Spiro-OMeTAD: Benchmark HTL for Perovskite Cells

• Spiro-OMeTAD is a small-molecule hole-transport material that enables selective hole extraction and transport in perovskite solar cells.
• Interface engineering with doping and additives optimizes energy-level alignment and enhances charge extraction and device efficiency.
• Research indicates that while Spiro-OMeTAD achieves high efficiency, its performance is limited by thermal degradation, humidity sensitivity, and additive volatility.

Spiro-OMeTAD, chemically 2,2',7,7'-tetrakis(N,N-di-p-methoxyphenylamine)-9,9'-spirobifluorene, is the benchmark organic hole-transport layer in regular $n$–$i$–$p$ perovskite solar cells. Its canonical functions are selective hole extraction from the perovskite absorber, hole transport to the metal electrode, and electron blocking; in this role it has remained central to high-efficiency device design, even as multiple studies have shown that the perovskite/Spiro-OMeTAD contact can dominate non-radiative recombination, stochastic transport bottlenecks, and thermally or ionically driven degradation (Paul et al., 17 Sep 2025, Stolterfoht et al., 2018).

1. Molecular identity, electronic role, and device placement

In the reviewed literature, Spiro-OMeTAD is consistently treated as the archetypal small-molecule hole-transport material for regular perovskite cells. In standard $n$–$i$–$p$ stacks it is placed between the perovskite absorber and a metal contact such as Au or Ag, as in FTO/TiO$_2$/MAPbI$_3$/Spiro-OMeTAD/Au, ITO/SnO$_2$/perovskite/Spiro-OMeTAD/Ag, or IO:H/SnO$_2$/perovskite/PEAI/Spiro-OMeTAD/Au (Davenport et al., 2020, Wu et al., 2 Jul 2025, Siegrist et al., 2021). A more elaborate variant is the double-hole-layer stack MAPbI$i$0/CuO/Spiro-OMeTAD/Au, where CuO is placed adjacent to the perovskite and Spiro-OMeTAD is used above CuO to align energy levels to Au and improve selectivity (Rahman et al., 3 Sep 2025).

Electronic descriptors reported for Spiro-OMeTAD depend on context and temperature. A thermal-stability review gives a HOMO of $i$1 eV at $i$2C and $i$3 eV at $i$4C, with the perovskite valence-band maximum in the same analysis at $i$5 eV; the corresponding hole-extraction offset $i$6 therefore increases from approximately $i$7 to $i$8 eV under heating (Paul et al., 17 Sep 2025). Typical hole mobilities cited there are $i$9–$p$0, while an ultrafast spectroscopy study notes that the Spiro-OMeTAD hole mobility is $p$1, sufficiently low that its direct THz photoconductivity contribution is negligible (Butler-Caddle et al., 2024).

The material’s continued prominence is not simply a consequence of historical use. A genetic-algorithm study that balanced stability, efficiency, and cost selected TiO$p$2/CH$p$3NH$p$4PbI$p$5Br$p$6/Spiro-OMeTAD as the most well-balanced stack under equal weighting, despite assigning Spiro-OMeTAD a normalized stability index of $p$7 and listing its price as 507 USD/g (Kim et al., 2020). This suggests that, within current perovskite design spaces, its efficiency contribution can outweigh known penalties in stability and cost when the objective is not dominated by either of those criteria alone.

2. Interfacial energetics and selectivity

A central theme in the modern literature is that Spiro-OMeTAD performance is controlled less by its bulk identity than by the energetic and defect landscape of the perovskite/HTL interface. Absolute photoluminescence measurements on perovskite/CTL heterojunctions showed that the perovskite/spiro-OMeTAD interface lowers the quasi-Fermi-level splitting relative to the neat perovskite: for perovskite/spiro-OMeTAD, the reported values are $p$8, $p$9 eV, and $n$0, whereas the neat perovskite yields $n$1 eV and $n$2 (Stolterfoht et al., 2018). In $n$3–$n$4–$n$5 cells with TiO$n$6 or SnO$n$7 ETLs, the full-stack QFLS is reported as approximately $n$8–$n$9 eV with $i$0 V, and the least-selective interface is identified as the perovskite/Spiro-OMeTAD junction (Stolterfoht et al., 2018).

Several interface-engineering studies quantify how modifying the perovskite side can improve alignment to Spiro-OMeTAD. In CsPbI$i$1 devices, adding TOPO on an OAI-passivated surface reverses OAI-induced downward band bending of $i$2 meV to upward band bending of $i$3 meV, reduces the perovskite work function by $i$4 meV in UPS and $i$5 meV in Kelvin probe, and increases the hole-extraction rate constant to $i$6 for TOPO + Spiro-OMeTAD, compared with $i$7 for the control + Spiro-OMeTAD; the selectivity ratio $i$8 rises to $i$9 (Iqbal et al., 2023). In MAPbI$p$0 devices modified by ribavirin, UPS gives a perovskite valence-band maximum shift from $p$1 to $p$2 eV and a work-function shift from $p$3 to $p$4 eV, which reduces the estimated hole-extraction barrier to a Spiro-OMeTAD HOMO near $p$5 eV from $p$6 to $p$7 eV; the associated $p$8–$p$9–$_2$0 device improves from $_2$1 V, $_2$2, $_2$3, $_2$4 to $_2$5 V, $_2$6, $_2$7, and $_2$8 (Wu et al., 2 Jul 2025).

Band engineering of Spiro-OMeTAD itself has also been modeled explicitly. In a CuO/Spiro-OMeTAD double-HTL design, I$_2$9O$_3$0 doping changes the Spiro-OMeTAD parameters from $_3$1 eV and $_3$2 eV to approximately $_3$3 eV and $_3$4 eV, shifting the HOMO from $_3$5 to approximately $_3$6 eV and reducing the CuO/Spiro valence-band offset from about $_3$7 to about $_3$8 eV (Rahman et al., 3 Sep 2025). The same study reports a planar double-HTL device progressing from $_3$9 PCE in the initial configuration to $_2$0 after thickness, doping, and band-matching optimization, and to $_2$1 in a hierarchically patterned design (Rahman et al., 3 Sep 2025). Taken together, these results indicate that Spiro-OMeTAD often functions effectively only when the interface is made highly selective by band bending, passivation, or both.

3. Doping, oxidation, and operando electronic states

Spiro-OMeTAD is rarely used in pristine form in efficient devices. The common strategy is oxidative $_2$2-type doping with additives such as LiTFSI, $_2$3-tert-butylpyridine, and cobalt complexes. An operando ESR study used a formulation of 73.0 mg Spiro-OMeTAD in 1 mL chlorobenzene with LiTFSI 9.2 mg, $_2$4BP 28.8 $_2$5L, and FK102 8.7 mg, and estimated a doping level of $_2$6 per monomer in the radical-cation model or $_2$7 in the diradical-dication model; by comparison, LiTFSI alone in previous work was associated with $_2$8 or $_2$9 (Watanabe et al., 2020). In a separate QFLS study, Spiro-OMeTAD films and devices were exposed to air overnight before encapsulation because the absence of $_2$0-doping led to $_2$1 and negligible photovoltaic performance (Stolterfoht et al., 2018).

Operando ESR established that Spiro-OMeTAD is the sole room-temperature ESR-active layer in a simplified planar cell, with a dominant signal at $_2$2 assigned to Spiro-OMeTAD holes, i.e. radical cations or diradical dications (Watanabe et al., 2020). Under simulated solar irradiation, the spin number $_2$3 increased with irradiation time, reaching on the order of $_2$4 spins in the plotted data. That increase tracked device metrics in a non-monotonic way: immediately after light-on, $_2$5 decreased as $_2$6 rose; over roughly 2 h, $_2$7 then gradually increased while $_2$8 continued to rise; and $_2$9 decreased monotonically with increasing $i$00 (Watanabe et al., 2020). The paper attributes the early $i$01 decrease to charge-carrier scattering by accumulated holes in Spiro-OMeTAD, the later $i$02 recovery to filling of deep trapping levels in the amorphous HTM, and the $i$03 loss to interfacial electric dipole formation at the Spiro-OMeTAD/Au interface (Watanabe et al., 2020).

The same operando study identified an ultraviolet-specific de-doping pathway. Under full-spectrum AM 1.5 illumination, $i$04 after light-off fell below the pre-illumination level, whereas with wavelengths $i$05 nm removed it returned only to the initial baseline. This was interpreted as direct evidence for reverse electron transfer from TiO$i$06 to Spiro-OMeTAD, expressed as $i$07, thereby lowering the effective hole density and conductivity (Watanabe et al., 2020).

A SCAPS drift-diffusion study modeled the effects of Li-TFSI, FK209, and $i$08BP by changing only Spiro-OMeTAD parameters. Relative to an undoped baseline of $i$09, $i$10, $i$11, and $i$12 V, the Li-TFSI-doped case gave $i$13, $i$14, $i$15, $i$16 V; FK209 gave $i$17, $i$18, $i$19, $i$20 V; and $i$21BP gave $i$22, $i$23, $i$24, $i$25 V (Dizaj, 2024). These results are simulation-specific, but they reinforce the broader conclusion that Spiro-OMeTAD performance is inseparable from its oxidation state, trap occupation, and interfacial electrostatics.

4. Carrier extraction, noise, and recombination dynamics

Measurements of Spiro-OMeTAD-related kinetics span a wide dynamical range and have not converged to a single universal picture. Time-resolved photoluminescence quenching on MAPbI$i$26 films coated with Spiro-OMeTAD reported extremely rapid hole-transfer signatures: for a $i$27 nm film, the neat lifetime was $i$28 ns and the Spiro-OMeTAD-coated decay was $i$29 ns, implying a quenching efficiency of about $i$30; for a 95 nm film, the neat lifetime was 12.4 ns and the coated decay $i$31 ns, implying about $i$32 quenching (Li et al., 2015). Fits to a one-dimensional diffusion model yielded a hole diffusion coefficient $i$33 and $i$34m in the thick film, versus $i$35 and $i$36m in the 95 nm film (Li et al., 2015).

By contrast, ultrafast THz and optical studies on triple-cation perovskite/Spiro-OMeTAD bilayers found negligible decay of photoconductivity or bleach amplitude over 3 ns and estimated an interfacial surface-loss velocity of only $i$37; low-injection TRPL changed from $i$38 ns for bare perovskite to an effective lifetime of only about 100 ns in the Spiro bilayer (Butler-Caddle et al., 2024). That study concluded that Spiro-OMeTAD exhibits slow hole extraction but does not measurably increase the perovskite surface recombination rate (Butler-Caddle et al., 2024). The juxtaposition of this result with sub-nanosecond PL quenching in MAPbI$i$39 films suggests that the extracted “charge-transfer rate” depends strongly on the observable, interfacial chemistry, and bilayer morphology.

Low-frequency and cross-correlation noise spectroscopy sharpen the interfacial picture. In MAPbI$i$40/Spiro-OMeTAD cells measured after 30 min light-soaking under short circuit, the frequency-independent term obeyed $i$41 with $i$42 in the thicker device, which was interpreted as near full-scale shot noise from a single dominant resistive element, likely the perovskite/Spiro-OMeTAD interface (Davenport et al., 2020). The same work reported strong $i$43 noise with non-ideal current scaling $i$44–$i$45, attributed to current-induced halide migration and mixed bulk/interfacial defect modulation, as well as generation-recombination noise whose inverse relaxation time scaled linearly with photocurrent, $i$46, identifying bimolecular recombination in the perovskite bulk and yielding an extrapolated $i$47s (Davenport et al., 2020).

Correlated low-frequency noise and impedance spectroscopy on SnO$i$48/MAPbI$i$49/Spiro-OMeTAD/Au devices further separated high- and moderate-performance Spiro-OMeTAD cells. High-PCE Spiro devices showed normalized noise four orders of magnitude lower than PTAA cells and a cyclostationary peak near 200 Hz, whereas moderate-PCE Spiro devices showed a Lorentzian feature centered near $i$50–200 Hz and two inductive loops between 0.6 and 1.0 V (Sangwan et al., 2019). The shared $i$51 Hz scale was linked to a crossover from electrode to dielectric polarization, and thus to interfacial ionic relaxation rather than bulk chemical capacitance (Sangwan et al., 2019). A common misconception is therefore that Spiro-OMeTAD performance is determined mainly by its bulk conductivity; the combined noise, QFLS, and ultrafast literature instead points to interface-dominated selectivity and interfacial ionic dynamics as recurring limiting factors.

5. Processing routes, scalable deposition, and device implementations

Although Spiro-OMeTAD is often associated with small-area spin-coated devices, the recent literature shows that it is compatible with scalable deposition. A hybrid PVD/blade-coating process deposited Spiro-OMeTAD in ambient air from p-xylene on $i$52 substrates, using 25 mg Spiro-OMeTAD in 1 mL p-xylene with $i$53BP 10 $i$54L and LiTFSI stock 6 $i$55L; blade coating was performed at 90 mm/s, a 100 $i$56m gap, and $i$57C (Siegrist et al., 2021). In the resulting IO:H/SnO$i$58/perovskite/PEAI/Spiro-OMeTAD/Au architecture, the champion device showed $i$59 V, $i$60, $i$61, and $i$62, with best $i$63 up to 1.16 V and PCE up to $i$64 (Siegrist et al., 2021). These data establish that Spiro-OMeTAD can be integrated into green-solvent, large-area workflows, although the study did not report HTL thickness or direct conductivity measurements (Siegrist et al., 2021).

Spiro-OMeTAD also remains effective in architectures where the dominant gains come from absorber-side modification rather than HTL reformulation. In ribavirin-modified MAPbI$i$65 $i$66–$i$67–$i$68 cells with ITO/SnO$i$69/perovskite/Spiro-OMeTAD/Ag, the champion metrics improved from $i$70 V, $i$71, $i$72, $i$73 to $i$74 V, $i$75, $i$76, and $i$77 after ribavirin addition to the precursor (Wu et al., 2 Jul 2025). The paper attributes the gain to larger grains, lower roughness, reduced trap density, higher defect formation energies, and a perovskite energy-level shift toward better alignment with Spiro-OMeTAD (Wu et al., 2 Jul 2025).

Modeled extensions beyond lead-iodide systems also preserve a role for Spiro-OMeTAD. In SCAPS simulations of SLG/FTO/In$i$78S$i$79/CH$i$80NH$i$81SnI$i$82/Spiro-OMeTAD/Au, the optimized Spiro-OMeTAD-based device reached $i$83, $i$84 V, $i$85, and $i$86, slightly outperforming the CuSCN alternative at $i$87 (Alam et al., 2020). In that model, Spiro-OMeTAD was assigned $i$88, $i$89 eV, $i$90 eV, and a 300 nm thickness (Alam et al., 2020). This does not establish experimental superiority in tin perovskites, but it does show that Spiro-OMeTAD remains a reference point even in lead-free device optimization.

6. Thermal instability, ionic interactions, and the search for alternatives

The strongest criticisms of Spiro-OMeTAD concern stability rather than efficiency. A 2025 thermal-degradation review summarizes several convergent mechanisms: heat-induced morphology change, temperature-driven energy-level shifts, hygroscopicity associated with Li-TFSI, and volatility associated with $i$91BP (Paul et al., 17 Sep 2025). In the cited data, Spiro-OMeTAD develops substantial voids at $i$92–$i$93C, while at $i$94C its HOMO shifts from $i$95 to $i$96 eV, increasing the barrier to hole extraction from a perovskite VBM at $i$97 eV (Paul et al., 17 Sep 2025). Song et al. are summarized as showing that under $i$98C stress in $i$99 up to 1032 h, Li-TFSI + $p$00BP formulations exhibit extensive domain crystallization of Spiro-OMeTAD and micrometer-deep trenches under Au, whereas Li-TFSI-only films remain largely amorphous with smaller aggregates or voids (Paul et al., 17 Sep 2025).

Device-level stability metrics in that review are comparably unfavorable. For unsealed FA-based PSCs with Spiro-OMeTAD + additives, the PCE loss at $p$01C reaches $p$02 by 600 h; even sealed devices retain only $p$03 of the initial PCE after 600 h (Paul et al., 17 Sep 2025). Spiro-OMeTAD without additives performs somewhat better at $p$04C, but still falls to $p$05–$p$06 by $p$07 h, whereas dopant-free P3HT retains $p$08 of initial PCE after $p$09 h across $p$10C in sealed and unsealed devices (Paul et al., 17 Sep 2025). The same review therefore recommends limiting or removing $p$11BP, preheating Spiro-OMeTAD before Au deposition, and introducing barrier layers between Spiro-OMeTAD and Au (Paul et al., 17 Sep 2025).

Operational degradation mechanisms under illumination are likewise interfacial. Operando ESR demonstrated UV-driven reverse electron transfer from TiO$p$12 to Spiro-OMeTAD that neutralizes $p$13 and lowers the effective doping level (Watanabe et al., 2020). In CsPbI$p$14 cells, the TOPO-modified perovskite/Spiro interface was proposed to suppress transport of electrons and mobile iodide into the HTM; under cyclic maximum-power-point tracking in nitrogen, TOPO-treated cells produced $p$15 more total energy than controls, with a $p$16 surplus after 342 h, and retained $p$17 and $p$18 more robustly (Iqbal et al., 2023). These results indicate that “Spiro-OMeTAD instability” is not a single failure mode but a compound problem involving additive volatility, humidity sensitivity, heat-driven crystallization, UV-induced de-doping, and ionically active interfaces.

The resulting consensus is conditional rather than categorical. Spiro-OMeTAD remains the benchmark HTL because it supports very high efficiencies, can be processed in scalable formats, and can be tuned by interfacial and chemical engineering. At the same time, QFLS analysis, operando ESR, ultrafast spectroscopy, low-frequency noise, and thermal-aging studies all identify the perovskite/Spiro-OMeTAD region as a frequent locus of loss and degradation (Stolterfoht et al., 2018, Watanabe et al., 2020, Butler-Caddle et al., 2024, Sangwan et al., 2019, Paul et al., 17 Sep 2025). A plausible implication is that future use of Spiro-OMeTAD will depend less on its status as a default HTL than on whether interface engineering, dopant redesign, and thermal management can preserve its efficiency advantages without reproducing its characteristic interfacial liabilities.