Surface engineering of zinc phthalocyanine organic thin-film transistors results in part-per-billion sensitivity towards cannabinoid vapor

Phthalocyanines (Pcs) and metal phthalocyanines (MPcs) are stable, conjugated macrocycles widely used in optoelectronic applications and as active semiconductors in organic thin-film transistors (OTFTs). Pc-based OTFTs have shown sensing capability for various analytes, including cannabinoids such as Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD), which are pharmacologically distinct and require accurate, low-cost quantification and speciation. Conventional methods (HPLC, GC–MS) can be impractical for many users. Prior work linked OTFT sensing responses to analyte-induced changes in Pc thin-film crystallinity and electrochemical interactions, emphasizing the importance of material selection. Because Pc properties can be tuned via central metal and peripheral substitution (e.g., fluorination), and film morphology can be engineered through deposition and annealing, the study investigates how central metal choice (Cu vs Zn), degree of fluorination, deposition conditions, morphology, thickness, and polymorphism affect sensitivity to THC vapor. The central hypothesis is that controlling thin-film structure and surface morphology will modulate analyte-induced physical and electronic effects and thereby enhance OTFT sensitivity to THC vapor.

The paper reviews that Pcs have historically been used as dyes/pigments and more recently in OTFTs and OPVs due to favorable spectral and electronic properties. Peripheral fluorination (F4, F8, F16) tunes band structure, crystal morphology, solubility, and can drive n-type behavior in OTFTs. Deposition and post-deposition annealing (thermal or solvent vapor) optimize charge transport by improving intermolecular packing and reducing defects and unfavorable interface energetics. Prior studies showed analyte exposure can alter Pc thin-film crystallinity similarly to annealing, affecting device performance. Central metal identity and degree of fluorination influence analyte interactions and resulting electronic/structural changes. ZnPc has been highlighted as highly tunable and sensitive, with analyte-induced structural modifications impacting charge transport pathways.

Materials included CuPc, ZnPc, F16-CuPc, F16-ZnPc (commercial), and synthesized F4-ZnPc (verified by mass spectrometry). All phthalocyanines were purified by train sublimation. BGBC OTFTs were fabricated on Si/SiO2 (230 nm dielectric) with pre-patterned Au electrodes (W = 2000 μm, L = 10 μm). Thin films for XRD/GIWAXS were deposited on Si/SiO2 (300 nm) substrates. Films were deposited by physical vapor deposition (PVD) at controlled rates and substrate temperatures using an Angstrom EvoVac evaporator; quartz crystal microbalance measured rate/thickness, thermocouple controlled substrate temperature. Screening: 400 Å films of CuPc, F16-CuPc, ZnPc, F4-ZnPc, and F16-ZnPc were deposited at 25 °C and 0.2 Å/s on OTS-treated Si/SiO2. Pre-exposure characterization used GIWAXS and XRD to assess crystallinity and orientation; OTFT transfer/output curves were recorded to extract mobility (μ), threshold voltage (VT), hysteresis, on/off ratio, and defect density (N). Vapor exposure: THC was dissolved in methanol, dried on a steel wool frit, and vaporized using a Volcano Medic at 210 °C. The vapor (collected in an 8 L balloon) was delivered into a 50 mL chamber containing the devices/films. Initial screening exposures were 4 ppm THC for 90 s. For morphology/thickness studies, lower concentrations (400 ppb or 40 ppb) were used as specified. Morphology control: α-ZnPc films (400 Å) with varying crystallinity were prepared by adjusting deposition rate and substrate temperature, and by using a p-sexiphenyl (p-6P) monolayer as a patterning agent: low crystallinity (1 Å/s, 25 °C), medium (0.05 Å/s, 25 °C), high (0.2 Å/s, 140 °C), very high (0.2 Å/s, 180 °C on p-6P). AFM quantified grain size/surface area; XRD assessed peak intensities; OTFT transfer curves were measured pre/post exposure to 400 ppb THC (90 s). Thickness effects and real-time sensing: Low-crystallinity α-ZnPc films with 200 Å and 800 Å thicknesses were fabricated (1 Å/s, 25 °C) and exposed to 40 ppb THC for 90 s. Real-time measurements were performed by operating devices at fixed saturation biases (e.g., VSP = −50 V, VGS = −40 V) while continually or periodically flowing 40 ppb THC over the device in a 50 mL chamber; current vs time was recorded and slope extracted from 20–120 s. Control exposures to heated air and to 40 ppb cigarette smoke were conducted. Polymorphism: β-ZnPc films were obtained by exposing 400 Å α-ZnPc films (deposited at 0.2 Å/s, 25 °C) to toluene vapor at 50 °C for 24 h in a custom chamber, followed by 70 °C vacuum baking for 45 min. β-phase confirmation used XRD (2θ peaks at 7.04° and 9.32°), GIWAXS (orientation and planes), and SEM (μm-scale rectangular crystals). These β-ZnPc OTFTs were exposed to 400 ppb THC (90 s) and characterized pre/post exposure; real-time response was also measured. Characterization: OTFTs were measured with a Keithley 2614B; mobility extracted from saturation-regime transfer curves: IDS = μ Ci (W/2L) (VGS − VT)^2. Bias-stress mitigation used pulsed gate (20 ms on/80 ms off). Defect density N estimated from subthreshold slope S using N = (q/kT ln 10) S. XRD used a Rigaku Ultima IV with Cu Kα (λ = 1.5418 Å), 5° < 2θ < 11°, 0.5° min−1, 30 rpm spin. GIWAXS performed at CLS (15.1 keV, Rayonix MX300, 416 mm, θ = 0.3°) and SOLEIL (10 keV, PILATUS3 S 1M, 330 mm, θ = 0.1°), calibrated to silver behenate and P3HT standards; analyzed with GIXSGUI (polarization and solid-angle corrections). AFM (Bruker Dimension Icon, ScanAsyst-Air tips) and SEM (Tescan Vega II, 20 kV) characterized surface morphology.

• Material screening (4 ppm THC, 90 s): Among CuPc, F16-CuPc, ZnPc, F4-ZnPc, and F16-ZnPc (400 Å films), unsubstituted ZnPc showed the largest changes across electrical metrics (μ, VT, N, on/off, hysteresis) and XRD intensity, indicating greatest sensitivity to THC vapor. Perfluorination (F16) limited structural changes and responses; central metal influenced VT shifts (CuPc and F16-CuPc showed +ΔVT consistent with electron trapping; ZnPc, F4-ZnPc, and F16-ZnPc showed −ΔVT indicating deep, short-lived hole traps).
• Morphology dependence in α-ZnPc (400 Å): Lower crystallinity (small grains, lower XRD peak) yielded the strongest sensor responses upon 400 ppb THC exposure (90 s): decreased μ (−%Δμ), negative VT shift (−ΔVT), increased hysteresis, decreased on/off, increased N, and reduced XRD peak intensity. Increasing crystallinity reduced sensitivity; the most crystalline films showed minimal electrical/XRD changes. AFM-derived surface area was inversely proportional to α-crystallinity (very high: 6.43; high: 6.52; medium: 6.61; low: 6.81 μm surface length in a 2.5×2.5 μm scan), suggesting increased surface area in low-crystallinity films enhances analyte interaction and structural change.
• Thickness effects at low crystallinity: 200 Å and 800 Å α-ZnPc films (1 Å/s, 25 °C) exhibited similar pre-exposure transfer characteristics, but thinner films showed larger THC-induced changes. Under 40 ppb THC (90 s), both thicknesses had +ΔVT, increased hysteresis, −ΔN, and decreased XRD peak intensity; 200 Å films showed larger on-current decreases and greater −%Δμ, indicating greater relative crystallinity changes.
• Real-time sensing (40 ppb THC): Operating α-ZnPc OTFTs at VSP = −50 V, VGS = −40 V showed an immediate, partially reversible current drop at vapor onset (hole trapping), followed by a sustained, thickness-dependent current decrease (structural defects). Measured current decay slopes were thickness-dependent (e.g., approximately −0.068 μA/s for 200 Å, −0.040 μA/s for 400 Å, −0.017 μA/s for 800 Å from 20–120 s). After ~90 s exposure, operating current matched post-exposure bias-characterized values, indicating limited recovery with prolonged exposure. Controls with heated air showed negligible change; 40 ppb cigarette smoke produced small, noisy responses, evidencing selectivity for THC.
• β-ZnPc polymorph behavior: β-ZnPc OTFTs (converted via toluene vapor) exhibited a turn-on response to 400 ppb THC (90 s): +10.3 μA increase in peak current, −7.4 V VT shift, and +4.1 V hysteresis change. XRD showed no change at 2θ = 7.04°, a 53% decrease at 9.32°, a 20× increase of a shoulder at 6.84°, and emergence of a new peak at 7.38°, indicating significant morphological transformation. SEM revealed transition from narrow rectangular crystals to broader sheet-like structures. GIWAXS showed decreased (10−2) intensity at q = 0.68 Å−1, peak splitting around q ≈ 0.50 Å−1 ((100) plane), narrowed arcs, and a change in preferential molecular orientation from ~45° to ~74°. Post-exposure defect density was low (N = 5.4×10^−12, from subthreshold swing), consistent with improved transport along sheet-like structures.
• Sensitivity enhancement: By selecting ZnPc and engineering thin-film morphology and thickness (low-crystallinity α-phase, 200 Å), the devices achieved sensitivity to 40 ppb THC vapor and a 100× sensitivity increase compared to initially screened conditions and prior CuPc-based devices.

The findings support the hypothesis that both material selection and thin-film engineering critically dictate OTFT sensor response to THC vapor. ZnPc exhibits stronger interactions with THC than CuPc, with the central metal mediating trap formation (hole traps in Zn versus electron trapping in Cu), and peripheral fluorination reducing film restructuring and sensitivity. In α-ZnPc, lower crystallinity increases grain boundary density and surface area, promoting THC-induced structural perturbations that strongly modulate charge transport, yielding larger electrical responses. Thickness modulates sensitivity by altering the relative amount of analyte per semiconductor volume and the extent of structural change, with thinner films responding more strongly. Real-time measurements reveal a two-component sensing mechanism: a rapid, partially reversible decrease in current due to hole trapping by THC upon exposure onset, followed by slower, largely irreversible declines driven by THC-induced structural defects that degrade transport pathways, limiting reuse. In highly crystalline β-ZnPc, THC triggers pronounced recrystallization and reorientation, likely involving partial conversion towards α-like packing and formation of sheet-like domains that can improve transport, producing a turn-on response. Overall, optimizing polymorphism, crystallinity, and thickness provides a pathway to maximize sensitivity while understanding the mechanistic origins of the electrical signals in Pc-based OTFT sensors.

Non-fluorinated ZnPc OTFTs showed the highest electrical and structural sensitivity to THC vapor among the phthalocyanines tested. Peripheral fluorination reduced analyte-induced restructuring, while central metal identity controlled the direction of VT shifts. Engineering α-ZnPc film morphology demonstrated that lower crystallinity enhances sensor response; thickness also influenced sensitivity, with 200 Å low-crystallinity films detecting 40 ppb THC. Real-time studies revealed immediate, partially reversible hole trapping at vapor onset and sustained, irreversible current decreases due to structural changes. β-ZnPc films, although initially highly crystalline and lower performing, underwent significant morphology and orientation changes upon THC exposure and exhibited a turn-on response. Through targeted film engineering (material choice, morphology, thickness, polymorphism), the work achieved a 100-fold sensitivity improvement over prior CuPc-based devices, underscoring the importance of thin-film structural control in Pc-OTFT sensor implementations.

THC-induced structural changes produce largely irreversible decreases in operating current in α-ZnPc, limiting device reusability and making the sensing response partially irreversible. Real-time β-ZnPc responses showed inconsistencies attributed to uneven distribution and orientation of β-crystals between electrodes. The most sensitive low-crystallinity films were rendered inoperable under higher-concentration exposures (e.g., 4 ppm), necessitating lower ppb-level exposures for characterization.