Detrimental impact of aqueous mobile phases in 18F-labelled radiopharmaceutical analysis via radio-TLC

1. INTRODUCTION

Positron-emission tomography (PET) is a non-invasive molecular imaging technique that harnesses radiopharmaceuticals to quantify biochemical processes in vivo. The radiopharmaceutical (or tracer) is a bioactive molecule labelled with a short-lived positron-emitting radionuclide. The most commonly used radionuclide is fluorine-18 due to its favorable physical and chemical properties.^1,2 Currently, most PET scans measure glucose metabolism with the radiopharmaceutical 2-[¹⁸F]fluoro-2-deoxy-D-glucose ([¹⁸F]FDG) to diagnose a myriad of diseases. However, additional types of scans may become more prevalent as several new PET tracers that are more specifically targeted to disease phenotypes have recently garnered clinical approval, including those targeting amyloid plaques (Neuraceq, Amyvid, Vizamyl), dysfunctional tau protein (Tauvid), prostate cancer (Axumin, Pylarify), and Parkinson’s disease ([¹⁸F]FDOPA).³ The list of new ¹⁸F-labeled tracers under development also grows with the discovery of new biological targets and therapeutic strategies.⁴

The successful development and production of PET tracers rely on analytical techniques such as radio-high performance liquid chromatography (radio-HPLC) and radio-thin layer chromatography (radio-TLC) to assess radiochemical conversion (during radiosynthesis development) or radiochemical purity (during quality control testing of tracers produced for clinical use). A shortcoming of radio-HPLC, however, is the propensity for retention of free [¹⁸F]fluoride in the column (i.e., not reaching the detectors), which can lead to the underestimation of this species in the output chromatogram.⁵ In contrast, radio-TLC does not suffer from this issue as the entirety of the TLC plate is scanned.

When using typical silica TLC plates to separate ¹⁸F-labelled mixtures, [¹⁸F]fluoride is usually sequestered near the origin through strong interaction with surface silanol groups, and the mobile phase moves the radiopharmaceutical away from the origin. The stationary phase, silica gel (polysilicic acid), is well known in the literature to possess the ability for acidic hydrogen bonding, basic hydrogen bonding, and dipolar interactions with analytes⁶, and the mobile phase plays an important role in modulating these interactions. The reported mobile phases vary widely in literature, even for analysis of the same compound (Table 1), and we noted with interest that many studies use water as a polar mobile phase additive to enhance the migration of polar radiopharmaceutical compounds.

Table 1.

Mobile phases reported in the literature for silica-based TLC separation of various radiopharmaceuticals.

Radiopharmaceutical	Aqueous Mobile Phase	Non-Aqueous Mobile Phase
[18F]Fallypride	(60:40 MeCN: 25 mM NH4HCO2, 1% TEA)10,11, (95:5 MeCN:H2O)17, (90:10 MeCN:H2O)18	(10:90 MeOH:DCM)19, (50:50 MeOH:EtOAc, 1% TEA)20
[18F]FET	(80:20 MeCN:H2O)10,21, (67:16.5:16.5 MeCN:MeOH:H2O)22	(90:10 MeOH:AcOH)23, (67:33 Hexanes:EtOAc)24
[18F]FBB	(90:10 MeCN:H2O)25	NR
[18F]FDOPA	(95:5 MeCN:H2O)26, (67:16.5:16.5 MeCN:MeOH:H2O)27	(90:10 DCM:EtOAc)28, (40:60 EtOAc:Et2O)29
[18F]FEPPA	NR	(8:10:82 MeOH:Hexanes:EtOAc)30
[18F]FPEB	NR	(95:5 EtOAc:EtOH)31
[18F]FLT	(95:5 MeCN:H2O)17,32	(90:10 DCM:MeOH)33
[18F]FMZ	(80:15:5 EtOAc:EtOH:H2O)34,35	(80:20 EtOAc:EtOH)36
[18F]FMISO	NR	(95:5 MeOH:NH4)32, (MeOH)37
[18F]FNB	(60:40 MeCN:H2O)38	NR
[18F]FBA	(95:5 MeCN:H2O)38	(67:33 Hexanes:EtOAc)39
[18F]DFA	(95:5 MeCN:H2O)38	NR
[18F]AlF-2-AMPDA-HB	(75:25 MeCN:H2O)40	NR
[18F]AlF-NOTA-HL	(50:50 MeCN:H2O)41	NR
[18F]FTP	(20:80 MeOH: 1M NH4OAc)42	NR
[18F]Altanserin	(80:20 MeCN:H2O)43	NR
[18F]MPPF	(90:10 MeCN:H2O)44	NR

NR = not reported.

However, a drawback of using water is that it can alter the stationary phase itself through direct interactions of the water with surface silanol groups of the TLC plate. These modifications can disrupt the possible intermolecular interactions of the surface silanol groups, impeding plate-analyte interactions and adversely affecting chromatographic behavior.⁷ In fact, under some conditions, the [¹⁸F]fluoride-silica interaction can be disrupted, leading to the movement of free radionuclide away from the origin, which could lead to confusion and ambiguities in the analysis of ¹⁸F-radiopharmaceuticals if the TLC method is not carefully validated. We use systematic studies with different aqueous compositions to illustrate the potential detrimental impacts of using mobile phases with significant aqueous component on the radio-TLC analysis of tracers labeled with [¹⁸F]fluoride and argue that the effects could also apply to tracers labelled with other radionuclides.

2. EXPERIMENTAL

2.1. Materials and Methods

All reagents and solvents were obtained from commercial suppliers. Acetonitrile (MeCN; anhydrous, 99.8%), methanol (MeOH; anhydrous, 99.8%), water (H₂O; suitable for ion chromatography), 2,3-dimethyl-2-butanol (thexyl alcohol; anhydrous, 98%), N-methyl-2-pyrrolidone (NMP; anhydrous, 99.5%), 4,7,13,16,21,24-hexaoxa-1,10-diazabicyclo[8.8.8]hexacosane (K₂₂₂; 98%), and potassium carbonate (K₂CO₃; 99.995%), were purchased from Sigma-Aldrich (St. Louis, MO, USA). Tetrabutylammonium bicarbonate (TBAHCO₃; 75mM in ethanol), (2S)-O-(2′-tosyloxyethyl)-N-trityl-tyrosine-tert-butyl ester (TET; precursor for [¹⁸F]FET, >95%), O-2-fluoroethyl-L-tyrosine (FET-HCl; reference standard, >95%), ethyl-5-methyl-8-nitro-6-oxo-5,6-dihydro-4H-benzo[f]imidazo[1,5-a][1,4]diazepine-3-carboxylate (nitromazenil; precursor for [¹⁸F]Flumazenil, >97%), Flumazenil (FMZ; reference standard, >99%), (S)-2,3-dimethoxy-5-[3-[[(4-methylphenyl)-sulfonyl]oxy]-propyl]-N-[[1-(2-propenyl)-2-pyrrolidinyl]methyl]-benzamide ([¹⁸F]Fallypride precursor, >90%), Fallypride (reference standard, >95%), were purchased from ABX Advanced Biochemical Compounds (Radeberg, Germany). Silica gel 60 F₂₅₄ sheets (aluminum backing, 5 cm × 20 cm) were purchased from Merck KGaA (Darmstadt, Germany). Glass microscope slides (76.2 mm × 50.8 mm, 1 mm thick) were obtained from C&A Scientific (Manassas, VA, USA).

No-carrier-added [¹⁸F]fluoride was produced by the (p, n) reaction of [¹⁸O]H₂O (98% isotopic purity, Huayi Isotopes Co., Changshu, Jiangsu, China) in an RDS-111 cyclotron (Siemens, Knoxville, TN, USA) at 11 MeV, using a 1.2-mL silver target with havar foil.

2.2. Preparation of samples of [¹⁸F]fluoride and complexes

To illustrate the impact of aqueous mobile phases on the migration of [¹⁸F]fluoride, several samples were prepared.

[¹⁸F]fluoride samples were prepared by diluting [¹⁸F]fluoride/[¹⁸O]H₂O with ion chromatography-grade water to a concentration of 0.75–1.1 MBq/μL.

[¹⁸F]KF/K₂₂₂ samples were prepared by adding K₂CO₃ (0.5 mg, 3.6 μmol) and K₂₂₂ (5 mg, 13.3 μmol) to a volume of 0.5 mL of ion chromatography grade water spiked with [¹⁸F]fluoride, yielding a 0.75–1.1 MBq/μL solution with 7.2 mM K₂CO₃ and 26.2 mM K₂₂₂.

[¹⁸F]TBAF samples were prepared by adding TBAHCO₃ (75 mM; 1.2 μL, 0.7 μmol) to 99 uL of ion chromatography grade water spiked with [¹⁸F]fluoride to yield a 0.75–1.1 MBq/μL solution with 7.2 mM TBAHCO₃.

2.3. Preparation of samples of ¹⁸F-radiopharmaceuticals

To prepare mixed samples of radiotracers and [¹⁸F]fluoride, several radiopharmaceuticals were prepared using droplet radiochemistry methods on Teflon-coated silicon surface-tension trap chips as previously described⁸; except that optimal reaction conditions were altered to increase the amount of [¹⁸F]fluoride in the crude reaction mixture.

Mixtures of [¹⁸F]FET-intermediate/[¹⁸F]TBAF were prepared by depositing an 8 μL droplet of [¹⁸F]fluoride/[¹⁸O]H₂O (70–90 MBq [1.9–2.5 mCi]; mixed with 240 nmol of TBAHCO₃) and drying at 105 °C for 1 min. Then, the fluorination step was performed by adding a 10 μL droplet containing 80 nmol of FET precursor dissolved in thexyl alcohol:MeCN (1:1; v/v) to the dried [¹⁸F]fluoride residue and reacting at 90 °C for 5 min. The crude product was collected by dispensing 10 μL of MeCN to the reaction site and aspirating the volume. This process was repeated 6x for 60 μL of collected crude product.

Mixtures of [¹⁸F]FMZ/[¹⁸F]TBAF were prepared similarly, except that the fluorination step was performed by adding an 8 μL droplet containing 280 nmol of FMZ precursor dissolved in NMP to the dried [¹⁸F]fluoride residue and reacting at 200 °C for 0.5 min.

Mixtures of [¹⁸F]Fallypride/[¹⁸F]TBAF were prepared similarly, except that (i) the initial [¹⁸F]fluoride droplet contained 480 nmol of TBAHCO₃, and the fluorination step was performed by adding a 6 μL droplet containing 234 nmol of Fallypride precursor dissolved in thexyl alcohol:MeCN (1:1; v/v) to the dried [¹⁸F]fluoride residue and reacting at 110 °C for 1 min.

Stock solutions of non-radioactive reference standards were prepared at 20 mM concentration. 5 mg of Fallypride standard was added to 685 μL of MeOH. 5 mg of FMZ standard was added to 825 μL of MeOH. 5 mg of FET standard was added to 1100 μL of MeOH. MeOH was chosen due to the high solubility of these reference standards (as suggested by the manufacturer), and due to its relatively low boiling point that allowed rapid drying after spotting onto the TLC plate.

2.5. TLC spotting, developing, and readout

TLC plates were cut (6 cm long × 3 cm wide), then marked with a pencil at 1 cm (origin line) and 5 cm (development line) from the bottom edge. 1 μL of the relevant sample was applied to the plate via a micro-pipette. In cases where radiopharmaceutical solutions were separated, an adjacent lane on the plate was spotted with the corresponding reference standard. The sample spots were then dried under a gentle stream of nitrogen for 1 min. Spotting was repeated on multiple plates to compare the effect of different mobile phases containing MeCN with different amounts of H₂O (all compositions expressed as v/v). After developing, plates were dried under a gentle stream of nitrogen for 3 min.

To better visualize the location of phase transfer catalysts (TBAHCO₃ and K₂CO₃/K₂₂₂), some plates were stained after developing by exposure for 1 min to a mixture of iodine crystals and silica gel in a sealed container.⁹

Plates were visualized via Cerenkov luminescence imaging (CLI) as previously described.^10,11 Briefly, the radio-TLC plate was positioned inside a light-tight chamber, then the plate was covered with a glass microscope slide. The Cerenkov light emission was detected by a cooled (−10 °C) scientific camera (QSI 540, Quantum Scientific Imaging, Poplarville, MS, USA) equipped with a 50 mm F/1.2 lens (Nikkor, Nikon, Tokyo, Japan) for a 60 s exposure. In addition to previously described corrections, background subtraction was further performed by selecting a small region of the image (approximate size 20 pixels) not containing any radioactive species, computing the average pixel intensity, and subtracting this average from the entire image.

After CLI imaging, the glass microscope slide was removed, and a UV lamp installed inside the light-tight chamber was illuminated while acquiring another image of the plate (7 ms exposure time). This enabled visualization of chemical species on the plate (which appear as darker bands due to indicator present on the TLC plate), as well as capturing an image of pencil markings and iodine-stained bands on the plate.

2.6. Analysis of TLC plates

A MATLAB program (MathWorks, Natick, MA, USA) was written to generate TLC chromatograms and display TLC plate images. The user is first prompted to select the CLI image to be analyzed, followed by a corresponding darkfield image, and corrections are applied as previously described.¹² The program then asks the user to select a corresponding UV image, followed by selecting a flat field correction UV image (taken in advance with a blank TLC plate installed). The UV image is corrected in a similar fashion (i.e., dividing by the flat field correction image and applying 3×3 median filtering), and the user can further adjust brightness and contrast for viewing. The CLI image is then redisplayed, and the user is asked to draw a line to define the width of the widest radioactivity band, which is used as the lane width. The UV image is then redisplayed, and the user is asked to draw a line from the bottom of the TLC plate to the solvent front. The program then generates an average line profile along the lane, taking the pixel intensities versus distance from the CLI image along the user-defined centerline and averaging with adjacent lines automatically generated at 1-pixel intervals along the entirety of the selected lane width. The program displays the final averaged line profile (chromatogram), the corrected CLI image, and the UV image.

The chromatogram was then exported and loaded in OriginPro (OrignLab, Northampton, MA, USA) to normalize the chromatograms to the highest intensity and plot groups of chromatograms. To compute the percentage of each species in a chromatogram, OriginPro was used to find the area under each band (peak) after fitting to a sum of Gaussian curves and then dividing the area corresponding to a particular band by the sum of areas for all bands.

3. RESULTS AND DISCUSSION

Samples of [¹⁸F]fluoride (with and without phase transfer catalysts) and various crude radiopharmaceuticals with different polarities were prepared and separated on silica TLC plates with different aqueous mobile phase compositions to illustrate the effect of water on the mobilization of different species and the potential pitfalls in analysis.

3.1. Effect of Aqueous Mobile Phases on Migration of [¹⁸F]fluoride

Samples of [¹⁸F]fluoride/[¹⁸O]H₂O were initially spotted on TLC plates and developed under mobile phases of increasing aqueous composition (Figure 1, Table 2). When the water content is low, [¹⁸F]fluoride remains at the origin as expected. For ≥40% water, free [¹⁸F]fluoride begins migrating away from the origin. Notably, and with great potential for ambiguity in the radio-TLC analysis of radiopharmaceuticals, two distinct bands of radioactivity are observed when using a mobile phase with ≥50% water. For ≥80% water composition, there is again only a single band observed, but it is located at the solvent front, near where the radiopharmaceutical species would be expected.

Figure 1.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the migration of [¹⁸F]fluoride. (A) CLI images of TLC plates. (B) TLC chromatograms generated from the CLI images. F denotes [¹⁸F]fluoride.

Table 2.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the retention factor (R_f) of [¹⁸F]fluoride.

MeCN (%)	H2O (%)	Rf
100	0	0.0
90	10	0.0
80	20	0.0
70	30	0.0
60	40	0.0, 0.13
50	50	0.0, 0.44
40	60	0.0, 0.76
30	70	0.0, 0.83
20	80	0.92
10	90	0.92
0	100	0.93

Recognizing that phase transfer catalysts are typically used in ¹⁸F-radiosyntheses, the behavior of complexed [¹⁸F]fluoride was also investigated. Figure 2 and Table 3 show the effects of water on the movement of [¹⁸F]TBAF (i.e., [¹⁸F]fluoride in the presence of TBHACO₃). The overall trend is similar to free [¹⁸F]fluoride, though there are increased signs of chromatographic fronting and band widening for some mobile phase compositions. This fronting may be explained because [¹⁸F]fluoride and the phase transfer catalyst interact with one another yet have different migration behavior (Figure 2A). The migration of [¹⁸F]KF/K₂₂₂ (i.e., [¹⁸F]fluoride in the presence of K₂CO₃ and K₂₂₂) was also explored (Supplementary Figure S1 and Table S1) and shown to behave similarly to [¹⁸F]TBAF.

Figure 2.

The effect of aqueous mobile phases (MeCN:H₂O) on the migration of [¹⁸F]TBAF. (A) For each mobile phase composition, two images are shown: a UV image of the TLC plate stained with I₂ to visualize TBAHCO₃ (top), and a CLI image of a TLC plate spotted with [¹⁸F]TBAF (bottom). (B) Normalized TLC chromatograms generated from the CLI images. F denotes [¹⁸F]TBAF. The Supplementary Information contains an additional figure showing the migration of samples of [¹⁸F]KF/K₂₂₂.

Table 3.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the Rf of [¹⁸F]TBAF.

MeCN (%)	H2O (%)	Rf
100	0	0.0
90	10	0.0
80	20	0.0
70	30	0.0
60	40	0.0, 0.16
50	50	0.0, 0.40
40	60	0.0, 0.53
30	70	0.0, 0.77
20	80	0.88
10	90	0.92
0	100	0.93

3.3. Effect of Aqueous Mobile Phases on the Analysis of [¹⁸F]Fallypride

Noting the effects that water composition had on the chromatographic behavior of [¹⁸F]fluoride and its complexes, the effect of aqueous mobile phases on the chromatographic behavior of crude [¹⁸F]Fallypride samples (containing unreacted [¹⁸F]TBAF) was investigated (Figure 3, Table 4). Interestingly, while [¹⁸F]Fallypride is non-polar (cLogP = 3.3), using a mobile phase of 100% MeCN did not lead to migration away from the origin, and the [¹⁸F]TBAF and [¹⁸F]Fallypride bands could not be resolved. Increasing the water content to ≥20% caused the migration of [¹⁸F]Fallypride toward the solvent front and allowed the bands to be resolved. However, for ≥40% water, there was decreased migration of [¹⁸F]Fallypride (and further increased mobilization of [¹⁸F]TBAF), and at 50% water, there was a significant overlap of the two species. Further addition of water (≥60%) led to a reversal in the order of the bands.

Figure 3.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the migration of [¹⁸F]Fallypride. (A) Images of TLC plates. For each mobile phase composition, a CLI image of the TLC plate spotted with crude [¹⁸F]Fallpyride is shown (top), along with a UV image of an adjacent lane spotted with Fallypride standard (bottom). The Fallypride standard band is enclosed with a yellow dashed line for clarity. (B) Normalized TLC chromatograms generated from the CLI images. F denotes [¹⁸F]TBAF, P denotes [¹⁸F]Fallypride, and I denotes impurity.

Table 4.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the R_f of different radiochemical analytes in the crude synthesis of [¹⁸F]Fallypride. The R_f of [¹⁸F]Fallypride was confirmed by parallel spotting of [¹⁹F]Fallypride.

		Rf values
MeCN (%)	H2O (%)	[18F]TBAF	[18F]Fallypride	Impurity
100	0	0.0	0.07	0.75
90	10	0.0	0.27	0.84
80	20	0.0	0.50	0.86
70	30	0.0	0.68	0.90
60	40	0.0, 0.19	0.57	0.79
50	50	0.0, 0.40	0.40	ND
40	60	0.0, 0.53	0.18	ND
30	70	0.0, 0.77	0.10	ND
20	80	0.88	0.05	ND
10	90	0.92	0.0	ND
0	100	0.93	0.0	ND

ND = Not discernable.

Under most TLC analyses performed, the radioactivity at the origin is generally assumed to be the free radionuclide, and the migrated species as the intermediate or product. Without careful assessment and validation of which bands correspond to which species, the overlapped or reversed results under the moderate or high proportion of water could introduce significant ambiguity and errors during analysis (Supplementary Figure S2).

3.4. Effect of Aqueous Mobile Phases on the Analysis of [¹⁸F]FMZ

The effects of water composition were further investigated using samples of a more polar compound, [¹⁸F]FMZ (cLogP = 1.0), containing [¹⁸F]TBAF (Figure 4, Table 5). Though [¹⁸F]FMZ is more polar than [¹⁸F]Fallypride, the use of 100% MeCN led to the complete separation of [¹⁸F]FMZ (near the solvent front) from [¹⁸F]TBAF (at the origin). For ≥40% water, the [¹⁸F]TBAF band begins to migrate away from the origin and split into two distinct radioactive bands, and for ≥60% water, mobilization of [¹⁸F]FMZ begins to be adversely impacted. Significant overlap of the bands was observed using a 70% water mobile phase. Mobile phases with higher water content resulted in the migration of [¹⁸F]TBAF to the solvent front, and [¹⁸F]FMZ remained close to the origin. As mentioned previously, improper assumptions about the band locations could lead to significant misinterpretation of results (Supplementary Figure S3).

Figure 4.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the migration of [¹⁸F]FMZ. (A) Images of TLC plates. For each mobile phase composition, a CLI image of the TLC plate spotted with crude [¹⁸F]FMZ is shown (top), along with a UV image of an adjacent lane spotted with FMZ reference standard (bottom). The FMZ standard band is enclosed with a yellow dashed line for clarity. (B) Normalized TLC chromatograms generated from the CLI images. F denotes [¹⁸F]TBAF, P denotes [¹⁸F]FMZ, and I denotes impurity.

Table 5.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the R_f of different radiochemical analytes in the crude synthesis of [¹⁸F]FMZ. The R_f of [¹⁸F]FMZ was confirmed by parallel spotting of [¹⁹F]FMZ.

		Rf values
MeCN (%)	H2O (%)	[18F]TBAF	[18F]FMZ	Impurity
100	0	0.0	0.84	0.89
90	10	0.0	0.89	0.79
80	20	0.0	0.78	ND
70	30	0.0	0.82	ND
60	40	0.0, 0.18	0.88	ND
50	50	0.0, 0.42	0.88	ND
40	60	0.0, 0.58	0.82	ND
30	70	0.0, 0.80	0.70	ND
20	80	0.84	0.60	ND
10	90	0.84	0.53	ND
0	100	0.93	0.18	ND

ND = Not discernable

3.5. Effect of Aqueous Mobile Phases on the Analysis of [¹⁸F]FET-intermediate

Next, the impact of aqueous mobile phases was assessed for crude mixtures of [¹⁸F]FET-intermediate (cLogP = 3.9) containing [¹⁸F]TBAF. Spotting of FET reference standard (cLogP = −0.9) in an adjacent lane allowed visualization of the impact of mobile phase composition on the analysis of 3 species (i.e., [¹⁸F]TBAF, [¹⁸F]FET-intermediate, and FET). With 100% MeCN, both [¹⁸F]TBAF and FET remain at the origin, while the non-polar [¹⁸F]FET-intermediate moves with the solvent front (Figure 5, Table 6). With ≥10% water, FET begins to move away from the origin, and with ≥40% water, [¹⁸F]TBAF moves away from the origin, travelling close to FET. Interestingly, using ≥60% water impairs the mobility of the [¹⁸F]FET-intermediate, which ends up closer to the origin with increasing water content, while [¹⁸F]TBAF and FET remain overlapped at the solvent front. Supplementary Figure S4 shows the potential for the erroneous determination of radiochemical yield if the bands were improperly identified.

Figure 5.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the migration of [¹⁸F]FET-intermediate and FET. (A) Images of TLC plates. For each mobile phase composition, a CLI image of the TLC plate spotted with crude [¹⁸F]FET-intermediate is shown (top), along with a UV image of an adjacent lane spotted with FET reference standard (bottom). The FET standard band is enclosed with a yellow dashed line for clarity. (B) Normalized TLC chromatograms generated from the CLI images. F denotes [¹⁸F]TBAF, Int denotes [¹⁸F]FET-intermediate, and I denotes impurity.

Table 6.

Impact of water composition in aqueous mobile phases (MeCN:H₂O) on the R_f of different radiochemical analytes in the crude synthesis of [¹⁸F]FET.

		Rf values
MeCN (%)	H2O (%)	[18F]TBAF	[18F]FET-intermediate	FET
100	0	0.0	0.97	0.0
90	10	0.0	0.95	0.11
80	20	0.0	0.97	0.48
70	30	0.0	0.95	0.69
60	40	0.0, 0.13	0.95	0.79
50	50	0.0, 0.43	0.91	0.80
40	60	0.0, 0.55	ND	0.82
30	70	0.0, 0.77	ND	0.82
20	80	0.79	ND	0.81
10	90	0.82	0.0	0.76
0	100	0.90	0.0	0.71

ND = Not discernable.

3.6. Discussion

These findings underscore water’s complex role in chromatographic selectivity and highlight two major concerns about using water as a mobile phase additive for radiopharmaceutical analysis on silica TLC plates. Firstly, the use of increasing amounts of water leads to a decreasing ability of the silica TLC plate to sequester [¹⁸F]fluoride (and [¹⁸F]fluoride complexes) at the origin, as well as the possibility that the [¹⁸F]fluoride can form multiple bands when using a certain range of mobile phase compositions. Secondly, the water content can greatly affect the migration of the radiopharmaceutical species. Evidently, water plays a more important role in TLC plate selectivity than as a purely polar additive to affect analyte retention. These effects can lead to the overlap of bands at moderate water content and reversal of expected band positions at high water content, potentially leading to ambiguous or inaccurate determination of radiochemical compositions if TLC methods are not carefully validated.

A possible explanation for the first observation could be that water can convert silanol groups (Si-OH; isoelectric point ~2–3) to silanolate (Si-O⁻) groups.¹³ While silanol groups can act as weak ion exchangers and bind anionic [¹⁸F]fluoride, silanolate groups do not possess this ion exchange capacity¹⁴, potentially explaining the inability of the silica plates to sequester [¹⁸F]fluoride at the origin when using mobile phases with moderate to high water percentages. This change in the TLC plate may also explain the reduced migration of polar radiopharmaceuticals for mobile phases with high water content, i.e., modification of the silanol groups could lead to greater analyte affinity.

Interestingly, the water content of the mobile phase also appears to strongly influence the migration of cationic radionuclides like [⁶⁸Ga]Ga³⁺ on silica TLC plates. In a recent publication, researchers studied the effects of various mobile phases in the analysis of a ⁶⁸Ga-labeled radiopharmaceutical¹⁵, finding that with a 50% aqueous mobile phase, [⁶⁸Ga]Ga³⁺ remains at the baseline, but for 100% aqueous mobile phases, [⁶⁸Ga]Ga³⁺ migrated with the solvent front. This trend of mobilization for cationic species also appears to be true for other radionuclides like [⁶⁴Cu]Cu²⁺, which have also been shown to move with the solvent front when using purely aqueous mobile phases¹⁶, and it is possible that similar effects could be possible for other charged radionuclides (e.g., Sc-47, Zr-89, I-124, Lu-177, Ac-225). Further study is needed to better understand the chromatographic behavior of radiometals on silica TLC plates due to the complexity of these systems (e.g., different charge states of metal ions, possible coordination of metal ions with anions or solvents, and possible coordination with the surface functional groups).

4. Conclusions

Via systematic studies of different TLC mobile phase compositions and different ¹⁸F-labelled radiopharmaceuticals, we investigated the potential pitfalls of using water-containing mobile phases in TLC analysis of radiopharmaceuticals on silica TLC plates. Aqueous mobile phases with ≥30% water composition led to the migration of [¹⁸F]fluoride (or complexes) away from the origin, with higher water content (~50%), leading to the splitting of the [¹⁸F]fluoride band, and further increase of water content pushing [¹⁸F]fluoride to the solvent front. Secondarily, it was found that moderate amounts of water could hinder the migration of the radiopharmaceutical and even cause overlap with the [¹⁸F]fluoride band. While water is often used as a polar mobile phase additive in radiochemical analysis, the observations in this work highlight that water has, in fact, a more complex role in chromatographic selectivity, and care is needed in radio-TLC interpretation when using mobile phases containing significant amounts of water. To avoid these complex effects, we are exploring facile methodologies for purely organic mobile phase optimization that can efficiently separate radiopharmaceuticals from radionuclides.