The characteristic bands and what they genuinely prove, why a sharp 2D band does not confirm a monolayer, how strain and doping are separated, and a mapping protocol for a transferred film.
A transferred film of chemical vapour deposited graphene looks continuous and clean under the optical microscope, the transfer went smoothly, and the first Raman spectrum shows a defect band that should not be there. That sequence happens in every group that starts working with 2D materials, and the spectrum is usually right. Raman is the fastest and least destructive way to find out what a 2D film really is, provided the bands are read for what they demonstrate rather than for what the group hopes.
| Band | Approximate position with green excitation | Origin | What it tells you |
|---|---|---|---|
| G | near 1580 cm-1 | In-plane stretching of the carbon bonds | Present in all sp2 carbon. Its position and width respond to doping, strain and disorder |
| D | near 1350 cm-1 | Ring breathing mode that requires a defect to become active | Disorder, edges, vacancies, sp3 sites and grain boundaries. Its position shifts with laser energy |
| D prime | near 1620 cm-1, on the shoulder of G | A second defect activated process | Its ratio to D indicates the type of defect rather than only the amount |
| 2D, also written G prime | near 2700 cm-1 | A two phonon process that needs no defect | Shape and width report layer number and stacking. Strongly dependent on laser energy |
| D plus D prime | near 2940 cm-1 | Combination band, defect activated | Corroborates that a D band is real disorder rather than an artefact |
| Silicon substrate | near 520 cm-1 | The substrate itself | A convenient internal check that the calibration has not drifted |
Report the excitation wavelength every time. The D and 2D bands move with laser energy, so a position quoted without the wavelength cannot be compared with anything, and a reviewer will ask.
For graphene, the 2D band carries the layer information. A monolayer gives a single, sharp, symmetric band that fits well to one Lorentzian and is typically taller than G. Adding a second layer with the usual stacking splits the process into several contributions, so the band broadens, becomes asymmetric and no longer fits a single component. Through three and four layers it broadens further, and beyond roughly five layers the shape stops changing usefully. Raman is not a thickness gauge for thick films.
Now the case that produces wrong papers. Multilayer graphene with randomly rotated, or turbostratic, stacking is electronically decoupled, and it gives a single sharp 2D band that looks like a monolayer. A single Lorentzian 2D is therefore consistent with a monolayer but does not prove one. Anyone claiming monolayer coverage from Raman alone is over-reading the data. Confirm with optical contrast on a substrate chosen for it, a step height from atomic force microscopy, or electron microscopy.
The second trap is the ratio of 2D intensity to G intensity, widely used as a layer indicator. That ratio also responds to doping, so a monolayer sitting on polymer residue or on a charged substrate will show a suppressed ratio and be counted as bilayer. Use the band shape and width first, and treat the intensity ratio as supporting evidence.
The ratio of D intensity to G intensity is the standard defect measure, and it is not monotonic. Starting from perfect material, the ratio rises as defects are introduced, reaches a maximum, and then falls again as the material becomes genuinely disordered and approaches amorphous carbon. Two very different samples can return the same ratio. The way out is to look at the G band width alongside it, which broadens continuously with disorder and therefore tells you which side of the maximum you are on.
Both strain and charge doping move the G and 2D bands, which is why single-band interpretations of either are unreliable. Tensile strain shifts bands down in wavenumber and compressive strain shifts them up. Under sufficient uniaxial strain the G band splits into two components whose separation grows with the strain and whose polarisation behaviour identifies the strain axis. The 2D band responds more strongly per unit strain than G, which makes it the more sensitive indicator.
Doping behaves differently. The G band stiffens and narrows for both electron and hole doping, while the 2D band responds in a way that depends on the sign of the carriers. Because the two bands respond to strain and doping in different proportions, the standard treatment is to plot the 2D position against the G position for every pixel of a map. Strain and doping displace the data along two distinguishable directions in that plane, so the cloud can be decomposed into a strain component and a doping component. That single plot answers questions that no table of average band positions can, and it takes minutes to produce from a map you already have.
Local heating shifts and broadens the bands, and on a suspended or poorly heat-sunk flake it appears at power levels that seem harmless on a bulk sample. The resulting shift is then interpreted as strain or doping, and the conclusion is published. Run a power series at a representative point on every new sample type, plot band position against power, and work within the region where the position is flat. Record the power measured at the sample, not the laser output, because the objective and the optical path change it substantially.
For the same reason, take a spectrum, move to a fresh spot, and take another with the same settings. If the second differs from the first at the same point after exposure, the measurement is changing the sample.
On a transferred film, this sequence answers the questions that matter for device work: whether coverage is continuous, where the film is folded or wrinkled, where the transfer left residue, whether the grain structure of the growth survived, and whether the strain field is uniform across the area you intend to pattern.
The transition metal dichalcogenides, of which molybdenum disulphide is the most studied, show two prominent modes, one in-plane and one out-of-plane. The separation between them increases as layers are added, and that separation is the standard thickness gauge for those films. Monolayer molybdenum disulphide also becomes strongly photoluminescent, and the same instrument records that emission with the appropriate grating and detector range. A group planning work on these materials should therefore specify a Raman microscope that can also collect photoluminescence, rather than a Raman-only spectrometer, because the two measurements together identify a monolayer far more convincingly than either alone.
Hexagonal boron nitride gives a single strong in-plane mode and is much weaker scatterer than graphene, so it needs longer acquisitions. For every one of these materials the band positions and relative intensities depend on excitation wavelength, substrate and strain state, which is why the method section carries as much weight as the spectra.
Groups in Bangladesh working on graphene and related materials are usually competing for publication against laboratories with far larger facilities. Measurement discipline is the part of that competition that costs nothing. A clean power series, a recorded calibration, a stated wavelength and a strain against doping plot make a modest data set defensible, and their absence makes an excellent data set arguable.