Chapter 8 — Transflection Spectroscopy
In transflection geometry a thin sample layer is deposited on a reflective (“metallic”) substrate; the incident radiation passes through the layer, is reflected by the substrate, and traverses the layer a second time before reaching the detector. Because the radiation crosses the sample twice, transflection spectra are often read as transmission spectra with an effective path length of about twice the film thickness. This picture is intuitive but valid only under restrictive conditions: the reflective substrate imposes electromagnetic boundary conditions, and the incident and reflected waves interfere to form a standing-wave field whose intensity varies with position. Because absorption is governed by the local field intensity rather than by geometric path length alone, transflection spectra generally cannot be treated as Beer–Lambert spectra with a doubled path. (38), (54), (36)
These effects are most pronounced when the sample thickness becomes comparable to or smaller than the wavelength, and they have long been recognized in the spectroscopy of ultrathin films — films from a fraction of a monolayer to a few nanometers thick — where the measured spectra depend strongly on the optical properties of the substrate and on the measurement geometry. (54) The same situation arises elsewhere: in interference-enhanced Raman spectroscopy a thin dielectric layer is used to amplify the surface field, (55) and the infrared analogue is sometimes termed buried-metal-layer IRRAS. (54) That nearly identical geometries can either distort infrared absorption or enhance Raman scattering shows that the underlying phenomenon is a general consequence of electromagnetic wave propagation in layered media, not a peculiarity of one technique.
8.1 Concept and Geometry of Transflection Measurements
In the transflection geometry a sample layer of thickness \(d\) sits on a highly reflecting substrate, typically a metal-coated surface. Incident radiation of intensity \(I_0\) passes through the layer, reflects at the substrate, and traverses the layer again before reaching the detector (Figure 8.1).
Fig. 8.1
Transflection geometry. Incident radiation of intensity \(I_0\) passes through a sample layer of thickness \(d\) on a reflective (metallic) substrate, is reflected, and traverses the layer a second time before reaching the detector. The naive picture treats this as transmission with a doubled path length \(2d\); the following sections show why the reflective boundary makes that picture incomplete.
If wave-optical effects are ignored, the two traversals give an optical path of roughly \(2d\) at near-normal incidence — or \(2d/\cos\theta_1\) if the oblique path inside the layer is taken literally, with \(\theta_1\) the refraction angle in the film — and the detected intensity is written as
\[ I_{\mathrm{T}} = I_0\, \exp(-2 \alpha d) \tag{8.1} \]
where \(\alpha\) is the Napierian absorption coefficient of the sample. This is simply a doubled-path Beer–Lambert law, and it cannot be generally correct: in the limit of an optically thick layer essentially no radiation reaches the substrate, the second traversal disappears, and the detected signal approaches the Fresnel reflection response of the front surface — independent of any supposed path length \(2d\). The same misconception appears in ATR spectroscopy, where the evanescent penetration depth is sometimes read as an effective path length; in both cases the signal is set by electromagnetic boundary conditions, not by a geometric path (ATR is treated in Chapter 9). (38)
8.2 Optical Description of Transflection
\[ E(z) = E_{f}\, e^{i k_z z} + E_{b}\, e^{-i k_z z} \tag{8.2} \]
where \(E_{f}\) and \(E_{b}\) are the forward and backward amplitudes and \(k_z\) is the component of the wavevector normal to the interfaces. Their interference makes the local intensity
\[ I(z) \propto \left| E(z) \right|^{2} \tag{8.3} \]
vary with position. Because absorption is proportional to \(|E|^{2}\), the effective absorption is governed by this spatial field distribution, not by path length alone. For a highly reflecting substrate the backward amplitude approaches the forward one, so the field is strongly modulated — the physical origin of the standing-wave effects below.
Such systems are described by the well-established methods of thin-film optics, in which the stratified structure is solved from the complex refractive indices (or dielectric functions) of the materials and the interface boundary conditions. (36), (56) These models predict how the measured spectra depend on thickness, refractive index, absorption coefficient, and substrate, and so separate intrinsic molecular absorption from purely optical modifications.
8.3 Ultrathin Layers, the Long-Wavelength Limit, and the Surface Selection Rule
A particularly instructive case is the ultrathin layer, whose thickness is much smaller than the wavelength,
\[ d \ll \lambda . \tag{8.4} \]
Here the phase of the wave barely varies across the layer, and the response is governed by the boundary conditions imposed by the substrate rather than by propagation through the material. This is the regime of infrared reflection–absorption spectroscopy (IRRAS) and of ultrathin-film spectroscopy, which address the same optical situation of a thin layer on a reflective substrate. (54), (57)
The reflective boundary also fixes the orientation of the field, and this is decisive for ultrathin films. At a good metal surface the tangential (in-plane) component of the electric field is screened to nearly zero, while the perpendicular (out-of-plane) component is reinforced and can exceed the incident field. With p-polarized radiation at oblique incidence, an ultrathin film therefore experiences essentially only the perpendicular field: vibrational modes whose transition dipole has a component normal to the surface are excited, whereas modes lying parallel to the surface are suppressed.
This is the metal surface selection rule of IRRAS. (54), (36) Its practical consequence is that polarized transflection and IRRAS spectra of an ultrathin film encode molecular orientation: a band can be strong or absent depending on how its dipole is tilted relative to the surface, independent of how much material is present. Quantitative interpretation of such spectra therefore requires accounting for both polarization and orientation, not concentration alone (Figure 8.2).
The field enhancement just described concerns the perpendicular component selected by p-polarized light at oblique incidence; it should be distinguished from the tangential (in-plane) field treated in the next section, which is screened to a node at an ideal metal surface. The two pictures describe different field components of the same boundary-value problem and are not in conflict: a p-polarized perpendicular field can be reinforced at the metal even while the tangential field there vanishes.
Fig. 8.2
The metal surface selection rule for an ultrathin film: a 20 nm PMMA layer on gold, calculated using an oblique three-layer Fresnel model. (a) Polarized reflection–absorbance at \(80^{\circ}\) incidence. With p-polarized light the reinforced perpendicular field excites the full PMMA band pattern (carbonyl near 1730 cm\(^{-1}\), ester C–O–C near 1150–1270 cm\(^{-1}\)), whereas with s-polarized light the tangential field is screened and essentially no signal appears. (b) The carbonyl band measured with p-polarization grows steeply toward grazing incidence, while the s-polarized signal stays near zero — which is why ultrathin-film reflection spectra are taken with p-polarized light at grazing incidence.
8.4 Standing-Wave Effects and Their Spectroscopic Consequences
The interference between the incident and substrate-reflected waves produces a standing electromagnetic field inside the layer, with field intensity varying between nodes and antinodes across the thickness (Figure 8.3). The pattern is set by the optical constants of the materials and shifts with both wavelength and layer thickness.
Fig. 8.3
Electric-field standing wave above a gold mirror at \(\tilde{\nu} = 1700\) cm\(^{-1}\), computed from measured optical constants of gold. (a) The field intensity \(|E(z)|^2/|E_0|^2\) is a standing wave: pinned near a node at the metal surface, rising to an antinode (about four times the incident intensity) a quarter-wavelength away, and repeating every \(\lambda/2\). Because absorption follows the local \(|E|^2\), the signal a film produces depends on how much of this wave it occupies — a 20 nm film is trapped at the node and absorbs very little, a \(\lambda/4\)-thick film reaches the antinode, and a \(\lambda/2\) film spans a full node-to-node lobe and averages over it, so apparent band intensities depend on thickness and not on composition alone. (b) A zoom on the first 80 nm: the real metal (solid) departs slightly from a perfect conductor (dashed), leaving a small but non-zero field at the surface. An ultrathin film therefore samples a very weak field — orders of magnitude below the antinode value in this example — which is one reason such films give disproportionately small signals. This is the tangential (scalar) standing-wave picture; it should be distinguished from the perpendicular field component selected in p-polarized oblique IRRAS (§8.3).
Contrary to the textbook assertion that a metal suppresses interference entirely, the fringes are never completely absent in transflection; their amplitude depends on the conductivity of the substrate and vanishes only in the perfect-conductor limit. (38), (58) Their origin is instructive. Because a metal transmits nothing, reflectance and absorptance must sum to unity, \(R + A = 1\), so for a non-absorbing film the reflectance dips precisely where the standing wave drives a strong field against the metal and the metal itself absorbs.
The fringe amplitude therefore scales with how imperfectly the metal reflects, increasing from gold through silver to aluminum — for a low-index (\(n \approx 1.5\)) organic film the absorptance fringes stay below about 2% on gold but reach several percent on aluminum. This substrate dependence is weak for such low-index films but grows with the refractive index of the film, because a higher index raises the reflection at the buried film–substrate boundary (Figure 8.4).
Fig. 8.4
The substrate fringes are absorption in the metal. (a) For a non-absorbing film on gold the reflectance \(R\) and the absorptance \(A = 1 – R\) are mirror images: the transmittance is zero, so \(R + A = 1\), and where the standing wave drives a strong field against the metal the reflectance dips because the metal absorbs. (b) The same absorptance for gold, silver and aluminum. Its amplitude grows as the metal departs more strongly from the perfect-conductor limit — smallest for gold, largest for aluminum — and would vanish only for an ideal conductor; it also increases with the refractive index of the film.
Because absorption follows the local field intensity, this standing wave has direct spectroscopic consequences. The detected signal reflects the spatial overlap between the absorbing material and the field: material near an antinode contributes strongly, material near a node weakly. This differs fundamentally from the Beer–Lambert assumption of a uniform field. Two consequences follow.
First, the apparent band intensity depends on layer thickness, since changing the thickness moves the antinodes relative to the material; small thickness variations can therefore produce large intensity changes.
Second, because the standing-wave pattern is wavelength dependent, the apparent band strength does not track the intrinsic absorption coefficient — the field can enhance or suppress bands across the spectrum, producing thickness-dependent intensities, band-shape distortions, and apparent shifts of maxima.
Where the field is strong at the metal interface, some energy is dissipated in the substrate, which can add weak features even where the sample itself barely absorbs. For a layer with a high background dielectric function the effect is sharper still: the standing wave can produce additional maxima that coincide with the antinodes of \(|E|^2\) rather than with any vibrational resonance, so that a peak in a transflection spectrum need not correspond to a molecular transition at all. All of these originate in the electrodynamics of the layered system, not in the intrinsic optical constants of the material.
8.5 Spectral Distortions and Thickness Effects
The size of these effects depends on two distinct measures of the layer, not on physical thickness alone. The absorptive optical depth, \(\alpha d\) — the product of the absorption coefficient \(\alpha\) and the physical thickness \(d\) — determines how much radiation survives propagation through the layer, while the phase thickness, of order \(n d / \lambda\), sets the standing-wave pattern. Because the absorption coefficient varies strongly with wavenumber, a layer can be optically thin in one spectral region and optically thick near a strong band, so different parts of one spectrum fall into different regimes.
For thin and moderately thick layers the standing wave dominates: intensities, band shapes, and apparent peak positions all depend on thickness and wavelength, often giving asymmetric or derivative-like features. As the layer becomes optically thick, radiation is largely absorbed during the first pass before reaching the substrate; the reflected wave and the standing wave weaken, and the signal saturates — further thickness no longer increases the absorbance, and the spectrum approaches a front-surface reflection spectrum with derivative-like bands (Figure 8.5). This saturation is a key difference from transmission, where absorbance can in principle grow without bound as path length increases.
Fig. 8.5
Spectroscopic consequences for a model layer (a damped-harmonic-oscillator model of a PMMA-like material) on gold. (a) At a fixed thickness, the transflection absorbance \(-\log_{10}(R/R_0)\) is not the doubled transmission absorbance \(2\alpha d\): bands are reshaped and their relative intensities redistributed. (b) Tracking one band against layer thickness, the naive \(2\alpha d\) law rises without bound, whereas the real transflection signal is modulated by the standing wave and then saturates once the layer becomes optically thick.
A debate followed over whether an electric-field standing wave could really be responsible; (62) the debate was clarified by showing that the effect is genuine and is the same standing-wave phenomenon seen in free-standing films, merely far stronger because of the highly reflecting metal interface. (38) Thickness variations across a specimen can therefore mimic compositional differences, and where they must be removed a wave-optics-based correction is required. (47)
8.6 Practical Consequences for Spectral Interpretation
The central practical point is that transflection spectra are not simple absorption spectra: band intensities and shapes depend on layer thickness, substrate, and wavelength through the standing-wave field, so the Beer–Lambert law cannot be applied directly. In infrared microspectroscopy this matters most, because thickness naturally varies across a specimen and identical material can then yield visibly different spectra — easily misread as chemical heterogeneity. Band shapes and positions can likewise differ from transmission or ATR results, so spectra from different geometries are not directly comparable.
For many qualitative purposes, recognizing the standing-wave field and the thickness regimes above already explains the observed features. Quantitative work, however, requires modeling the sample and substrate as a stratified optical system, which reproduces the distortions and separates the material absorption response from geometry-induced optical contributions. The deeper resolution is to determine the material’s optical constants: once the complex refractive index or dielectric function is known, the description becomes geometry-independent, and transmission, ATR, transflection, and reflection spectra are seen as different manifestations of the same dielectric response. In this sense the various geometries do not probe different properties of matter, but different optical projections of one underlying response.
References
- Mayerhöfer TG. Wave Optics in Infrared Spectroscopy: Theory, Simulation, and Modeling. Elsevier; 2024. doi:10.1016/C2023-0-00262-X
- Mayerhöfer TG, Popp J. The electric field standing wave effect in infrared transflection spectroscopy. Spectrochim Acta A Mol Biomol Spectrosc. 2017;191:283-289. doi:10.1016/j.saa.2017.10.033
- Mayerhöfer TG, Pahlow S, Hübner U, Popp J. Removing interference-based effects from the infrared transflectance spectra of thin films on metallic substrates: A fast and wave optics conform solution. Analyst. 2018;143(13):3164-3175. doi:10.1039/c8an00526e
- Tolstoy VP, Chernyshova I, Skryshevsky VA. Handbook of Infrared Spectroscopy of Ultrathin Films. Wiley; 2003. doi:10.1002/047123432X
- Solonenko D, Gordan OD, Milekhin A, Panholzer M, Hingerl K, Zahn DRT. Interference-enhanced Raman scattering of F16CuPc thin films. J Phys D Appl Phys. 2016;49(11):115502. doi:10.1088/0022-3727/49/11/115502
- Yeh P. Optical Waves in Layered Media. John Wiley & Sons; 1988.
- Tano T, Umemura J. Deviation from the Lambert law in ultrathin films such as black soap films. Appl Spectrosc. 1997;51(7):944-948. doi:10.1366/0003702971941593
- Park KC. The extreme values of reflectivity and the conditions for zero reflection from thin dielectric films on metal. Appl Opt. 1964;3(7):877-881. doi:10.1364/AO.3.000877
- Dannenberg H, Forbes JW, Jones AC. Infrared spectroscopy of surface coatings in reflected light. Anal Chemistry. 1960;32(3):365-370. doi:10.1021/ac60159a019
- Gunde MK, Orel ZC. Thickness-dependent optical effects in infrared reflection-absorption spectra of a fairly thick polymer layer. Appl Spectrosc. 2002;56(1):24-30. doi:10.1366/0003702021954395
- Filik J, Frogley MD, Pijanka JK, Wehbe K, Cinque G. Electric field standing wave artefacts in FTIR micro-spectroscopy of biological materials. Analyst. 2012;137(4):853-861. doi:10.1039/c2an15995c
- Lee J. On the non-existence of the so-called “electric field standing wave effect” in transflection FTIR spectra. Vibr Spectrosc. 2017;90:104-111. doi:10.1016/j.vibspec.2017.03.007
The Field Guide is a living document. Questions, corrections, and suggestions for this chapter are very welcome — join the discussion in the Coblentz Field Guide forum.





