Following the Long-Term Evolution of sp$^3$-type Defects in Tritiated Graphene using Raman Spectroscopy

We report on the evolution of tritium-induced sp$^3$-defects in monolayer graphene on a Si/SiO$_2$ substrate, by comparing large-area Raman maps of the same two samples, acquired just after fabrication and twice thereafter, about 9-12 months apart. Inbetween measurements the samples were kept under standard laboratory conditions. Using a conservative classification of sp$^3$-type spectra, based on the D/D’ peak intensity ratio, we observed almost complete depletion of sp$^3$-type defects over the investigation period of about two years. This by far exceeds the ~5.5% annual reduction expected from tritium decay alone (~3x larger). This change in the defect composition is accompanied by a recovery of the 2D-band of graphene and an overall decrease in defect-density, as determined via the D/G intensity ratio. Hydogenated graphene is reported to be reasonably stable over several months, when kept under vacuum, but suffers substantial hydrogen loss under laboratory air conditions. While the results shown here for tritiated graphene exhibit similarities with hydrogenated graphene, however, some distinct differences are observed.

💡 Research Summary

This paper investigates the long‑term evolution of sp³‑type defects in monolayer graphene that has been functionalized with tritium (T‑graphene). The authors prepared CVD‑grown graphene on Si/SiO₂ substrates, exposed four 1 cm × 1 cm samples to a 400 mbar mixture of T₂ gas (97.2 % T₂) for 55 h, and thereby loaded the graphene with an activity of ≈19 MBq (≈7.6 × 10¹² Bq total). Two of the samples (A and B) were selected for detailed Raman analysis. Sample A was stored under normal laboratory atmosphere (≈21 °C, 40 % RH) after exposure, while sample B received additional thermal anneals (300 °C for 22 h and 500 °C for 22 h) to produce a “heated T‑graphene” (hTG) reference. Both samples were Raman‑mapped three times: (i) shortly after fabrication (low‑resolution data, labeled TG‑0), (ii) after ≈337 days (Campaign I), and (iii) after an additional ≈350 days (Campaign II). The Raman system employed a 532 nm laser (120 mW for Campaign I, 100 mW for Campaign II) and a 10× objective (NA 0.25), yielding a ≈7 µm spot size. Spectra were intensity‑calibrated with NIST SRM2242a, and peak fitting used pseudo‑Voigt functions.

The analysis relies on the Eckmann model, which uses the intensity ratio I_D/I_D′ to discriminate between sp³‑type defects (carbon‑hydrogen bonds) and vacancy‑type defects. In the low‑defect regime, I_D/I_G serves as a proxy for overall defect density, while I_D/I_2D and the full‑width at half‑maximum (FWHM) of the D and 2D bands provide additional insight into defect stage (Stage 1: low‑defect, Stage 2: high‑defect). The authors extracted a uniform 40 × 40 grid of spectra from each map to enable direct statistical comparison.

Key findings:

1. Initial state (TG‑0) – The low‑resolution data showed a high I_D/I_2D ratio and broadened D/2D peaks, indicating that the sample was in Stage 2 with a relatively high defect density.
2. After ~1 year (TG‑I) – Sample A displayed I_D/I_G ≈ 0.70 and I_D/I_2D ≈ 3.5, with I_D/I_D′ in the range 2–3, confirming that sp³ defects dominate. The 2D band was suppressed, reflecting significant disorder. Sample B (hTG‑I) showed a much lower I_D/I_G (≈0.15) and reduced D intensity, demonstrating that the annealing step had already removed most sp³ defects.
3. After an additional year (TG‑II) – Sample A’s I_D/I_G increased to 1.28, the 2D band intensity recovered to near‑pristine levels, and I_D/I_2D fell to 2.90. Crucially, I_D/I_D′ dropped dramatically, indicating that sp³ defects had almost completely vanished. Sample B’s spectra changed only marginally between Campaign I and II, confirming the stability of the annealed state.

The observed reduction of sp³ defects (~15–20 % per year) far exceeds the ~5.5 % loss expected solely from tritium β‑decay (half‑life 12.3 y). The authors attribute the accelerated defect loss to the energy deposited by decay electrons, which can break C–H (or C–T) bonds, and to ambient moisture/oxygen that facilitate hydrogen (or tritium) desorption under laboratory air conditions. This behavior mirrors earlier reports on hydrogenated graphene, which is stable only under vacuum but degrades rapidly in air.

Implications:

• Device reliability – For applications such as tritium‑based β‑electron sources, isotope‑separation membranes, or tritium‑resistant coatings, the rapid loss of functional groups could compromise performance over the intended service life.
• Mitigation strategies – Thermal annealing appears to “lock in” a low‑defect configuration that remains stable under ambient conditions. Alternative storage (vacuum, low temperature, inert atmosphere) or chemical passivation may also preserve sp³ functionality.
• Fundamental insight – The work demonstrates that Raman spectroscopy, combined with the D/D′ intensity ratio, is a powerful, non‑destructive tool for monitoring defect chemistry in radioactive graphene systems over multi‑year timescales.

In conclusion, the study provides the first systematic, long‑term Raman‑based assessment of sp³‑defect dynamics in tritiated graphene. It shows that sp³ defects are not only chemically unstable but also highly susceptible to decay‑induced bond breaking, leading to a rapid transition toward a more pristine graphene lattice. The findings highlight the need for careful material engineering and storage protocols when employing tritium‑functionalized graphene in practical technologies.