Of the imaging techniques available, MALDI imaging is the workhorse: it reaches the widest range of analyte classes, and it is the approach most laboratories encounter first. It is also the one where sample preparation does most of the work. The instrument matters, but a poorly chosen or poorly applied matrix will limit an experiment more than any specification will.
Key Takeaways
|
The MALDI Imaging Principle
A thin tissue section is mounted on a conductive slide and coated with a matrix, a small organic compound that absorbs at the laser wavelength. A pulsed laser is then fired at a defined position. The matrix absorbs the energy, desorbs from the surface carrying analytes with it, and transfers charge to them in the resulting plume. The ions are extracted into the mass spectrometer and a spectrum is recorded for that position. The stage steps, and the process repeats across a grid.
Two features of that sequence explain most of MALDI’s behaviour. The ionisation is soft, so large molecules survive largely intact and fragmentation is minimal, which is why the accessible analyte range extends from small metabolites through to intact proteins. And it produces predominantly singly charged ions, which yields considerably less crowded spectra than electrospray would — a real advantage when there is no chromatographic separation to spread the mixture out in time, as discussed in Mass Spectrometry Imaging: Principles, Techniques, and Applications.
Working in analytical science?
Register for a FREE Separation Science account to subscribe to the Separation Science Newsletter.
The matrix is doing several jobs at once, and this is worth being explicit about because it explains why the choice is constrained. It absorbs laser energy so the analytes do not have to. It isolates analyte molecules from one another, limiting aggregation. It participates in charge transfer. And its crystals physically define the spatial units from which material is sampled. A compound that performs the first three well but crystallises coarsely will still cap your resolution.
How Do You Choose a Matrix?
By analyte class and ion mode first, then by crystallisation behaviour. There is no universal matrix, and the choice is genuinely consequential rather than a formality.
Matrix | Typical Mode | Commonly Applied To | Notes on Behaviour |
DHB, 2,5-dihydroxybenzoic acid | Positive | Lipids, metabolites, and one of the two most widely used matrices overall | Crystallises coarsely relative to others, so deposition method matters most here |
CHCA, alpha-cyano-4-hydroxycinnamic acid | Positive | Peptides, lipids, small molecules | Fine crystallisation, and the matrix used in sub-micron single-cell work |
DHAP, 2,5-dihydroxyacetophenone | Positive | Lipids, proteins | Preferred over sinapinic acid for sublimation and spray coating |
Sinapinic acid | Positive | Proteins | More reliable for spotting applications than for sublimation |
9-aminoacridine | Negative | Acidic species | Established negative-mode option |
Norharmane and DAN, 1,5-diaminonaphthalene | Both | Lipids, including high-resolution work | Among the matrices evaluated for 10 micron lipid imaging |
Table 1. Matrices commonly used in MALDI imaging, compiled from the peer-reviewed studies cited in this article. Matrix performance is tissue- and analyte-dependent, so treat this as a starting point for optimisation rather than a lookup table.
Three practical points follow. A study evaluating six matrices for 10 micron lipid imaging assessed them not only on signal intensity but on lipid coverage, ability to support on-tissue tandem MS, and achievable useful spatial resolution in both ion modes, which is the right set of criteria and a reminder that intensity alone is a poor selection basis. Ion mode is a real fork rather than a setting, since acidic and basic species favour opposite polarities and a single acquisition commits you to one. And matrix stability under vacuum matters for long acquisitions: a high-resolution image across a large section can take hours, and a matrix that sublimes away during the run will produce a gradient that looks like biology.
Sublimation or Spraying?
This is the decision that most determines what a MALDI image can resolve, and it is usually presented as a matter of laboratory preference. The published comparisons are more useful than that, because they quantify the cost in both directions.
The governing rule is stated plainly in a 2025 Analytical Chemistry study of matrix application methods: to achieve images with high spatial resolution, uniform application of fine matrix crystals is necessary, with matrix crystal size kept below the intended pixel size. Because crystal size is influenced by the presence of solvent, sublimation, a dry deposition, is regarded as the gold standard for producing ultrasmall crystals.
| Sublimation | Spraying |
Deposition | Dry. Matrix vaporised and condensed onto the section | Wet. Matrix in solvent applied as an aerosol |
Crystal size | Ultrasmall. Around 400 nm reported for sublimed CHCA in single-cell work | Larger, and influenced by solvent and spray parameters |
Resolution outcome | With DHB at 10 micron imaging, significantly better resolution; structural boundaries clearly discernible | With DHB at 10 micron, hippocampus and corpus callosum boundaries not easily discerned |
Coverage outcome | Lower. Lack of solvent leaves some compounds undetected | Higher. An optimised automatic sprayer with DHB detected roughly double the metabolites of sublimation or airbrush |
Analyte delocalisation | Less, since there is no solvent to move analytes laterally | More, though only a subset of the lipid spectrum proved susceptible |
Reproducibility | Good, and largely operator-independent once the apparatus is set | Automatic sprayers were more reproducible and caused less diffusion than airbrush application |
Mitigation available | Post-sublimation recrystallisation in a humidity chamber may recover higher-mass compounds | Optimise solvent composition and flow to limit spreading |
Table 2. Sublimation against spraying, using figures from the two Analytical Chemistry studies cited in this section. Both examined DHB among other matrices; outcomes are matrix- and tissue-dependent.
Read That Table as a Resolution Against Coverage Decision The two findings are not in conflict. Sublimation produces finer crystals and therefore better spatial detail, and the same absence of solvent that keeps crystals small also means less extraction, so some compounds are never mobilised into the matrix layer at all. Spraying extracts more effectively and detects more species, and the solvent that achieves this also permits lateral movement and coarser crystals. So the question is not which method is better but which loss you can afford. Suppose you are imaging a small number of abundant lipids at fine resolution, sublimate. If you are surveying a metabolite panel and 30 or 50 micron pixels will answer the question, an optimised automatic sprayer will show you more. This is the same resolution against coverage axis that governs the technique as a whole, arriving here as a sample preparation decision. |
One further note on the airbrush, since it remains common in laboratories without a dedicated sprayer. In the comparison above, it performed worse than the automatic sprayer on both reproducibility and analyte diffusion, which is unsurprising given the operator dependence involved. If manual application is your only option, treat reproducibility as something to demonstrate rather than assume. Deposition methods are covered further in Matrix Selection and Application in MALDI Imaging.
Laser Rastering and Spatial Resolution
Achievable resolution is set by the smaller of two things: the laser spot size and the matrix crystal size. Optimising one while ignoring the other wastes effort, which is why the previous section matters as much as the instrument specification.
Four parameters govern the acquisition.
- Spot size. The laser focus, which modern instruments allow to be adjusted. It sets the physical area sampled per shot.
- Raster step. The distance the stage moves between positions. Setting the step smaller than the spot size oversamples, which can improve apparent image quality but also samples partly ablated material.
- Shots per pixel. Multiple laser shots are typically summed at each position to build usable signal. More shots means better spectra and longer acquisitions.
- Laser fluence. Energy per pulse, which trades signal intensity against fragmentation and matrix background.
The single-cell end of this is worth knowing about because it shows what the parameters can deliver when all of them are aligned. A MALDI imaging protocol for spatial bottom-up proteomics at single-cell resolution used sublimation of CHCA followed by a dip in ice-cold ammonium phosphate to produce peptide-rich spectra while maintaining matrix crystal sizes around 400 nanometres. That enabled imaging of proteins in single cells on a conductive slide at a throughput of roughly 7,800 cells per day, detecting 89 peptide-like features from a single breast cancer cell, and identifying 24 peptides corresponding to 17 proteins when the imaging data was combined with conventional LC-MS/MS on cell pellets.
Two things in that result are instructive beyond the headline. The crystal size was the enabling variable, not the laser. And the identifications came from combining imaging with a solution-phase experiment, which is a pattern worth internalising: MALDI imaging localises, and a complementary technique confirms.
Detectable Molecular Classes
MALDI reaches the broadest range of any imaging ionisation approach, which is the main reason for its dominance. A review of mass spectrometry imaging for spatially resolved multi-omics mapping places metabolites, lipids, peptides, proteins, glycans, and drugs within its scope, and a society-published guide to mass spectrometry imaging makes the same point about its breadth relative to the alternatives.
In practice, the classes behave quite differently, and it is worth knowing which you are attempting.
- Lipids. The most productive class. Abundant, ionise readily, strongly compartmentalised, and well served by several matrices. Most method development papers use lipids for good reason.
- Metabolites. Accessible but harder, because they sit in the mass range most affected by matrix-derived ions and because ionisation efficiency varies enormously across chemical classes.
- Peptides. Usually require on-tissue enzymatic digestion first, which adds a wet step and a delocalisation risk but opens up protein-level questions.
- Intact proteins. Possible, with sinapinic acid or DHAP among the usual choices, though sensitivity and resolution are both harder to achieve at higher mass.
- Drugs and their metabolites. A major application area, and one where MALDI has no real competitor among spatial techniques because no antibody or probe exists for a xenobiotic.
On-tissue tandem MS is worth pursuing wherever the instrument supports it. The six-matrix lipid study cited earlier identified 76 peaks using on-tissue tandem MS, which is a considerably stronger basis for assignment than accurate mass alone — particularly for lipids, where isobaric and isomeric overlap is the dominant identification problem.
What Are the Real Limitations?
Five, and knowing them in advance turns most of them into design decisions rather than disappointments.
Limitation | Why It Arises | What To Do |
Matrix ions crowd the low mass range | The matrix is a small organic compound present in vast excess, producing clusters and fragments across the low m/z region | Use high mass resolving power, choose a matrix with less interference in your window, or consider matrix-free approaches |
Resolution capped by crystal size | Sampling cannot be finer than the physical units of matrix present | Sublimate, or optimise spray parameters, and verify crystal size rather than assuming it |
Analyte delocalisation | Solvent in wet deposition can move analytes laterally before crystallisation | Prefer dry deposition for fine work, and validate with a structure of known geometry |
Ion suppression varies across tissue | Local composition affects ionisation efficiency, so intensity is not straightforwardly proportional to concentration between regions | Treat images as qualitative unless a quantitative strategy is built in deliberately |
Long acquisitions at fine resolution | Pixel count scales with the square of resolution, and shots per pixel multiply it further | Specify the resolution the question needs, and consider whether a larger step answers it |
Table 3. The practical limitations of MALDI imaging and the usual responses. None is a reason to avoid the technique; all are reasons to design the experiment before running it.
The first row is the one that surprises newcomers from a chromatographic background. In LC-MS, the matrix, in the general sense, is separated away before detection. In MALDI, it is deliberately present in large excess and ionises alongside the analytes, so the low-mass region is genuinely congested. That is a strong argument for mass resolving power in small-molecule imaging, and part of why the analyser choice discussed in the section hub matters so much.
For the ambient alternative that avoids matrix application altogether, along with what that costs in resolution, see DESI and Ambient Ionization Imaging. For where MALDI imaging sits alongside the antibody and sequencing-based spatial techniques, see Spatial Analysis in Analytical Science: Mass Spectrometry Imaging and Spatial Omics.
This article was produced under Separation Science's AI Editorial Guidelines.
