Exploring Approaches for Accelerated Mass Spectrometric Detection of Fluorochemicals from Aqueous Solutions Using Indirect, Adsorbent-Facilitated Desorption Ionization

Abstract

There is a growing need for rapid, low-cost analytical tools to screen water supplies for persistent contaminants. Due to their ubiquitous presence in the environment and their detrimental effects on human health, there is a public interest in widespread testing for perfluoroalkyl substances (PFAS). However, the most common methods for PFAS analysis can be time-intensive, requiring a preconcentration step followed by liquid chromatography–mass spectrometry (LC–MS), and are cost-prohibitive for most members of the public. To address these challenges, this study explores a novel method for rapid analysis of aqueous analytes by concentrating analytes on solid adsorbents and detecting them using mass spectrometric analysis via desorption ionization directly from the adsorbent surface. We demonstrate this general approach of indirect solution analysis (ISA) with desorption electrospray ionization (ISA-DESI) and laser desorption ionization (ISA-LDI). Various adsorbent materials were screened, including commercial mesoporous carbons, alumina, and metal–organic frameworks (MOFs). Using ISA-DESI, we consistently detect concentrations as low as 5–100 parts per trillion (ppt) for perfluorooctanoic acid (PFOA) and 10–100 ppt for perfluoroundecanoic acid (PFUnDA). This approach enables rapid detection in under 5 min, making it suitable for applications requiring trace detection with fast response times. We further demonstrate the use of ISA-DESI on a range of environmentally relevant contaminants. Advantages of this indirect solution analysis approach include simplified sample preparation, accelerated analysis, and low cost with readily available materials, making it a promising tool to accelerate mass spectrometric PFAS detection for identifying contaminated sites, performing laboratory studies, and monitoring industrial waste streams.

Introduction

There is a growing demand for rapid and low-cost access to analytical testing for persistent environmental pollutants in public and private water systems. This demand is being driven by increased awareness and concern regarding per- and polyfluoroalkyl substances (PFAS), which are a class of anthropogenic compounds comprised of a fluoroalkyl chain. Due to the chemical stability of the carbon–fluorine bond, PFAS are persistent in the environment and can take up to 1000 years to degrade under ambient environmental conditions. PFAS continue to see widespread use in both industrial and military applications, despite some species showing toxicity at parts-per-trillion (ppt) levels. , Detection of PFAS in groundwater has been observed across the USA, including several sites at parts-per-million (ppm) levels, , contributing to increased regulation and public awareness of these compounds. ,

Due to the widespread concern associated with PFAS exposure, there has been significant research focused on increasing throughput and lowering cost for PFAS testing. However, commonly used methods for targeted testing can be time-consuming for sample preparation and costly to R&D scientists and the general public. For example, depending on sample composition, a single sample can take anywhere from 20 min to several days using the solid phase extraction (SPE), liquid chromatography (LC)- mass spectrometry (MS) method in EPA method 537 or 1633, and can cost more than 100 USD per sample to send for analysis, which can be cost-prohibitive for broad applicability and wide-scale testing. , Lower cost and higher throughput would benefit the broader public as well as scientific communities. For example, lower costs can increase access to testing, which in turn permits small agricultural businesses who want to ensure the safety of their goods easier access to testing. Meanwhile, increased throughput can benefit R&D scientists collecting degradation kinetics of target PFAS in a laboratory setting. −

Desorption ionization methods are promising approaches for high-throughput, targeted mass spectrometric screening because spectra can be collected quickly by scanning spatially across multiple samples. Methods such as matrix-assisted laser desorption ionization (MALDI)-MS or its variants, including surface-assisted laser desorption ionization (SALDI)-MS, , have been explored for their application in screening for perfluorinated compounds. , However, MALDI and SALDI require specialized sample preparation to analyze from solution (i.e., the addition of a carefully selected matrix), which lowers throughput, raises costs for each sample analysis, and reduces access. , Another possibility is desorption electrospray ionization mass spectrometry (DESI-MS), which offers a rapid analysis of either a single point or a 1D or 2D distribution across a surface. − Indeed, some PFAS compounds have been identified and their distribution across a surface has been visualized with this method. However, DESI is typically used on a solid surface rather than to analyze trace aqueous analytes.

Here, we report on preliminary testing of indirect solution analysis (ISA) using adsorbent-enhanced desorption ionization MS approaches (Figure ). We demonstrate this approach using several common adsorbents to sequester analytes from solution for analysis by surface desorption ionization methods. Perfluorooctanoic acid (PFOA) was selected as the primary analyte due to its widespread environmental presence, significant health concerns, and regulatory focus. Additionally, to demonstrate sensitive detection for a broader range of PFAS, perfluoroundecanoic acid (PFUnDA) was also used as a primary analyte. The adsorbents preconcentrate the target analyte for rapid screening using desorption ionization methods. We test both laser desorption ionization (LDI) and desorption electrospray ionization (DESI). There is no need to desorb analytes from the adsorbents prior to analysis nor mix with a matrix molecule, which greatly simplifies sample preparation. Further, these methods can be applied directly to adsorbents used at wastewater treatment facilities for the physical removal of contaminants. This approach still relies on access to mass spectrometry but accelerates sample preparation and lowers costs relative to established methods. Using ISA-DESI, we demonstrate ppt-level detection limits in laboratory samples, demonstrating it is a promising method for rapid, trace detection of PFAS. Overall, the advantages include (1) sample preparation is greatly simplified relative to conventional SPE methods, (2) it is an approach for using surface desorption ionization for trace aqueous analytes from a liquid solution, and (3) it works with low-cost adsorbent materials.

1.

Graphical overview of ISA-DESI analysis procedure, illustrated here for the rapid-detection process demonstrated with alumina, emphasizing the rapid nature of this analysis method and the PFOA chemical structure demonstrating where the [M – H]⁻ peaks (purple line) and [M – CO₂H]⁻ peaks arise from and their respective isotopic modeling for these expected PFOA ions at m/z = 412.97 and 368.98.

Experimental Section

Materials

A range of adsorbents were selected to represent a variety of material types that are widely available and inexpensive. Metal organic frameworks (MOFs) were also investigated on a limited basis based on their prevalence in literature surrounding PFAS sequestration. − All non-MOF adsorbents were sourced from commercial suppliers. These included mesoporous carbon CMK-3 (ACS materials), disordered carbon (DC) (ACS materials), molecular sieves (3 Å, 8–12 mesh, Fisher Sci), and activated alumina (8 – 14 mesh, Fisher Sci). Both MOF materials (MOF-808 and NU-1000) were synthesized using protocols found in literature. − A table of material properties can be found in the Supporting Information (Table S1). Stock solutions of PFOA (Strem catalog) and PFUnDA (>95%, Sigma-Aldrich) was prepared in methanol (Avantor Performance Materials). Working solutions of PFAS were prepared by diluting these stocks in 18 MΩ water. Aqueous solutions of ibuprofen (99%, Acros Organics), Eriochrome black T (indicator grade, Sigma-Aldrich), and triethylamine (99.5%, Sigma-Aldrich) were prepared in 18 MΩ water.

Sample Preparation for Initial Screening of Adsorbents for ISA

Initially, a slower and more methodical analytical approach was taken to screen for the most promising adsorbents and determine the optimal experimental parameters. Following this, a revised approach was adopted to optimize analysis time (see Experimental, Demonstrating Rapid Use of ISA-DESI). The fastest recorded analysis, from receiving a fresh solution to finishing collecting data, was 3.5 min total using a 500-ppb solution.

For all indirect solution analyses, 5 mg of fresh adsorbent was measured by weight and added to a 50 mL polyethylene tube with 50 mL of PFOA solution. Typically, adsorbent-solution mixtures were equilibrated for 1 h with agitation. Shorter equilibration times were used when demonstrated to be effective (e.g., 100 ppb PFOA on alumina pellets as shown in Figure S1). For powdered adsorbents, equilibrated samples were centrifuged (∼20,000 RCF), resulting in adsorbent separation from the solution, and the PFOA solution was decanted. The solid residue was then dried at 50 °C until visually dry.

ISA-DESI

All adsorbents were tested in ISA-DESI. For initial screening of adsorbents, the materials were fixed to a substrate (glass slide) using double-sided tape. Powders required a drying step to allow proper adhesion to the tape. Powders were dried at 70 °C for 12 h. For pellets (alumina), two methods of drying were tested. (1) The pellets were dried at 70 °C for a minimum of 12 h, or (2) excess liquid was removed immediately by mechanical action (i.e., lightly wicking off excess liquid with a fresh Kimwipe). As shown in Figures S2 and S3, there was no difference in signal depending on how the adsorbent was dried. Based on this outcome, the more rapid mechanical drying was adopted for pellets.

ISA-LDI

For ISA-LDI, only powder adsorbents were tested, including mesoporous carbons CMK-3 and DC, and a metal organic framework (MOF) MOF-808. Drying was performed at 70 °C for 12 h. The dried adsorbents were either adhered with tape or drop-cast onto individual wells on a commercial MALDI plate. Drop-casting was performed with either an acetone suspension, a water suspension, or an acetone-suspension with a concentration of 10 mg mL^–1 of MALDI matrix (1,5-diaminonaphthalene). The plated material was dried under ambient laboratory conditions (∼22 °C, ∼20% relative humidity).

Instrumental Methods for ISA

The ISA-DESI and ISA-LDI experiments were performed using a Waters quadrupole time-of-flight (QToF) mass spectrometer, Synapt HDMS XS (Waters Corporation, MA, USA), with an ionization interchangeable source to switch between the MALDI and DESI XS High-Performance Ion Sources. Data acquisition was in MassLynx Version 4.2 (Waters Corporation, MA, USA). While instrument parameters varied for optimization specific to PFAS and adsorbent material, typical parameters for ISA-LDI and ISA-DESI are included in the following sections.

ISA-DESI

ISA-DESI samples on a glass slide were loaded onto a two-dimensional stage and analyzed in negative ion mode. Typical settings on the DESI XS source were: capillary voltage 0.16 kV, cone voltage 35 V, source temperature 150 °C, gas pressure 0.005 MPa, heated transfer line 350 °C, and spray solvent composition of 98:2 methanol/water with 0.1% formic acid. Depending on the adsorbent, optimal capillary voltage ranged from 0.16 to 32 kV and cone voltage ranged from 20 to 44 V. Positioning of the sample stage and DESI spray source was optimized to maximize the PFOA ion signal for each surface depending on adsorbent geometry. MS spectra were collected in resolution mode from 50 m/z to 1200 m/z with a scan rate of 0.5 s per scan. The average background signal from the blanks was subtracted from raw signals, and then converted to signal-to-noise using the average baseline noise ± 0.15 m/z on either side of the peak of interest.

For MS–MS analysis, the following parameters were used: capillary voltage 0.28 kV, sampling cone voltage 10 V, and fragmentation collision energy (CE) 14.5. MS–MS analysis of PFOA was performed on the parent peak (m/z = 412.97) to detect the major fragment ion (m/z = 368.98) and spectra were collected in the range of 50–500 m/z. MS–MS analysis of PFUnDA was performed on the parent peak (m/z = 562.95) to detect the major fragment ion (m/z = 518.94) and spectra were collected in the range of 50–600 m/z.

ISA-DESI (Alumina)

For use with alumina pellets, which are amorphous in shape, it was necessary to adjust the height of the inlet capillary so that the center of the inlet was level with the top of the pellet, compensating for the irregular geometry and added height of the pellet. Initial tests with alumina indicated there was significant run-to-run carryover with this geometry, particularly at high concentrations (>100 ppb) likely due to splashing of the DESI spray off the irregular facets of the alumina onto the capillary. Thus, to accurately test the performance of this material, the capillary was rinsed in between trials using a squirt of methanol. This carryover effect was not observed for flat surfaces, so no capillary rinse was necessary for any materials other than alumina, suggesting this step can be omitted when using flat, pelleted alumina.

ISA-LDI

The laser source for ISA-LDI was a frequency-tripled Nd:YAG laser (355 nm, pulsed at ≥100 μJ/pulse) with a pulse frequency of 2 Hz during MS data collection. Spectra were collected in negative ion mode and Waters resolution mode from 50 m/z to 1200 m/z with a scan rate of 0.5 s per scan.

Data Analysis

For a single ISA-DESI run, data was typically averaged for 10 s after the maximum ion signal when the DESI spray was directed at the adsorbent, as indicated in Figure , unless otherwise indicated. Repeat measurements were then averaged (±1 SD) weighting each data point equally. For a single ISA-LDI run, 10 s of data encapsulating the areas with the largest increases in ion count were averaged ±1 SD.

2.

ISA-DESI analysis of alumina showing (a,b) ease of sampling and handling, (c) clear mass spectral signal from 1 ppm PFOA initial solution, and (d) both a rapid response time and long-lived PFOA signal. The time where data collection was initiated and stage was moving are indicated in the plot. Signal appears as soon as the DESI spray contacts the adsorbent at ∼0.2 min. Maximum signal occurred when the spray was centered on the sample. The signal persists even when the stage is not moving. This long-lived signal facilitates data analysis and collection.

Demonstrating Rapid Use of ISA-DESI

Since DESI is an ambient technique that does not require vacuum for sample introduction, it was selected to demonstrate the utility of the ISA technique for rapid analysis. For reasons discussed later, alumina was used as the fastest effective adsorbent. The final method selected for ISA-DESI began with receiving a solution, followed by submerging a pellet of alumina in the solution for 3 min. The analyte-laden pellet was then removed from the solution using tweezers, gently dried using a Kimwipe, and placed in front of the DESI ion source for analysis where data collection was initiated. For more details on selecting these parameters, see the Results and Discussion.

Results and Discussion

PFOA was the principal analyte used as proof-of-principle for the indirect solution analysis method using readily available adsorbents. As shown in Figure , PFOA under the conditions used in this study was associated with two ions: m/z = 412.97, corresponding to deprotonated PFOA, and m/z = 368.98, a fragment arising from the loss of the carboxylic acid headgroup. Both ions were observed simultaneously at varying respective intensities depending on the energy or voltage applied to the sample.

ISA-DESI

Desorption electrospray ionization (DESI) is an ionization method for surface analysis, typically used on solid samples. Here, we use DESI for indirect solution analysis (ISA) to detect PFOA in aqueous solutions by sequestering the analyte onto a adsorbent surface and extracting the solid for analysis. We demonstrate the ISA-DESI approach in detail here using alumina as an example. As illustrated in Figure , where a 1 ppm PFOA solution was concentrated onto alumina pellets (Figure a,b), when the DESI ion source was operated over the PFOA-laden adsorbent, a clear PFOA signal was observed (Figure c). The PFOA signal persisted for many seconds, as shown in Figure d. In contrast, no signal was detected when 1 μL of the same 1 ppm solution was drop-cast and dried on a DESI substrate (Figure S4). Additional analysis comparing drop casting to ISA-DESI is discussed in the following section (Screening the Performance of Other Adsorbents in ISA-DESI). This demonstrates that the ISA-DESI approach amplifies analyte signal relative to direct solution analysis and that the adsorption of PFOA onto an adsorbent facilitates the use of DESI for PFOA detection.

Using MS–MS analysis of the parent peak (m/z = 412.97) and its major fragment ion (m/z = 368.98), ISA-DESI successfully detected signal above the blank from 10 ppt solutions for PFOA on alumina (Figure , with raw data shown in Figure S5). This represents a significant first step for this technique and highlights the potential of this method for ultratrace analysis. The calculated limit of detection (LOD) is 5–100 ppt, depending on how the LOD was calculated. Using a log–log plot of log­(signal) as a function of log­([PFOA]), which yields an extensive linear relationship from 1 ppm to 10 ppt (R² = 0.990), the LOD is estimated to be ∼5 ppt (Figure S6a) whereas a linear calibration curve treatment indicates an LOD of ∼100 ppt when calibrating to the lowest concentration standards (Figure S6b). The observed nonlinear (log–log) relationship between signal-to-noise ratio and concentration may stem from the nonlinear behavior of the adsorption process, uncharacterized desorption processes specific to DESI analysis on porous adsorbents, or a combination of these factors (see the Supporting Information for additional discussion). , The limiting factors preventing even lower LOD is attributed to residual contamination (see Experimental, Instrumental Methods for ISA, ISA-DESI Alumina) and the inconsistency of the total signal due to morphological variations from experiment to experiment. This suggests that pellets with flattened, regular surfaces may improve the LOD. The consistent increase in signal over the blank at 10 ppt underscores the robustness of this method, even at concentrations approaching the lower limits of detection. Future efforts to address residual contamination and optimize adsorbent materials could further improve sensitivity and reduce noise, enabling even lower detection limits.

3.

ISA-DESI analysis of PFOA on alumina (mechanically dried, 1 min MS/MS collection, average triplicate experiments (±1 SD), background subtracted) at decreasing initial concentrations of PFOA, showing an increase in signals over the blank from samples at 10 ppt and above.

Detection of PFUnDA was also demonstrated using MS–MS analysis of the parent peak (m/z = 562.95) and its major fragment ion (m/z = 518.94). ISA-DESI confidently detected signal above the blank from 100 ppt solutions for PFUnDA on alumina (Figure S7). The LOD for PFUnDA was calculated in the same manner as for PFOA and was determined to be 10–100 ppt, again depending on the calculation method (Figure S8). These results reinforce the efficacy of the ISA-DESI technique for detecting PFAS species and demonstrate its potential for broader applications in ultratrace environmental analysis. Future work will focus on further optimizing adsorbent materials and morphology and further minimizing contamination sources to improve detection limits for both PFOA and PFUnDA.

With these LODs, ISA-DESI technique could be applied in laboratory conditions to detect anything above the EPA limit (e.g., EPA limit for PFOA is 4 ppt). , Additionally, where it is applicable this method is estimated to reduce costs for analysis of a single sample by 46% compared to traditional analysis as a result of reduced costs related to consumables, steps, and time (Table S2). This cost reduction could be further amplified as automated sampling across many samples (e.g.,100 samples) on a single plate with a sampling rate of mere seconds is possible with current systems. Notably, the ISA-DESI approach using standard laboratory adsorbents is a significant cost savings over conventional SPE cartridges, regardless of prep time, although reduced labor costs is another expected advantage.

The signal lifetime for PFOA decreased with decreasing solution concentration, lasting several minutes for 1 ppm samples, as shown in Figure d, less than a minute for 100 ppb samples, and only a few seconds for 100 ppt samples before decaying to background levels (Figure S9). This reduction in signal lifetime at ultratrace concentrations is attributed to the depletion of local PFOA from the adsorbent surface during electrospray analysis, an effect observed broadly in DESI. , Moving to a new spot sufficiently far away from the initial location, typically at least two spot diameters (∼1 mm), resets the signal by targeting a fresh region of adsorbate. While this approach allows continued analysis at lower concentrations, it is recommended to initiate data collection prior to spraying the adsorbent surface with the DESI source to maximize signal stability. These findings highlight the importance of optimizing adsorbent materials and electrospray parameters to maintain reliable signal at ultratrace levels and further improve the sensitivity of ISA-DESI.

Screening the Performance of Other Adsorbents in ISA-DESI

ISA-DESI analysis was further tested on different powdered adsorbents. Figure shows the ISA-DESI signal-to-noise ratio for a range of different common laboratory adsorbents that were used to sequester PFOA from 100 ppm solutions. Even at high concentrations some adsorbents did not perform well; e.g., a 3 Å molecular sieve produced a signal similar to the background levels observed in the unaltered adhesive used to secure samples to the glass slide. This is likely due to the small pore size of the material which may allow only a trivial quantity of PFOA to be adsorbed. The signal-to-noise ratio for 100 ppm PFOA drop-cast onto the slide was nearly zero. The results of the tape and glass slide emphasize the impact of preconcentration onto an adsorbent for DESI analysis.

4.

A comparison of ISA-DESI analysis of PFOA on adsorbents (12 h drying, 10 s MS collection times, average of triplicate experiments (±1 SD), background-subtracted) subjected to a 50 mL, 100 ppm PFOA solution.

Mesoporous carbon materials are advantageous due to their low cost, availability, and safety, but did not prove as effective as alumina in this initial survey. Both CMK-3 and disordered carbon performed around 1 order of magnitude lower than alumina at 100 ppm. Further, the powdered form made handling more difficult and increased the possibility of particles becoming dislodged and introduced into the system, potentially blocking the capillary. As demonstrated in Figure , PFOA was detectable to at least 1 ppm on CMK-3. In this brief survey, signal strength diminished greatly between 10 and 1 ppm, where the signal strength approaches that of the blank.

5.

(a) Porous carbon CMK-3 exposed to solutions of PFOA. (b) ISA-DESI imaging analysis of CMK-3 subjected to decreasing initial concentrations of PFOA, still showing spatial resolution (dark blue regions) at 1 ppm. (c) Interrogating the performance of CMK-3 for ISA-DESI analysis at reduced concentrations of PFOA.

Among the tested adsorbents, MOF-808 alone demonstrated slightly better sensitivity at comparable concentrations than alumina. While the raw signal for PFOA on MOF-808 was more than an order of magnitude higher than on alumina, the signal-to-noise ratio proved to be only marginally better. It is suspected that this may, in part, be a result of the powdered form which allows the electrospray solvent to also interact with adhesive between particles. Regardless of culpability in increasing noise, the powdered form introduced challenges to handling like the porous carbons.

In contrast, while the signal-to-noise ratio was slightly lower for alumina, this material provides several practical advantages over MOFs. Alumina is readily available, affordable, and often produced as large (∼1–2 cm), insoluble particles, making it easy to handle and analyze. Further, alumina pellets can be manufactured with smooth surfaces, which likely will improve signal consistency during analysis compared to the amorphous particles used in this work. These benefits led to the decision to focus the bulk of this study on alumina as the primary adsorbent. Still, even well below subppm levels, both MOFs maintain clearly resolved signal, as can be seen in Figure S10. These findings suggest that while MOFs may be a promising candidate for applications requiring low detection limits, alumina remains the most practical and effective adsorbent for this initial ISA-DESI work.

Differences in sensitivity from one adsorbent to another are likely to be strongly related to the adsorbate–adsorbent interaction potential as well as other relevant adsorption mechanisms (e.g., adsorbate–adsorbate stabilization due to mesopore filling). Acquiring relevant adsorption isotherms for the system being studied could greatly assist in optimizing adsorbent selection for any particular analyte. However, collecting isotherms for every available adsorbate–adsorbent may be impractical. In such instances, inferences based on available literature or experimental data can be used, as was the case with this work on PFOA. , Additionally, it is possible that during the electrospray process, charge transfer to the interior of the porous structure may be better facilitated by the electronic structure of some adsorbents relative to others. Regardless, it is likely that for application to other PFAS or other analytes, the adsorbent and mass spectrometer settings will require being tuned on an individual basis.

ISA-DESI for Other Environmental Contaminants

ISA-DESI analysis was further tested on a range of environmental contaminants to demonstrate the broader applicability of this technique. In addition to PFOA, these included other PFAS compounds and representative species from textile dyes, pharmaceuticals and personal care products, and industrially relevant species.

As discussed earlier, perfluoroundecanoic acid (PFUnDA) was tested using the ISA-DESI technique as shown in Figure S7 and was shown to have a LOD of 10 ppt based on the logarithmic treatment of the calibration curve. In addition to fluorinated alkyl chains with carboxylic headgroups, perfluorobutanesulfonic acid (PFBS) was tested with this technique. As shown in Figure S11, ISA-DESI on alumina subjected to a 500 ppb PFBS solution produces a strong signal, positively identifying it at m/z = 298.7. This demonstrates the versatility of this approach for PFAS species with different chain lengths and functional groups.

Beyond PFAS, there are many categories of environmental contaminants of concern that could benefit from this rapid analysis, such as industrial chemicals, pharmaceuticals, and textile dyes. For example, Eriochrome black T (EBT), a common textile dye and industrially significant species, produced a strong signal at 500 ppb initial concentration (Figure S12).

Ibuprofen, representing pharmaceuticals and personal care products, was positively identified at a 500 ppb initial concentration also. The signal was not as strong as it was for EBT but was still distinct, as seen in Figure S13. It is possible that a different electrospray solvent or adsorbent may improve signal here, as the methanol–water mixture used in this study was primarily optimized for PFAS species and alumina is considered a polar adsorbent and therefore may not have adsorbed much of the very nonpolar ibuprofen. An industrially significant and highly volatile species was also selected for analysis, triethylamine (TEA), and was positively identified (Figure S14). Further development will be required to understand the sensitivity of ISA-DESI to the diverse range of chemicals found as environmental contaminants and how chemical properties, e.g., volatility, influence sampling. However, application of ISA-DESI to such a broad range of target analytes demonstrates the method’s promise as a technique for broad environmental and industrial monitoring.

ISA-LDI

Given the ease of use of ISA-DESI, we extended the concept of indirect solution analysis to another surface-desorption ionization method using laser desorption analysis. As shown in Figure a,b using mesoporous carbon CMK-3 as the adsorbent, ISA-LDI can be used to detect PFOA. Figure b shows the time-dependent ion intensity as the laser is scanned across the LDI substrate. Here, the solution concentration was 200 ppm PFOA. The MOF materials were also tested for the ISA-LDI technique at initial PFOA concentrations of 100 ppm (Figure c,d). Powders were adhered to the sample plate using double sided tape to mitigate loss. Unlike DESI, the laser in LDI must be continuously scanning because the sample was ablated by the laser, depleting the region of analyte with each laser pulse. This method thus produced a less predictable and stable signal relative to ISA-DESI, as demonstrated by the time series in Figure b,d, which increased the time scale of analysis. For example, as seen if Figure b, significant PFOA signal is not seen until 0.4 min. In contrast, the ISA-DESI signal appears quickly after the spray is directed onto the adsorbent and can last extended periods of time depending on concentration as seen in Figure S9. Although these time differences are small for a single analysis, this added time can significantly reduce automated throughput with large numbers of samples. Further, the LDI signal is less consistent, which may complicate quantification since it is unclear what bursts of intensity should be used as the representative ion signal. This inconsistency may be due to inhomogeneities in the sample’s distribution or ability to absorb the laser.

6.

ISA-LDI analysis of PFOA on mesoporous carbon CMK-3 (green, a and b) and MOF-808 (orange, c and d). The mass spectra (a and c) show a much noisier system compared to ISA-DESI and time series associated with the intensity at 368.98 m/z (b and c) reflect the less stable signal of ISA-LDI relative to ISA-DESI, as seen in (d) where ISA-DESI analysis of PFOA MOF-808 (blue, dashed) is overlaid to highlight the relative signal stability.

Decreasing the initial concentration of PFOA down to 10 and 1 ppm resulted in no positive identification of PFOA (Figure S15), demonstrating that although the method works for high concentrations, it lacks sensitivity. The low sensitivity of ISA-LDI was likely due to the simultaneous ablation of the adsorbent which produces an abundance of background ions that generate noise and interfered with the generated PFOA ions, reducing analyte signal. This background noise can be clearly seen in Figure and juxtaposed against the low noise generated during ISA-DESI analysis.

7.

Background signal observed with CMK-3 on double-sided scotch tape in ISA-LDI (top) compared to ISA-DESI (bottom) showing a major advantage to the ISA-DESI approach for direct analysis of adsorbents.

Ultimately, due to high noise, low signal intensity, inconsistent signal compared to ISA-DESI, longer analysis time than ISA-DESI, and dispersion of adsorbent in the MALDI chamber by ablation, it was determined that for indirect solution analysis of PFOA, the ISA-LDI technique was not as effective as the ISA-DESI technique in its current form and no further testing was performed. This testing did, however, still demonstrate that ISA-LDI may be effective at high PFOA initial concentrations, providing an alternate route of analysis for individuals with access to a MALDI source but no DESI source. The challenges noted with ISA-LDI are not insurmountable but would require more development. For example, different adsorbents or laser sources could be explored further to reduce background noise and ultimately improve detection limits for ISA-LDI.

Concluding Remarks

There is a growing need to accelerate PFAS analysis to enable widespread, low-cost testing for environmental contaminants. This study presents a novel application of surface desorption techniques, specifically ISA-DESI, to analyze aqueous analytes from adsorbent surfaces. ISA-DESI demonstrated exceptional sensitivity, achieving consistent signal above the blank at ppt-level concentrations for both PFOA and PFUnDA. Additionally, it was demonstrated that this technique can be applied to many other environmental contaminants.

While ISA-DESI demonstrated high sensitivity in laboratory generated samples, further characterization with multicomponent solutions is required to understand how LODs for target analytes might be affected by other species. Residual contamination is a limiting factor for achieving even lower detection limits. Future work will focus on mitigating cross contamination by optimizing adsorbent morphology, i.e., exploring custom, flat pellets as these pellets are expected to reduce spray being redirected from the adsorbent surface toward the inlet capillary. Future work can also explore additional adsorbent materials with higher adsorption capacity to improve sensitivity.

Compared to conventional methods such as SPE/LC–MS, ISA-DESI offers significant advantages in throughput, simplicity, and cost. The sample preparation process is straightforward, does not require a desorption step, and allows for automated mass spectrometric analysis of large numbers of samples. For example, ISA-DESI can screen 100 samples on a standard DESI plate within minutes, whereas liquid injections or LC–MS runs may take significantly longer. Additionally, the estimated cost reduction for single-sample analysis is approximately 46%, with further savings possible through automated sampling. The low cost and simplicity of ISA-DESI, along with the demonstrated sensitivity on fast time scales, highlights ISA-DESI as a promising tool for rapid, ultratrace analysis and screening of PFAS and other environmental pollutants.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c08128.

• Material properties (Table S1); ISA-DESI signal comparison at variable equilibration time scales (Figure S1); comparison of ISA-DESI analysis with different adsorbent drying processes (Figures S2, S3); comparison of PFOA signal adsorbed on alumina vs dropcast onto a slide (Figure S4); unprocessed data from Figure (Figure S5); calibration curve plots (Figures S6, S8); extended discussion of LOD calculations; PFUnDA ISA-DESI analysis (Figure S7); cost comparison (Table S2); ISA-DESI time series (Figure S9); ISA-DESI on MOFs (Figure S10); ISA-DESI of PFBS; ISA-DESI of non-PFAS analytes (Figures S11–S14); ISA-LDI analysis (Figure S15) (PDF)

¥.

Oak Ridge National Laboratory, Oak Ridge, TN 37830 United States

$.

Critical Materials LLC, Albuquerque, NM 87123 United States

#.

Perspective Therapeutics, Inc., Seattle, WA, 98121 United States

The authors declare the following competing financial interest(s): A related patent application is in process where coauthors R.D.D., N.R.B., S.M.K., and B.A.B. are named inventors. There are no other conflicts to declare.