In comparison to traditional one-dimensional spectroscopy, where absorption or emission is measured as a function of energy (typically wavelength or frequency), two-dimensional spectroscopy allows analysis of the spectrum along two energy axes, exposing detail that is not present in one-dimension. These details consist of connections between electronic transitions, exciton – exciton coherences, population transfer, separate measurements of the homogeneous and inhomogeneous linewidth, or structural change to the molecules of interest. In a complex system, where there is an ensemble of different molecular configurations, this allows for tracking the evolution of subsets of the ensemble in time, whereas the one-dimensional spectrum is only the average of the ensemble. In other words, one can track the changes to absorption or emission of a sample at a given probe frequency caused by excitation in a narrow bandwidth.
Two-dimensional spectroscopy with ultraviolet, visible, or infrared pulses in the time domain is only possible with the use of femtosecond pulses. The first pulse creates a coherent superposition between the ground and excited state of the sample, which oscillates and decays over the course of the coherence time, often denoted as τ. The second pulse converts and stores the superposition as a population in the ground and excited state. After a waiting time, denoted as T, another pulse, or series of pulses, recreates a superposition of states, which emits radiation that is typically measured in the spectral domain. The time between the third pulse and the measurement of the signal is referred to as the detection time and is denoted as t. Often, the data is converted to the frequency domain by Fourier transform with respect to τ and plotted versus the detection frequency.
It is important that there is a known and controllable phase relation between electric fields of the first two pulses. It is also important to have phase-resolved detection for the emitted signal, which also entails a high degree of phase stability. With visible pulses, this corresponds to less than a femtosecond stability between the pulses. It is then required to use a pulse shaper to form the first two pulses and in some cases to have active measurement or control of the relative phase of the pulses. The same can be said for the last two pulses, however, as previously discussed, when recorded in the frequency domain by interference , the phase stability is passively ensured.
2DSpec requires the use of Fastlite's Dazzler™ acousto-optic programmable dispersive filter to form the first two "pump" pulses with high precision in the control of relative phase and delay. The pump pulses take the optical path; they are collinear. The third pulse, or "probe," is crossed at a small angle with the pumps and the emitted signal is collinear and phase matched with the probe beam for subsequent heterodyne detection, with the probe serving as the local oscillator. Other 2D implementations use non-collinear phase matching of all pulses. While there are some advantages to this technique, the experimental implementation and resulting data analysis are much more difficult. In the "pump-probe" geometry the "absorptive," narrowed, line shape is directly extracted after Fourier transform. It is also possible to implement phase-cycling to extract other signals of interest.