacuum oven for 5–10 minutes, using 1 mL of APTMS in a vial under a confined glass jar, followed by an anhydrous chlorobenzene or toluene rinse (Jariwala et al., 2020). A solution-processable variant based on spin-coating 5 v% APTMS in chlorobenzene also increases PL lifetime, whereas chlorobenzene alone does not, indicating that APTMS is the active passivating agent (Jariwala et al., 2020).

A later device-optimization study emphasizes that the process window can be narrow for FA-containing perovskites. In $\mathrm{Si\!-\!O\!-\!Si}$1, room-temperature vacuum deposition at about $\mathrm{Si\!-\!O\!-\!Si}$2 inches Hg relative to atmosphere was investigated for 30 s, 90 s, and 360 s (Huang et al., 3 Sep 2025). On films, APTMS increases TRPL lifetime from 30 s to 90 s treatment, but the 360 s treatment does not further increase lifetime relative to 90 s (Huang et al., 3 Sep 2025). UV–Vis absorbance declines progressively with increasing APTMS deposition time, and XRD peak intensities decrease at higher treatment times, with a clear reduction at 900 s (Huang et al., 3 Sep 2025). In devices, performance is optimized at short treatment durations of about 30 s: $\mathrm{Si\!-\!O\!-\!Si}$3 and FF improve relative to unpassivated controls, but with longer deposition times, $\mathrm{Si\!-\!O\!-\!Si}$4, FF, and PCE decline even though $\mathrm{Si\!-\!O\!-\!Si}$5 often continues to increase (Huang et al., 3 Sep 2025).

The interpretation offered in that work is twofold. First, overly long APTMS exposures can cause subtle perovskite decomposition or structural disruption, consistent with the absorbance and XRD losses. Second, amino-silanes are insulating, so thicker coatings increasingly hinder charge extraction (Huang et al., 3 Sep 2025). This is consistent with the earlier caution that thick APTMS layers would be insulating in typical n–i–p or p–i–n stacks and should therefore remain ultrathin and, where appropriate, patterned or confined to passivating regions (Jariwala et al., 2020).

The same dose sensitivity appears in comparison with [3-(2-aminoethylamino)propyl]trimethoxysilane (AEAPTMS). Under optimized short treatments, both APTMS and AEAPTMS improve $\mathrm{Si\!-\!O\!-\!Si}$6 and FF, but AEAPTMS shows a wider, more robust processing window and higher overall PCE (Huang et al., 3 Sep 2025). APTMS is more strongly affected by overexposure, with more pronounced absorbance loss and lifetime reduction at long deposition times (Huang et al., 3 Sep 2025).

6. Reactivity, generality, and limitations

APTMS is often described as a passivator that acts primarily through amine–$\mathrm{Si\!-\!O\!-\!Si}$7 coordination and siloxane-based interfacial stabilization, but more recent work qualifies that picture by identifying possible reactivity with formamidinium-derived species (Huang et al., 3 Sep 2025). In solution-state NMR of APTMS with FAI in DMSO-$\mathrm{Si\!-\!O\!-\!Si}$8 at 0.04 M, the reaction appears about 50% complete after 5 min and then stalls over many hours (Huang et al., 3 Sep 2025). Proposed products include N-(3-(trimethoxysilyl)propyl)formamidinium plus ammonia, followed by a proton-transfer pathway yielding N-(3-(trimethoxysilyl)propyl)formamidine and a 3-(trimethoxysilyl)propylammonium ion; the formamidine can be stabilized by coordination of the deprotonated nitrogen to silicon to form a five-membered ring adduct (Huang et al., 3 Sep 2025). The reaction proceeds to completion in methanol-$\mathrm{Si\!-\!O\!-\!Si}$9 or in the presence of DBU (Huang et al., 3 Sep 2025).

However, the same study finds limited detectable APTMS–FA$$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$0 reaction on perovskite films during vacuum deposition. ToF-SIMS shows no clear evidence of N-(3-(trimethoxysilyl)propyl)formamidinium or N,N′-bis(3-(trimethoxysilyl)propyl)formamidinium on films, and no significant change in the FA$$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$1 signal with deposition time (Huang et al., 3 Sep 2025). This contrasts with AEAPTMS, which produces a strong increase in a 4,5-dihydroimidazolium fragment and a decrease in the FA$$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$2 signal (Huang et al., 3 Sep 2025). The implication drawn is that APTMS passivates mainly through defect coordination and surface bonding rather than extensive FA-reactive surface chemistry under the studied vacuum conditions.

The generality of APTMS passivation across perovskite compositions is supported but not unlimited. Significant PL and PL-lifetime improvements have been reported not only for $$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$3 ($$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$4 eV), but also for higher-bandgap FA/Cs compositions near 1.7 eV, 1.75 eV, and 1.81 eV, as well as a triple-cation FAMACs/I–Br composition (Jariwala et al., 2020). Another study shows increased QFLS to about 1.25 eV in $$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$5, indicating relevance for tandem architectures (Shi et al., 2022). At the same time, several limitations are explicitly noted across the cited literature: film thickness, solvent system, atmosphere, humidity, and detailed long-term stability metrics are often not fully specified in the main text (Akrami et al., 2023); SKPM studies do not always report relaxation times or exact tip/poling details beyond “a few seconds” (Pothoof et al., 2023); and direct spectroscopic identification of specific APTMS–surface bonds is not always provided, with mechanistic assignment relying partly on consistency between PL, SKPM, XPS, and prior chemistry (Pothoof et al., 2023, Shi et al., 2022).

Outside perovskites, the broader silanization literature also highlights a general risk intrinsic to free-amine trialkoxysilanes: uncontrolled hydrolysis/condensation, crosslinking, and multilayer formation can reduce monolayer quality and amine accessibility if water content and exposure are not tightly managed (Kosovari et al., 2024). Although APTMS was not empirically studied in that silicon-wafer work, the argument made there is that, because methoxy groups are more hydrolytically labile than ethoxy groups, APTMS is expected to be at least as prone as APTES and likely more prone to uncontrolled hydrolysis/condensation, oligomerization, and multilayer formation under poorly controlled conditions (Kosovari et al., 2024). This suggests that the process sensitivity observed in perovskites is chemically plausible and not solely device-specific.

7. Significance for optoelectronics

APTMS occupies a distinct position among perovskite passivators because it combines chemical defect passivation with interfacial robustness. In the most favorable reports, it drives perovskite thin films close to radiative limits, with external PLQE above 20%, quasi-Fermi-level splitting near the Shockley–Queisser limit, and SRV reduced to values more commonly associated with highly passivated inorganic semiconductors (Jariwala et al., 2020). In device stacks, it reduces surface recombination velocity and decouples the absorber from detrimental interactions at the $$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$6 interface, enabling higher $$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$7, higher fill factor, and higher PCE (Shi et al., 2022). In wide-bandgap mixed-halide compositions relevant to tandem top cells, it also suppresses bias-induced and photoinduced halide migration, mitigates phase segregation, and produces more homogeneous local electronic landscapes (Pothoof et al., 2023, Akrami et al., 2023).

The principal caveat is that APTMS is not merely a benign molecular ligand. It is a reactive silane whose benefit depends on balancing ultrathin passivation against the risks of overexposure, excessive insulation, and possible chemical disruption of FA-containing perovskites (Huang et al., 3 Sep 2025). The comparative data with AEAPTMS show that not all amino-silanes share the same process latitude, and that APTMS can have a narrower optimum than structurally related alternatives (Huang et al., 3 Sep 2025).

Taken together, the literature portrays APTMS as an effective but process-sensitive passivator for mixed-halide perovskites. Its amine functionality is associated with passivation of undercoordinated $$\mathrm{(RO)_3SiR + H_2O \rightarrow (HO)_3SiR + 3\,ROH}$$8 and halide-vacancy-rich defect sites, while its trimethoxysilane functionality enables formation of a robust, surface-confined siloxane network (Jariwala et al., 2020, Pothoof et al., 2023). Where deposition is optimized, this combination suppresses non-radiative recombination, slows ion migration, reduces halide segregation, and improves photovoltaic performance; where deposition is excessive, the same chemistry can become detrimental through insulation, deeper penetration, or structural disruption (Shi et al., 2022, Huang et al., 3 Sep 2025).