Bidistilled water was obtained with an Aquatron water still purification system model A4000D (Barloworld Scientific, Nemours, France). Elemental standards of Y, Zr, Nb, Mo, Ru, Rh, Pd, Cd, and As (all 1000 mg/L) were purchased from Merck KGaA (Darmstadt, Germany). Nitric acid (65%, AnalaR NORMAPUR®, (w/v)) and sulfuric acid (95%, AnalaR NORMAPUR, (w/v)) were purchased from VWR International (Radnor, PA, USA). Ammonium hydroxide solution (25%, Analytical Reagent Grade, (w/v)) was purchased from Fluka Chemie GmbH (Buchs, Switzerland). Ammonium nitrate (p.A.) was acquired from AppliChem GmbH (Darmstadt, Germany). Polypropylene sample tubes (15 and 50 mL) were obtained from Th. Geyer (Renningen, Germany) and syringe filters (0.45 µm, 25 mm, hydrophilic PTFE) were purchased from BGB Analytik AG (Boeckten, Switzerland). Empty cartridges for extraction chromatography were purchased from Merck KGaA (Darmstadt, Germany). TEVA resin used for the extraction chromatographic purification of [99Tc]Tc\(}_^\) was provided by Triskem International SAS (Bruz, France).

All ICP-MS experiments were conducted on an Agilent 7700 ICP-MS (Agilent Technologies, Santa Clara, CA, USA) with x-lens configuration and platinum sampler and skimmer cones using MassHunter software (MassHunter 4.6, Version C.01.06). The ICP-MS was tuned daily to obtain an optimal signal-to-noise-ratio (S/N) for the m/z of 99 and 101. Typical instrument parameters were as follows: RF power, 1600 W; RF matching, 1.52 V; sample depth, 5.5 mm; first extraction lens, − 10 V; second extraction lens, − 200 V; omega bias, − 105 V; omega lens, 5.4 V; cell entrance, − 40 V; cell exit, − 56 V; deflect, 12.6 V; plate bias, − 40 V; octopole bias, − 8.5 V; octopole RF, 180 V; energy discrimination, 7.1 V. The measurements were conducted without the need for a collision gas flow as backgrounds were sufficiently low and polyatomic interferences were not expected. Detector dead time was calibrated using an elemental In standard (10 mg/L and 50 µg/L). Dwell times for the m/z ratios 99 and 101 were set at 0.4 s to decrease the relative standard deviation of the recorded signals and to improve precision for the recorded isotope ratio. To additionally detect potential interferences of Mo, the m/z of 97 was monitored with a dwell time of 0.1 s. Mass bias correction was performed using the four most abundant isotopes of Ru with the exponential approach suggested by Rodriguez Gonzalez et al. [24].

To increase sensitivity, aerosol desolvation nebulization was performed with an Apex 2 High Sensitivity Desolvating System equipped with a MicroFlow PFA ST nebulizer operated with Apex software (Version 1.0.1.2, all from Elemental Scientific, Omaha, NE, USA). The operating parameters were tuned concurrently with the instrumental parameters of the ICP-MS: nebulizer gas flow, 0.85 L/min; argon makeup gas flow, 0.311 L/min; N2 add-gas flow, 3.9 mL/min; spray chamber temperature, 140 °C; condenser temperature, 3 °C.

The IC separation was conducted using a fully automated single platform system for total metal analysis and syringe-driven chromatography (prepFAST IC, Elemental Scientific, Omaha, NE, USA) that was configured to fulfill the specific needs of the described analysis (Fig. 2). Separation was achieved with an IonPac AG9-SC column (4 × 50 mm, Thermo Fisher Scientific, Waltham, MA, USA). Method development and system operation were performed using ESI SC (Version 2.9.0.496, Elemental Scientific, Omaha, NE, USA). Compared to the standard setup that allowed auto dilution of standards and samples within two separate loops, only one sample loop of 50 µL was installed as no dilution of the sample was necessary [25,26,27]. Instead, four chromatographic syringes were used to hold different eluents or rinsing solutions to set up a step gradient consisting of 150 mM NH4NO3 solution (eluent A, set to pH 9.2 with NH4OH solution) and bidistilled water (eluent B). 500 mM NH4OH solution (eluent C, rinsing) and 20 mM nitric acid (eluent D, rinsing) were prepared for rinsing the column as well as the chromatographic system before and after each injection. The developed three-step gradient was run at a flow rate of 650 µL/min and initially consisted of eluents A and B in a composition of 10% A and 90% B. After 40 s, the concentration was increased to 50% A until it was changed to 100% A after another 90 s. For quantification, a Ru post-column internal standard (PCIS; concentration 1 µg/kg, in 10 mM HNO3) was added at a flow rate of 50 µL/min through another syringe, so that the overall flow rate was set at 700 µL/min. After the chromatographic separation, the column was first rinsed with eluent C (60 s, 700 µL/min) and the chromatographic system and the loop were finally rinsed with eluent D (60 s, 700 µL/min). Column recoveries were determined by comparing flow injections of different diluted solutions of the quantified 99Tc standard with respective column injections by changing the position of the column valve accordingly. The examined spiked sample of 99mTc-generator eluate was previously diluted by a factor of 10,000, to reduce the effects of its matrix containing 0.9% w/w NaCl and to test the method at a concentration closer to its lower detection capabilities.

Schematic depiction of the developed setup for on-line ion chromatographic quantification of [99Tc]Tc\(}_^\) with IBDA using ion suppression and aerosol desolvation nebulization coupled to ICP-MS

The application of the aerosol desolvation system requires considerations for used buffers and especially in cases where non-volatile salts are used, crystal build-up can be a limiting factor [23]. To avoid formation of crystalline NH4NO3 within the aerosol desolvation nebulization system, a Dionex ACRS 500 chemically regenerated suppressor (4 mm ID, Thermo Fisher Scientific, Waltham, MA, USA) was installed between the chromatographic system and the sample introduction system. H2SO4 (150 mM) was chosen as a regenerant using an external peristaltic pump at a flow rate of 2.3 mL/min to suppress the ammonium cations originating from the eluent. Individual recovery of the suppressor was determined by flow injection of a diluted solution of the quantified 99Tc standard either flowing through or bypassing the suppressor.

The urine sample was collected from an anonymous patient, who underwent a bone scintigraphy using a [99mTc]Tc-MDP tracer. The study was approved by the local ethics committee (No. 2007-467-f-S and No. 2016–585-f-S, Ethikkommission der Ärztekammer Westfalen-Lippe und der Universität Münster). The sample was collected, immediately frozen at − 80 °C and left in the dedicated environment at the university hospital for a period spanning several half-lives of 99mTc until no increased activity was detectable. Prior to analysis, the sample was thawed, filtered with a syringe filter (0.45 µm, 25 mm, hydrophilic PTFE) and directly analyzed with the same setup presented in the experimental section above, which was used for all presented chromatography data. This involved the injection of 50 µL raw and undiluted urine onto the IC column immediately after the sample was filtered, which represented a very low preparation and analysis effort even for samples containing a strong matrix composition.

For on-line quantification of 99Tc using IBDA, sensitivity correction between 99Tc and 99Ru was necessary. To increase the accuracy of the sensitivity correction, an in-house elemental 99Tc standard was prepared from 99mTc-generator eluates. The raw eluate was stored in glass bottles and left in the dedicated environment at the university hospital until no increased activity through γ-radiation was detectable. To decrease potential exposure and limit the risks, associated especially with the use of higher concentrations of the newly generated standard, all vessels as well as all unprotected instrumental setups containing higher concentrations of 99Tc were covered in aluminum foil, to prevent any β-radiation from penetrating. As the largest hazard of 99Tc originates from contamination through small particles or aerosols which can show persistent behavior in the lungs, the formation of aerosols as well as dried residues of 99Tc-containing solutions, which could potentially form particles, was avoided or performed under controlled conditions in closed containments with appropriate air exchange.

To obtain the standard, multiple decayed generator eluates were unified to a total amount of ~ 35 mL. 99Tc from the combined eluates was cleaned and preconcentrated using TEVA resin. The resin was densely packed into a 500-µL cartridge, preconditioned with HNO3 (10 mL, 0.1 M) and loaded with the unified eluates that were spiked with HNO3 to a final acid concentration of 0.1 M. After washing with HNO3 (10 mL, 1 M), 99Tc was eluted in five steps of 100 µL (HNO3, 8 M). The first two fractions contained the entire detectable amount of 99Tc and were consequently unified and quantified using TXRF analysis.

The quantitative characterization of the cleaned and preconcentrated 99Tc standard was carried out with TRXF analysis (S2 PICOFOX spectrometer operated with Bruker Spectra software, version 6.1.5.0; all from Bruker Nano GmbH, Berlin, Germany). The excitation settings for the analysis were set at 50 kV and 750 mA. For quantification, two aliquots of 22.8 µL from the combined TEVA eluates were each mixed with 1.2 µL of a diluted solution of an elemental As standard, so that the final concentration of As in the mixture was 500 µg/L. A volume of 5 µL was pipetted onto two separate quartz disks, dried, and analyzed in triplicate with a recording time of 30 min. Data were smoothened with Savitzky-Golay filtering at a second-order polynomial regression over 15 data points. The emitted signal of Tc was found to be a stack of multiple emission lines of L-alpha and L-beta transitions, which required subsequent deconvolution. Therefore, the most intense signal, originating from the L-alpha 1 emission was isolated by fitting its left flank, unaffected by any overlap, with a gaussian function. The resulting gaussian-shaped curves obtained from the X-ray fluorescence emission of Tc and As were integrated and enabled an accurate calculation of the Tc content in the combined eluates, without the need for a certified Tc reference material. The specific relative sensitivities for Tc and As (\(_}}\) and \(_}}\)), necessary for calculating the concentration of Tc (\(_}}\)), were taken from the instrument software. The concentration of Tc was calculated using Eq. 1, which additionally required the recorded net intensities of Tc and As (\(_}}\) and \(_}}\)) and the concentration of the internal As standard (\(_}}\)).

$$_}}=_}}\cdot \frac_}}/_}}}_}}/_}}}$$

(1)