Imaging Mass Spectrometry Using Post-Electron Ionisation: Differences in Secondary Ionisation Between the DESI and LDI Approaches

Post-photoionisation (PI) is not a generic module that can simply be ‘attached’ to a desorption method—there are significant differences in the operating conditions, physical processes and effects between the DESI and LDI desorption routes. Only by understanding these differences can one make full use of both imaging sources.

Two starting points: differences in desorption products

The neutral molecular clouds produced by DESI and LDI desorption differ in both their physical state and chemical composition.

DESI desorption produces a solvated aerosol plume. The desorbed molecules are encapsulated within micrometre-scale droplets, surrounded by a methanol/water solvent shell. The release of neutral molecules from the droplets requires solvent evaporation, which is precisely the same process by which ESI ionisation occurs. A key feature of DESI+PI is that the VUV beam encounters a solvent-containing cloud of neutral molecules in which the molecules are solvated and encapsulated.

LDI desorption (transmission mode, as used in MSI LDPI) directly converts the sample from a solid-state film into gaseous molecules and clusters. A laser pulse vaporises molecules at the substrate–sample interface within a few hundred femtoseconds to a few nanoseconds, forming a solvent-free, high-concentration cloud of neutral molecules in which the molecules are directly exposed.

The difference between these two ‘starting points’ permeates every subsequent step of post-photoionisation.

Path 1: Post-photoionisation in the DESI plume

In the DESI plume, the majority of the neutral molecules encountered by the VUV beam are still ‘hidden’ within the droplets. VUV photons (10.6 eV) can indeed penetrate methanol (the VUV absorption cross-section for a 116.5 nm photon is approximately 10⁻¹⁷ cm²/molecule), but photon attenuation within micrometre-scale droplets is significant—when the droplet diameter exceeds approximately 500 nm, the photon flux at the centre has already fallen to less than 1/e of that at the surface. This means that the probability of neutral molecules inside the droplets being ‘directly struck’ by VUV photons is far lower than that for surface molecules.

The good news is that the methanol/water solvent itself exhibits a certain degree of photoionisation response to 10.6 eV photons (methanol IE = 10.8 eV, which is just above the 10.6 eV main emission line of a Kr lamp and therefore will not be ionised); this means that the solvent itself does not compete for VUV photons. This is a key advantage in solvent selection—if a solvent with a lower IE (e.g. acetone, IE = 9.7 eV) were used, the solvent molecules would absorb a large number of VUV photons and be converted into solvent ions, thereby depleting the photon flux.

In practical DESI+PI operations (MSI DPI, Neo Source INSTRUMENT), the primary effect of PI stems from two sources: approximately 30–40 per cent comes from the direct photoionisation of neutral molecules on the surface, whilst approximately 60–70 per cent comes from neutral molecules exposed after droplet evaporation being captured by VUV photons—this ratio depends on the spray flow rate and the position of the VUV beam. The higher the flow rate and the larger the droplets, the greater the contribution from the latter; the closer the VUV beam is to the sputtering point, the greater the contribution from the former.

Pathway 2: Post-photoionisation in the LDI plume

The molecular cloud produced by LDI desorption contains no solvent; the molecules are ‘exposed’ from the moment laser ablation begins. This makes LDI+PI physically more direct than DESI+PI: there is no barrier provided by a solvent shell between the generation of neutral molecules and their irradiation by VUV photons, so the probability of photoionisation is theoretically higher.

However, LDI presents another issue: laser ablation simultaneously generates neutral molecules, primary ions and free radicals. Among these, primary ions and free radicals can undergo gas-phase ion–molecule reactions (such as charge transfer or proton transfer) with neutral molecules, which reduces the proportion of the ‘pure’ post-photoionisation contribution to the total signal. From a spectral perspective, the ion sources in the LDI+PI mode are more complex—some originate from direct ionisation by LDI, some from VUV photoionisation, and some from gas-phase reactions—whilst in the DESI+PI mode, the sources are relatively clear (ESI+PI).

One of the engineering advantages of transmission-mode LDI is evident here: the laser is incident from the rear, ablation occurs at the substrate–sample interface, and the ablated material consists predominantly of neutral molecules (compared to reflection-mode LDI), which reduces the dilution of the PI signal by primary ions and free radicals. The transmission-mode configuration of MSI LDPI complements post-photoionisation well in this respect.

Differences in signal gain between the two approaches

Based on published comparative data from mouse brain tissue:

• DESI→DESI+PI: Approximately 3–5-fold increase in PC-class signals, approximately 50-fold increase in cholesterol signals

• LDI→LDI+PI: Approximately 5–10-fold increase in lipid signals, with increases of up to 100-fold for certain small-molecule drugs

The overall signal enhancement under the LDI approach is higher than that of DESI, which is consistent with its physical characteristics of ‘no solvent barrier and direct molecular exposure’. However, the DESI approach still holds advantages in terms of tissue integrity (DESI desorption is gentler and causes less damage to the sample) and coverage of polar molecules.

Selection logic

If the primary objective is to detect non-polar small molecules (sterols, fatty acids, drug metabolites) and high spatial resolution is required, the LDI+PI approach (MSI LDPI) is more effective—the absence of a solvent barrier enhances PI efficiency, whilst the transmission configuration enables a resolution of 2–3 μm.

If the primary objective is comprehensive molecular coverage (ranging from polar to non-polar) and the sample consists of large-area tissue sections, the DESI+PI approach (MSI DPI) is more suitable—DESI itself offers high ionisation efficiency for polar molecules, whilst PI supplements the non-polar fraction; the spatial resolution of 20–200 μm is better suited to macroscopic tissue imaging.

These two approaches are not mutually exclusive, but rather complementary tools designed for different analytical objectives and spatial scales.

Related products: Neo Source INSTRUMENT’s MSI DPI (20–200 μm resolution, DESI+PI) and MSI LDPI (2–3 μm resolution, LDI+PI), both compatible with Agilent, AB SCIEX and Thermo mass spectrometers.