Analytical and Bioanalytical Chemistry

, Volume 410, Issue 3, pp 963–970 | Cite as

IR-MALDESI method optimization based on time-resolved measurement of ion yields

  • Måns Ekelöf
  • David C. Muddiman
Research Paper
Part of the following topical collections:
  1. ABCs 16th Anniversary


In the field of mass spectrometry imaging, typical experiments involve ionization directly from complex samples with no pre-ionization separation, relying on high resolving power mass analyzers to separate ions of interest. When an ion trapping step is involved in the analysis, the dynamic range of the analysis may be limited by the capacity of the ion trap, which is easily exceeded. To minimize collection of undesired ambient species while maximizing collection of analyte signal, accurate timing between ion generation and collection is a requirement. Here, a method for achieving synchronicity between infrared laser ablation and ion collection on a Q Exactive Plus mass spectrometer is described and demonstrated through measurement of ion accumulation at fixed time points following a laser ablation event with electrospray post-ionization of ablated material. In a model imaging experiment using infrared matrix-assisted laser desorption electrospray ionization, fixing the injection time at the minimum duration required to capture all ions generated by the last laser pulse in a sequence is shown to maximize target ion abundances. Using optimized timing is shown to yield a doubling or better of useful signal compared to previously used parameters.

Graphical abstract

Illustration of the effects of signal optimization on data quality for a single lipid species (cholesterol) measured from mouse liver tissue


IR-MALDESI Electrospray post-ionization Mass spectrometry imaging Laser ablation Q Exactive 


Electrospray post-ionization is an ambient ionization technique where a sample is ionized through interaction with an electrospray plume. The sample can be introduced in a number of different ways. Notable examples of strategies for sample introduction include direct contact with a sample surface [1], an intersecting aerosol [2], and the use of sampling probes such as a metal needle [3] or a laser beam [4, 5]. Interfacing electrospray post-ionization with spatially resolved sampling methods makes it suitable for mass spectrometry imaging (MSI) applications, where the recorded ion abundances are mapped back to a physical location.

In matrix-assisted laser desorption electrospray ionization (MALDESI) mass spectrometry [5], a laser is used as a sample probe to desorb analytes from a surface placed under an orthogonally oriented electrospray cone. In this manner, an ESI-like ionization is achieved, where ions are produced mainly through charge transfer in solvent droplets [5, 6]. By employing a mid-infrared (IR) laser, it is possible to use water or ice as an external matrix with the laser wavelength tuned to the 2940-nm absorbance maximum of the O-H stretching mode of water. In IR-MALDESI analysis of biological samples, an externally applied layer of ice is used to provide homogeneous sampling [7]. Keeping samples frozen throughout imaging experiments, which often require several hours, also serves to prevent chemical and enzymatic degradation as well as dehydration.

The IR-MALDESI ion source [8] used throughout this work was coupled to a Q Exactive Plus mass spectrometer (Thermo Scientific, Bremen, Germany). The Q Exactive series utilizes a bent quadrupole ion trap (C-trap) to accumulate ions prior to injection into the Orbitrap analyzer, making it suitable for use with continuous ion sources such as ESI and APCI [9, 10]. In typical ESI-MS using a Q Exactive Plus, the automatic gain control function (AGC) is used to keep the total trapped charge roughly constant between scans through varying the ion injection time (IT) based on a brief pre-scan of the total ion current (TIC). This has been shown to greatly improve mass measurement accuracy (MMA) combined with a strategy of using known ambient peaks as lock masses for continuous internal re-calibration [10].

To measure a pulsed injection from electrospray post-ionization, it is necessary to disable AGC, as ion accumulation must necessarily coincide with the burst of ions from the ionization event. In the IR-MALDESI source, C-trap accumulation is externally triggered, and IT is held at a constant value that is selected to ensure that all generated ions are captured. For animal tissue sections, two mid-IR laser pulses are typically required for complete sampling, which is a prerequisite for absolute quantification [11]. Using a commercially available 20 Hz laser, suitable injection times fall on the order of 100 ms, two orders of magnitude higher than typical LC-MS analysis using the same instrumentation. As a consequence, during a typical IR-MALDESI analysis sequence using a wide mass window, the C-trap is filled to levels far exceeding its capacity, which is on the order of 106 elementary charges [10]. The long trapping time leads to the collection of a large fraction of ambient ions of no analytical value, reducing the effective dynamic range of quantitative imaging experiments.

There are two obvious strategies for improving data quality through reducing total trapped charge: lowering trapping times to minimize the collection of ambient ions and limiting the m/z range allowed into the trap by means of a mass filter. The former has been investigated by Rosen and coworkers, who introduced a high repetition rate IR laser (100 Hz) in order to reduce the C-trap injection time of a Q Exactive instrument, noting a significant improvement to measured signal at lower injection times [12]. The latter strategy is suitable for targeted analysis, where broad coverage can be sacrificed for greater sensitivity to a particular ion of interest.

To provide a complete and detailed model of the generation and accumulation of ions from electrospray post-ionization, we here describe a method for synchronizing the Q Exactive Plus ion injection to the laser firing order, and detail several experiments measuring the accumulation of target ions at discrete times after the laser ablation. Using this system, we demonstrate a simple and rapid optimization method for finding both the minimum required accumulation time per ablation event and the maximum total trapping time for a given sample and desired mass range.

Materials and methods

Materials used

Animal tissues were stored at −80 °C until the time of analysis. Sectioning of tissues was done using a Leica (Buffalo Grove, IL, USA) CM1950 cryomicrotome. Tissue sections were thaw-mounted on glass microscope slides for IR-MALDESI analysis. LC-MS grade water and methanol were purchased from Acros Organics (Geel, Belgium). PEG-600 and MS-grade formic acid was purchased from Sigma-Aldrich (St. Louis, MO, USA). Nitrogen gas for enclosure purging and humidity regulation was purchased from ARC3 gases (Raleigh, NC, USA).

Control scheme

To achieve accurate synchronization between the laser ablation event and the start of Q Exactive ion collection, it was necessary to operate the instrument in the “low-latency handshake mode” provided by Thermo Fisher Scientific. In this mode, the instrument responds to an externally provided trigger signal much faster than the optional “handshake mode,” which introduces a variable delay of up to 100 ms between trigger and response. Detailed diagrams of signal schemes illustrating the difference between the modes are provided as Fig. S1 in the Electronic Supplementary Material (ESM).

To achieve the desired behavior in low-latency mode, a set of specialized external circuits were designed and constructed. In this design, a Quantum Composers (Bozeman, MT) 9214 Sapphire square pulse generator was used to control the flash lamp and Q-switch functions of an IR-Opolette 2371 OPO laser (Opotek, Carlsbad, CA, USA) and to provide a timing reference for ion injection triggering. Signal routing was handled by a control unit built around an Arduino Uno microcontroller (Arduino, Ivrea, Italy). With this system, MS acquisition could be synchronized to a laser trigger with microsecond precision as verified through repeated measurements on a THS-720A oscilloscope (Tektronix, Beaverton, OR). A complete circuit diagram and representative oscilloscope measurements are provided in the ESM as Figs. S2 and S3.

Safety considerations

The use of a class IV invisible laser for sampling necessitates a number of precautions for safe use. A LAZ-R-SHROUD barrier (Rockwell Laser Industries, Cincinnati, OH) was erected around the IR-MALDESI source at all times during the described experiments to protect other persons working in the laboratory. Additionally, OD 5+ rated protective goggles (Laser Safety Industries, Minneapolis, MN) were kept next to the potential exposure zone and provided to those required to be present in the area with the laser in operation. To minimize the risk of exposure to stray beams, the entire optical path before focusing optics was shielded by anodized aluminum protective screens (Thorlabs, Newton, NJ). All focusing optics were arranged in a protective 1-in. lens tube.

Measurements of ion accumulation and decay from solution

A 0.1-mg/mL solution of polyethylene glycol, average MW 600 Da (PEG-600) in 50:50 methanol/water, was analyzed directly from 100-μL microwell plates (Cat. No. 7816 20, Brand GMBH & CO KG, Wertheim, Germany) with the IR laser focused at the center of the meniscus in each sample well. The solution was aerosolized with a single laser shot synchronized to the trigger signal for MS acquisition, and the ion injection time instrument setting was changed between scans. In each experiment, IT was cycled between 1 and 1000 ms in 26 pre-defined steps with each IT setting measured between 40 and 50 times. The complete list of time points is included in the ESM as Table S1. These experiments were repeated for mass windows of 100, 200, 300, and 750 Th centered on the PEG oligomer ion of m/z 564.3596.

Measurements of ion yield from ablated animal tissue

The signal rise time following laser ablation was characterized by repeated measurements using 2-ms IT windows delayed in 1-ms increments by 0–25 ms after the ablating laser pulse. A 10-μm section of mouse liver was used as a model tissue. The analysis was otherwise performed as previously described for animal tissue imaging [7]. The average signal of each 2 ms bin was used as an estimate of momentary ion flux. An additional experiment was performed keeping all settings constant while cycling IT between 1 and 250 ms in 22 steps to measure actual ion accumulation at different times after the ablation event.

Method optimization for IR-MALDESI MS imaging

An injection time cycling experiment with 22 time points between 1 and 250 ms was performed on rat brain and mouse liver tissue samples, sectioned into 25- and 10-μm sections, respectively. For the rat brain, only visually homogeneous areas of gray matter were included in the optimization region of interest. To simulate the conditions of a typical imaging experiment, all tissue analysis was done in the mass range of 250–1000 Th, using a nominal resolving power of 140,000 at m/z 200. Two laser shots were used for sampling, even when the injection time was set too low (< 50 ms) to capture material from the second ablation event. The samples were kept between −10 and −8 °C throughout analysis, and an ice layer of controlled thickness was deposited according to previously published methodology [7].

After determining the optimal injection time, sections from the same tissues were immediately imaged. The sections were imaged in two different regions: one with the new optimized setting for low-latency handshake mode operation and one using a previously described procedure with the instrument in handshake mode using an IT of 110 ms [13]. Directly adjacent regions were imaged this way to allow direct visual comparison.

Data analysis

Abundances measured on the Q Exactive are normally reported in units of ion current, i.e., ion abundance per time unit. This is to ensure that reported abundances correspond to analyte concentration in a directly infused sample when AGC is used. In order to compare abundances from pulsed injections with AGC disabled, it is necessary to multiply the reported value of each scan by the injection time. Whenever abundances are reported in absolute terms throughout this article, the reported values have been multiplied by the injection time in seconds.

To generate imaging information, raw data was converted from .RAW format to .mzml using the MSConvert tool from the ProteoWizard toolkit [14] and subsequently from .mzml to .imzml using the imzMLConverter utility [15]. All ion images were generated using the open-source software package MSiReader [16]. All images here presented were generated with ± 2.5 ppm tolerance.


Ion overfill and signal decay

A solution of PEG-600 was used as a model system for ion accumulation and decay, providing a distinct and consistent series of oligomer peaks from each laser ablation to simulate the dense spectral population of lipid and metabolite peaks in a biological sample. The normalized ion abundance of the PEG oligomer ion at m/z 564.3596, at the center of each investigated mass range, was collected and converted to units of absolute abundance. A representative IR-MALDESI mass spectrum is provided as ESM Fig. S3. The average abundance plotted as a function of injection time for different mass ranges is shown in Fig. 1.
Fig. 1

Accumulation of m/z 564.3596 in four mass ranges. Acquired from direct analysis of 0.1 mg/mL PEG-600 using a single laser shot at t = 0. One hundred scans were collected at each IT setting. Side-by-side comparison illustrates the loss of signal over time observed for wider mass ranges

By allowing a larger range of ion masses to enter the C-trap for accumulation, it is possible to induce a loss of signal over time from ions injected at t = 0. The observed signal loss can be well described by an exponential decay model as shown in Fig. 2. The rate of decay is very similar where observed, and decay onset coincides with a total ion signal of approximately 3 × 106 AU, supporting the hypothesis of Rosen and coworkers that ion exclusion due to overfilling of the C-trap is the most significant cause of signal loss in imaging experiments [12]. The utility of gas-phase fractionation to increase sensitivity is seen in the measurements of a 100-Th mass range, where C-trap fill levels were not exceeded and no decay at all was observed.
Fig. 2

Comparison of ion accumulation curves for the PEG oligomer ion of m/z 564.3596 ± 2.5 ppm analyzed directly from 0.1 mg/mL PEG-600 solution. The curves include 95% confidence intervals for all measured points (n = 40) as well as an exponential fit to the data points colored in red. The dashed line indicates a peak ion fill of ~ 3 × 106 in the higher mass ranges. In the lower mass range of 564 ± 50 Th, total fill did not reach a peak value, and no significant decay was observed

A similar but more detailed experiment was performed on animal tissue samples, where the characteristic rates of signal rise was measured for a number of ions. Figure 3 shows a comparison of ion accumulation measured as the summation of momentary ion flux in 2 ms IT bins staggered over 25 ms, compared to accumulation data acquired using incremental IT synchronized to the ablation event. Data is shown for two tissue-specific lipids in the investigated range (200–800 Th). The lipids were putatively assigned as cholesterol [M+H+-H2O]+ (m/z 369.3516) and PC(36:4) [M+H+]+ (m/z 782.5702).
Fig. 3

Top: Estimated ion flux of two mouse liver tissue-specific ions in the 200–800 Th mass range. Each bar represents an average of 100 scans using a 2-ms IT window delayed from the laser pulse at t = 0 by 0–23 ms. The measured abundances have been normalized to sum to unity. Bottom: ion accumulation calculated as integrated flux (gray line) in comparison with direct measurement using the IT-cycling method (red line)

This experiment shows that there are very significant differences in the dynamics of ion injection between the two analytes. In this case, using an injection time of 5–10 ms would introduce a very significant systematic bias towards the faster rising species. This unexpected observation must be considered when selecting a suitable metric for method optimization. While reducing accumulation times is generally desirable as a means of minimizing accumulation of ambient background peaks, in so doing one may introduce a bias for fast-rising signals.

Optimizing settings for IR-MALDESI imaging

To demonstrate the utility of trapping time optimization in an imaging application, the signal of cholesterol [M-H2O+H+]+ at m/z 369.3516 from healthy rat brain was measured in a visually homogeneous region of gray matter. Injection time was cycled between 1 and 250 ms. The accumulation curve and signal comparison are shown in Fig. 4, suggesting an optimal cholesterol response at 70–80 ms after the first of two 20-Hz laser pulses. Based on this, a 75-ms IT was selected as an estimated optimum for comparison to previously established settings.
Fig. 4

Accumulation curves of target ions used to determine suitable imaging parameters. Top: cholesterol (m/z 369.3516 ± 2.5 ppm) abundance acquired from a 25-μm section of mouse brain using two laser pulses at 20 Hz. Error bars represent 95% confidence limits, based on 100 measurements at each time point. Bottom: IR-MALDESI images of cholesterol in rat brain acquired immediately following optimization from the same brain sample. Left side shows an optical image of a different section from the same individual rat for comparison. The ion images show a comparison between the optimized injection time of 75 ms and the previously used setting of 110 ms with the instrument in handshake mode

The imaging comparison in Fig. 4 shows an increase in target ion abundance by a factor of 2 with the optimized settings compared to the standard procedure for animal tissue samples (110 ms IT in handshake mode). The whole experiment for empirically selecting a suitable IT was performed in less than 30 min using less than 1 mm2 of representative sample, which allowed the imaging comparison to be performed on the same section of brain tissue under identical experimental conditions.

Given the previous observation of individual differences in optimal trapping time, it may be necessary to consider more than a single indicator ion to optimize an untargeted experiment. This was tested by simultaneously measuring multiple known tissue-correlated species. The accumulation curves of six lipid ions from IR-MALDESI analysis of mouse liver is shown in Fig. 5. Of particular note is that the second laser pulse in the sequence, at t = 50 ms, provides no significant signal increase for this sample. While rise times and decay rates vary significantly between analytes, there is an optimal region between 10 and 30 ms IT where most ions are present at > 75% of their peak abundance. The average normalized abundance has an apparent maximum at 10 ms IT, which was used for a comparison of imaging settings. The tissue was completely ablated by a single laser pulse, as confirmed by visual inspection after completed analysis. An inherent side benefit of reducing IT is the reduction of ambient peaks in the spectrum, as shown in Fig. 6. Reducing the collection of parasitic signal automatically improves the dynamic range of the experiment.
Fig. 5

Top: six known tissue-specific ions acquired simultaneously from a 10-μm section of mouse liver and normalized to unity. Total ion current (TIC) is included to illustrate the relationship between ion fill and ion decay from a pulsed injection. A second laser shot at t = 50 ms was used, which can be seen to provide no significant advantage in this case. Bottom: representative IR-MALDESI images comparing the optimized settings to the lab standard method

Fig. 6

Representative IR-MALDESI spectra of the 365–375 m/z range from mouse liver, with cholesterol peak annotated. With the lower injection time, a higher absolute abundance of the target peak is noted. The clear injection time dependence on abundance of peaks originating from the ambient background highlights the advantage of improved signal to background with lower injection time

Table 1 shows a comparison of average ion abundances acquired with either setting over the whole tissue. Two of the selected lipids with the lowest individual signal improvements, m/z 369.3516 and 300.2895 (putatively cholesterol [M-H2O+H+]+ and sphingosine [M+H+]+), have significantly longer rise times than the other selected ions, which emphasizes the importance of choosing suitable settings when designing untargeted experiments. As a rule of thumb for untargeted IR-MALDESI analysis, injection times may be set to include 25 ms after the last laser pulse required for complete ablation of the sample tissue.
Table 1

Average absolute abundance of tissue-specific ions before and after injection time optimization








10 ms IT







110 ms IT














Averages calculated over the whole tissue region as seen in Fig. 5


Precisely synchronizing the laser ablation event to the start of ion collection allowed a thorough investigation of when in the process of electrospray post-ionization the measured signal arises, including the surprising observation that there are very significant differences in rise and decay times for different lipids in the same tissue. Signal rise times were found to vary between 5 and 30 ms after ablation.

In a real experiment where reproducible quantitative sampling is desired, the rise time of the slowest rising target analyte effectively dictates the minimum acceptable accumulation time. When the C-trap is filled to higher levels than it is rated for, there is an accompanying loss of signal, which can be modeled as an exponential decay. This relationship between fill rate and signal loss supports the findings of Rosen and coworkers [12] regarding C-trap accumulation time dependence of signal and emphasizes the importance of minimizing ion trapping times.

An injection time cycling experiment was designed to find the best injection time settings for a typical imaging experiment. A comparison with experiments performed using previous standard operating procedure greatly favors the optimized timing settings and provides a significantly improved ratio of signal to background, owing to the overall reduced injection times. We recommend including this brief optimization step as a routine procedure for any imaging method using an ion trapping step, as it requires only a few minutes of data acquisition and a small amount of equivalent sample while potentially offering a significant increase in sensitivity. This is essential for applications of MS imaging within the fields of lipidomics and metabolomics, where a large number of analytes of very different abundances are analyzed simultaneously.



The authors wish to thank Prof. Troy Ghashghaei from the NCSU Department of Molecular Biomedical Sciences and Prof. Heather Patisaul of the NCSU Department of Biological Sciences for providing the mouse liver and rat brain tissues, respectively. Financial support for this work was received from the National Institutes of Health (R01GM087964) and North Carolina State University.

Compliance with ethical standards

The authors declare no conflicts of interest.

Animal tissue samples used were obtained from an internal repository of tissue from animals managed in accordance with the Institute for Laboratory Animal Research Guide. All husbandry practices were approved by North Carolina State University Institutional Animal Care and Use Committee (IACUC).

Supplementary material

216_2017_585_MOESM1_ESM.pdf (728 kb)
ESM 1 (PDF 728 kb)


  1. 1.
    Takats Z, Wiseman JM, Gologan B, Cooks RG. Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science. 2004;306:471–3.CrossRefGoogle Scholar
  2. 2.
    Chang DY, Lee CC, Shiea J. Detecting large biomolecules from high-salt solutions by fused-droplet electrospray ionization mass spectrometry. Anal Chem. 2002;74:2465–9.CrossRefGoogle Scholar
  3. 3.
    Mandal MK, Chen LC, Hashimoto Y, Yu Z, Hiraoka K. Detection of biomolecules from solutions with high concentration of salts using probe electrospray and nano-electrospray ionization mass spectrometry. Anal Methods-Uk. 2010;2:1905–12.CrossRefGoogle Scholar
  4. 4.
    Shiea J, Huang MZ, HSu HJ, Lee CY, Yuan CH, Beech I, et al. Electrospray-assisted laser desorption/ionization mass spectrometry for direct ambient analysis of solids. Rapid Commun Mass Spectrom. 2005;19:3701–4.CrossRefGoogle Scholar
  5. 5.
    Sampson JS, Hawkridge AM, Muddiman DC. Generation and detection of multiply-charged peptides and proteins by matrix-assisted laser desorption electrospray ionization (MALDESI) Fourier transform ion cyclotron resonance mass spectrometry. J Am Soc Mass Spectrom. 2006;17:1712–6.CrossRefGoogle Scholar
  6. 6.
    Dixon RB, Muddiman DC. Study of the ionization mechanism in hybrid laser based desorption techniques. Analyst. 2010;135:880–2.CrossRefGoogle Scholar
  7. 7.
    Robichaud G, Barry JA, Muddiman DC. IR-MALDESI mass spectrometry imaging of biological tissue sections using ice as a matrix. J Am Soc Mass Spectrom. 2014;25:319–28.CrossRefGoogle Scholar
  8. 8.
    Barry JA, Robichaud G, Bokhart MT, Thompson C, Sykes C, Kashuba ADM, et al. Mapping antiretroviral drugs in tissue by IR-MALDESI MSI coupled to the Q exactive and comparison with LC-MS/MS SRM assay. J Am Soc Mass Spectrom. 2014;25:2038–47.CrossRefGoogle Scholar
  9. 9.
    Makarov A, Denisov E, Kholomeev A, Baischun W, Lange O, Strupat K, et al. Performance evaluation of a hybrid linear ion trap/orbitrap mass spectrometer. Anal Chem. 2006;78:2113–20.CrossRefGoogle Scholar
  10. 10.
    Olsen JV, de Godoy LMF, Li GQ, Macek B, Mortensen P, Pesch R, et al. Parts per million mass accuracy on an orbitrap mass spectrometer via lock mass injection into a C-trap. Mol Cell Proteomics. 2005;4:2010–21.CrossRefGoogle Scholar
  11. 11.
    Bokhart MT, Rosen E, Thompson C, Sykes C, Kashuba ADM, Muddiman DC. Quantitative mass spectrometry imaging of emtricitabine in cervical tissue model using infrared matrix-assisted laser desorption electrospray ionization. Anal Bioanal Chem. 2015;407:2073–84.CrossRefGoogle Scholar
  12. 12.
    Rosen EP, Bokhart MT, Nazari M, Muddiman DC. Influence of C-trap ion accumulation time on the detectability of analytes in IR-MALDESI MSI. Anal Chem. 2015;87:10483–90.CrossRefGoogle Scholar
  13. 13.
    Nazari M, Muddiman DC. Polarity switching mass spectrometry imaging of healthy and cancerous hen ovarian tissue sections by infrared matrix-assisted laser desorption electrospray ionization (IR-MALDESI). Analyst. 2016;141:595–605.CrossRefGoogle Scholar
  14. 14.
    Chambers MC, Maclean B, Burke R, Amodei D, Ruderman DL, Neumann S, et al. A cross-platform toolkit for mass spectrometry and proteomics. Nat Biotechnol. 2012;30:918–20.CrossRefGoogle Scholar
  15. 15.
    Race AM, Styles IB, Bunch J. Inclusive sharing of mass spectrometry imaging data requires a converter for all. J Proteome. 2012;75:5111–2.CrossRefGoogle Scholar
  16. 16.
    Robichaud G, Garrard KP, Barry JA, Muddiman DC. MSiReader: an open-source interface to view and analyze high resolving power MS imaging files on Matlab platform. J Am Soc Mass Spectrom. 2013;24:718–21.CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  1. 1.W. M. Keck FTMS Laboratory for Human Health Research, Department of ChemistryNorth Carolina State UniversityRaleighUSA

Personalised recommendations