Skip to content

Through Diamond Robust Surface Enhanced Raman Spectroscopy

FieldValue
JournalAdvanced Functional Materials
Published2026-03-28
AuthorsKieran N. Twaddle, Rebecca J. Watkins, Ralph J. Moors, Massimiliano L.A. Ramsay, Richard B. Jackman
DOI10.1002/adfm.202525656

A 31.5 nm polycrystalline diamond film over Au nanoparticles produced a rear-illumination SERS enhancement factor of (6.2 ± 0.6) × 10³ under 532 nm excitation. The rigid substrate detected Rhodamine 6G at 1 × 10⁻⁶ M. Its 612 cm⁻¹ signal showed no significant loss after Au etching, acid cleaning, solvent sonication, six months in air, and 100 sandpaper abrasion passes. The structure separates the target environment from the spectrometer.

Only about 1 in 10⁷ incident photons undergo Raman scattering, so ordinary Raman spectroscopy is weak for trace analytes. SERS uses metal nanoparticles to concentrate excitation near adsorbed molecules through localized surface plasmon resonance. Exposed Au nanoparticles on a 2 mm single-crystal diamond window gave enhancement factors of 6.2 × 10⁴ for front illumination and 4.5 × 10⁴ for rear illumination. The rear signal was 0.7 times the front signal for 1 × 10⁻⁴ M Rhodamine 6G at 532 nm.

Rear illumination sends excitation and collection through the 2 mm window to nanoparticles under the analyte-side diamond film. This puts the spectrometer behind the rigid face. Encapsulation reduced signal compared with exposed particles, but it protected the plasmonic structure. Earlier protective coatings and a nickel-foam-caged Ag-Au structure improved some failure modes, but the authors state that diamond had not previously served as both the window and nanoparticle protection layer.

The starting optical-grade {100} single-crystal diamond was 12 mm in diameter, 2 mm thick, and polished on both sides. After acid cleaning, a 13.5 nm Au film was sputtered onto the surface. A hydrogen plasma dewetted the film into nanoparticles at 560 °C. Processing used 199.5 sccm H₂, 45 Torr, 930 W, and a 5 × 10⁻⁵ mbar base pressure for about 10 minutes.

SEM gave a mean nanoparticle diameter of 24.9 ± 0.8 nm and a median of 25.7 nm. Diameters ranged from 2 to 50 nm, and mean edge-to-edge separation was 15.3 nm. The authors did not control particle spacing. They chose dewetting because methods for tighter control had temperature and contaminant conflicts with diamond growth.

Microwave plasma CVD added polycrystalline diamond in two 40-minute steps at 560 °C. The gas mixture was 0.75% CH₄ in H₂, with 198.5 sccm H₂ and 1.5 sccm CH₄. The base pressure was 2 × 10⁻⁵ mbar. AFM and SEM tracked the nanoparticles during growth. The finished overgrowth was 103.9 nm thick and had Rq of 1.7 nm and Ra of 1.36 nm.

Raman quality tracking used the 1332 cm⁻¹ diamond line and the D band near 1345 cm⁻¹. The diamond line to D band ratio was 7.9 for the original single-crystal substrate and 3.8 for the grown film. The authors associate the lower ratio with more sp² carbon in the polycrystalline film and note that 532 nm excitation favors sp² carbon detection.

Before SERS testing, Au etchant removed exposed particles and the Au growth mask. The measured enhancement therefore came from encapsulated particles rather than intact exposed particles.

The 103.9 nm film suppressed the R6G signal because its thickness blocked the enhancement field from reaching the diamond and analyte interface. The authors etched the film with O₂ plasma at 200 W, 40 sccm, and 10 Pa. Etching over 30 to 180 seconds produced depths of 12.5 to 96.0 nm in masked areas. AFM step measurements gave each remaining thickness.

Larger nanoparticles breached the surface first. Exposed particles appeared after 64.0 nm of etching, with 39.9 nm of film remaining. All particles were exposed after 96.0 nm of etching, with 7.9 nm remaining. The fully exposed particles had a mean diameter of 24.5 nm, close to the original distribution. Au etchant then removed them.

The measured factors for films that retained enhancement were:

Diamond film thicknessFront-illumination EFRear-illumination EF
27.7 nm(3.3 ± 0.4) × 10³(5.0 ± 0.3) × 10³
31.5 nm(3.4 ± 0.3) × 10³(6.2 ± 0.6) × 10³
39.9 nm(2.9 ± 0.8) × 10³(4.3 ± 0.8) × 10³

At 31.5 nm, the rear factor was 1.8 times the front factor. The authors report at least 1.5 times the front value across the three thicknesses. They expect this difference because rear illumination focuses and collects through the single-crystal window rather than the polycrystalline film.

The encapsulated film reversed the exposed-particle result, with rear illumination producing the stronger signal. Nanopits left after larger particles were removed may also contribute to the measured enhancement. Films at or below 14.3 nm exposed nanoparticles that etchant removed, while films at or above 66.5 nm remained too thick for a detectable signal.

Each test area was immersed in 1 × 10⁻⁴ M R6G for 10 minutes, rinsed with deionized water, and dried with N₂. Measurements used 532 nm excitation at 45 mW, 20-second exposures, five accumulations, and six locations per area. The authors subtracted the fluorescence background. The enhancement factor compared the 612 cm⁻¹ peak with standard Raman intensity and corrected for the molecular counts in each sampled volume. Silicon reference measurements corrected for laser power drift.

Concentration tests covered 1 × 10⁻⁴ M to 1 × 10⁻⁸ M with the 31.5 nm film in rear illumination. No R6G peaks resolved at 1 × 10⁻⁷ M or below, but the 612 and 770 cm⁻¹ peaks resolved at 1 × 10⁻⁶ M. The 612 cm⁻¹ intensity was 948, above the blank threshold of 159, which represented three times the blank standard deviation. The authors therefore reported a 1 ppm limit of detection. They described this limit and the 6.2 × 10³ rear factor as comparable with commercial SERS substrates.

The optimized substrate underwent chemical, storage, and physical tests. It spent 5 minutes in Au etchant with at least 30% and less than 50% KI, plus at least 2.5% and less than 10% I₂. It then spent 20 minutes in an acid cleaning solution at 200 °C. It was sonicated for 5 minutes each in methanol, acetone, and isopropyl alcohol. It remained in ordinary storage in air at room temperature for six months.

For abrasion, a 10 g mass pressed the SERS face against 3000-grit sandpaper. The substrate moved 10 cm per pass for 20 passes, then 80 more passes, for 100 total. Each abrasion was followed by acid cleaning. After the tests, the 612 cm⁻¹ intensity at 1 × 10⁻⁴ M showed no significant variation in rear illumination. Silicon references corrected laser fluctuations.

The preserved particles support repeated use under these conditions. The authors describe this as “near-infinite reusability,” but the study measured performance over the stated test sequence rather than an unlimited lifetime.

The architecture separates the spectrometer from the target environment while keeping the plasmonic layer embedded in diamond. The authors propose an integrated device with optical and electronic components behind the active window, but they did not build or test one. They suggest that a thinner single-crystal window could raise the enhancement factor, with lower mechanical robustness as the trade-off.

The 31.5 nm optimum was measured with 532 nm excitation, R6G, and this broad nanoparticle distribution, so it should not be treated as a general design constant. Although the paper cites polycrystalline diamond resilience up to 750 °C in air, the study did not perform a thermal exposure test. The work also did not perform field deployment, high-pressure, or radioactive exposure tests. It did not establish performance for other analytes or excitation wavelengths. The integrated-system concept and broader extreme-environment use therefore remain unverified.

The paper needs a {100} single-crystal diamond SERS window, 12 mm in diameter and 2 mm thick, with Au nanoparticles encapsulated by a 31.5 nm MPCVD-grown polycrystalline diamond layer. The 6CCVD diamond materials catalog lists MPCVD-grown single-crystal and polycrystalline diamond, plates and wafers up to 125 mm, substrates up to 10 mm, and Au metallization. The linked engineering support can review the specified SERS window, Au coating, and encapsulation layer against these catalog options.

View Original Abstract

ABSTRACT Surface‐Enhanced Raman Spectroscopy (SERS) is a form of Raman spectroscopy which utilizes nanoparticle plasmonic interactions to enable trace‐level molecular measurement. However, conventional SERS measurements are restricted to the laboratory environment due to the fragile nature of the plasmonic nanostructures required. Therefore, to enable the use of SERS in extreme environments for in situ analysis, a robust and chemically stable plasmonic substrate is required. Moreover, the deployment of SERS systems in extreme environments would be greatly enhanced if the optical illumination technology could be placed behind a SERS active surface. Laboratory‐based growth of thin diamond films offers a unique opportunity for SERS substrates; the encapsulation of plasmonic nanostructures within a transparent, chemically stable, and physically robust shell. In this work, the diamond acts as both a window and protective layer, enabling illumination of the plasmonic nanostructures through the diamond itself, with light stimulating NPs encapsulated on the opposite face. This innovative ‘rear‐illumination’ approach enables spectrometer protection via through‐diamond measurements. nm Au nanostructures have been encapsulated in varied diamond film thicknesses, providing an optimized SERS enhancement of 6.2 10 3 . This enhancement is comparable to commercially available SERS substrates, while the novel diamond platform offers conceivably near‐infinite reusability in virtually any environment.