A Controlled System for Parahydrogen Hyperpolarization Experiments

1. Introduction

Nuclear magnetic resonance (NMR) and magnetic resonance imaging (MRI) are valuable techniques used throughout chemistry, physics, and medicine [1,2,3]. Yet, their impact is inherently limited by low sensitivity and poor signal-to-noise ratio (SNR). The nuclear spin polarization, representing the fractional imbalance between energy levels responsible for the NMR signal, follows a simple expression, proportional to ℏγB₀/2kT, where ℏ is the reduced Planck constant, γ the gyromagnetic ratio, B₀ the magnetic field, k the Boltzmann constant, and T the temperature [4]. Conventional strategies to overcome this limitation rely on higher fields and/or lower temperatures. Brute-force nuclear hyperpolarization follows this route [5,6,7,8]. Reaching very high fields, however, is technically demanding. Commercial NMR magnets based on superconducting alloys such as NbTi or Nb₃Sn require cooling with liquid helium. The largest commercially available systems operate at ~28 T, weigh several tons, and cost millions of euros, yet still cannot break the fundamental scaling granting only up to about 1 proton spin over 10⁴ to contribute to the NMR signal. On the other side, cryoprobes and microcoils reduce noise on the detection side [9,10]

Hyperpolarization provides an alternative [11,12,13]. Several distinct methods have been developed, depending on the source of polarization (i.e., the signal) to be transferred to the nuclear spins: from electron spins, as in Dynamic Nuclear Polarization (DNP) [12,14,15,16,17,18,19,20]; from parahydrogen, as in Parahydrogen-Induced Polarization (PHIP) and Signal Amplification by Reversible Exchange (SABRE) [21,22,23,24,25,26]; from noble gases, as in Xe- and He-polarization [27]; or from quantum tunneling effects, as in Quantum Rotor-Induced Polarization [28,29]. DNP remains the most widely adopted method, capable of yielding signal enhancements of several hundredfold—either in solids under magic-angle spinning at cryogenic temperatures or in liquids via stochastic electron–nuclear interactions (Overhauser effect) [17,30,31,32,33,34]. Dissolution DNP (dDNP) extends this signal amplification to several thousand fold to solution-state NMR by polarizing frozen samples at ~1.2 K and rapidly dissolving them, enabling enhanced detection of metabolic probes such as pyruvate [3,15,18,35,36]. Despite its impact, DNP remains complex, expensive, and difficult to maintain.

Parahydrogen-based methods offer a simpler path [21,22]. PHIP and its variant PHIP Side Arm Hydrogenation (PHIP-SAH) [37,38] can match the performance of dDNP for many targets, while SABRE [25,26] allows repeated experiments on the same sample at a fraction of the cost. These techniques can hyperpolarize nuclei in seconds to minutes, without cryogenics or high-power microwaves. Their efficiency, however, depends critically on controlling key variables: magnetic field, parahydrogen enrichment, bubbling time, pressure, and flow rate.

In this work, we present an easily implementable parahydrogen and bubbling apparatus that provides control over pressure, flow rate, and temperature during parahydrogen-based hyperpolarization experiments.

Several systems for parahydrogen at different enrichment levels have been reported in the literature. Some operate at pressures around 490 psi (≈33 bar) and reach enrichments above 48%, up to nearly 100%, with designs aimed at clinical applications [39,40,41,42,43,44,45]. Commercial solutions are also available from Bruker (e.g., the BPHG90, delivering ≈87% parahydrogen at a cost above €100 000), Hyperspin Scientific (http://www.hyperspin.biz/), and Xeus Technologies (https://www.xeus-technologies.com/), which additionally offers a complete parahydrogen distribution system.

Our goal here is not to introduce an alternative method, but rather to demonstrate that a wide range of experiments—with full control over key experimental parameters—can be performed in almost any research environment, including for educational purposes, using relatively modest resources.

In many laboratories, parahydrogen experiments are still performed by manually shaking a J-Young NMR tube in the fringe field of a magnet, followed by detection at high magnetic field. While this approach can be effective, it raises questions about how well-controlled such experiments can truly be. Conversely, the full automation of parahydrogen bubbling may appear as a technical barrier — one that this work aims to overcome. We describe in detail the construction of a custom switch box that can be operated manually for parahydrogen experiments, while also enabling fully automated high-field operation through the TTL trigger signals available from most NMR spectrometer consoles. Practical examples of implementation are also provided, including pulse sequences used on Bruker instruments.

The proposed system is then employed to evaluate the performance of a recently developed SABRE catalyst for sodium [1-¹³C]pyruvate experiments, Chloro(1,5-cyclooctadiene)[1,3-bis(2,6-diisopropylphenyl)imidazolidin-2-ylidene]iridium(I) (Ir-SIPr), using a 1.4 T benchtop NMR spectrometer (Oxford XPulse), demonstrating ¹³C hyperpolarization under controlled bubbling conditions, although the sample transfer to the detection field was still performed manually. The same setup is used to perform fully automated high-field experiments (9.4 T) with the well-known Ir-IMes, Chloro(1,5-cyclooctadiene)[1,3-bis(mesityl)imidazolin-2-ylidene]iridium(I), to accurately monitor the catalyst activation time by following the hydride ¹H signals, to conduct polarization transfer experiments from parahydrogen protons to the ¹³C nucleus of [1-¹³C]pyruvate, and to study the dependence of the ¹³C signal intensity on hydrogen flow at 6 bar pressure. Overall, the system presented here—operating even with a 50% parahydrogen fraction—enables a broad range of experiments both at low magnetic fields typical of benchtop systems and at high magnetic fields, thereby significantly widening the applicability of parahydrogen-based hyperpolarization methods.

2. Description of the System

2.1. Parahydrogen Generator

In many laboratories, the hydrogen source is typically represented by high-pressure gas cylinders (≈200 bar), which are commercially available at relatively low cost. In our case, however, considerations related to logistics, safety regulations, and the desire to design a compact and easily transportable setup—allowing parahydrogen experiments to be performed on different NMR instruments—led us to adopt a different solution. A detailed list of components and estimated costs (< 10 k € excluding the hydrogen generator) is provided in Appendix A.

Hydrogen gas is provided by a commercial hydrogen generator (AD-600 Cinel SRL- Italy) (Figure 1b(1)). This unit produces high-purity (> 99.99999%) H₂ gas via a PEM (Proton Exchange Membrane) electrolytic H₂ generator using deionized water (low conductivity (<1µS/cm). In our case the H₂ gas flow is 600 mL/min and it is produced at a user-controllable pressure up to 10 bars. In our experiments, the hydrogen pressure was set to 7.5 bar. The gas produced by the generator was delivered through a 1/8″ Teflon tube, which was subsequently coupled to a 6 mm o.d. copper line. A S-Lok manual ball valve (S-Lok SBV1-3B-S-6M) regulated the gas flow into the system (Figure 1d(2)), while a first pressure gauge (Figure 1d(4))monitored the internal pressure (Swagelok PGI-63B-BG16-CASX). In parallel with the gas inlet, a second 6 mm o.d. outlet copper line is connected via a second manual valve (S-Lok SBV1-S-6M) (Figure 1d(3)) to a diaphragm pump (Vacuubrand Model MZ 1C) (Figure 1b(12)), which is activated only at the end of the experimental session to purge residual H₂ gas from the circuit. To prevent system over pressurization when residual gas in the line warms from liquid-nitrogen to room temperature, a safety pressure relief valve (IMI Norgren 1002/BR008) (Figure 1d(9)) with an automatic release at 14 bar was installed.

Hydrogen gas enters a soft annealed copper serpentine (Figure 1b(10)) into which approximately 21 g of Iron(III) oxide powder (Merck, 371254-50G, 30–50 mesh, CAS 20344-49-4) had been manually introduced and packed prior to coiling the tube.

The serpentine consists of 10.5 turns with an outer diameter of 65 mm. On both ends of the copper coil, compressed cotton wool plugs (approximately 20 mm in length) were inserted to retain the catalyst within the tubing.

Before loading, the Iron(III) oxide powder was washed to remove fine particles that could otherwise be transported downstream and promote back-conversion of parahydrogen to orthohydrogen. Specifically, the powder was poured onto a glass funnel fitted with a Whatman No. 1 filter paper (retention ~11 µm) and washed three times with n-Hexane until the filtrate appeared clear. The powder was then dried overnight at 60 °C in an air oven.

The resulting copper serpentine filled with the cleaned Iron(III) oxide catalyst is immersed in a 2 L Dewar (Isotherm DSS 2000) (Figure 1b(11)) containing liquid nitrogen at 1 bar and 77 K, where it acts as the parahydrogen conversion unit. At both the inlet and outlet of the copper serpentine, two 50 µm in-line filters (Parker M6A-F4L-50-SS) (Figure 1d(5)) are installed to further minimize the likelihood of iron particles migrating into the warm horizontal section of the parahydrogen generator shown in Figure 1d. A pressure gauge monitors, after a S-Lok manual ball valve (Figure 1d(6)), the gas pressure downstream of the Iron(III) oxide-filled serpentine (Figure 1d(7)). A mass flow controller (Sierra Instruments C100, connected to a 24 VDC power supply) (Figure 1d(8)) is used to regulate the flow of gas entering the distribution system. During SABRE-SHEATH catalyst activation, a low flow rate of approximately 20–30 sccm, adjustable and displayed on the mass-flow controller, is maintained for 20 minutes to prevent excessive solvent evaporation. During hyperpolarization experiments, the flow is typically set to ~80 sccm and can be increased up to 150 sccm.

2.2. The pH₂ Distribution System

The distribution system is part of the apparatus responsible for delivering the pH₂ gas from the parahydrogen generator to the NMR tube (see Figure 1c). The 7.5 bar pH₂ pressure after the mass flow controller is reduced to 6 bars by pressure regulator valve (Festo MS2-LR-QS6-D6-AR-BAR-B) (Figure 1c(1)). However, we have also successfully tested a final pressure of 8 bar at the NMR tube during the test experiments. In order to deliver the pH₂ gas in a controlled way to the NMR tube three valves are necessary: an inlet, an exhaust and a bypass valve. Although we used manual valves during the initial testing session, in the actual setup they have been all replaced with solenoid valves (SMC SMVDW20HA) (Figure 1c(2, 6, 7)) that can be remotely controlled as described below. They require a 24 V DC power supply unit (in the expansion of Figure 1c(4)). The inlet and exhaust lines, schematically sketched in (Figure 1a), guide the pH₂ through the NMR tube creating the desired bubbling when the bypass valve is closed. When it is open, the bypass valve connects the inlet and outlet and the bubbling stops while keeping the system under pressure (see Figure 1a). The desired pressurization of the system is guaranteed by a backpressure regulator (Rometec SRL - Back Pressure Regulator, 1/4" NPTF, body 316 SS, seat retainer PEEK, diaphragm 316 SS, o-ring Viton, 0-17 bar) equipped with an in-line pressure gauge (Figure 1c(3)) that is manually adjusted in order to keep the pressure slightly below the value dictated by the pressure regulator valve (Figure 1c(1)).

The 4 mm o.d. flexible tubings used for the inlet and outlet gas lines were connected via a push-in Y-adapter (two 4 mm and one 6 mm ports in Figure 1a(1)). The 6 mm port was fitted with a Teflon tube terminated with a nickel-plated brass compression fitting, allowing easy disconnection when required (Figure 1a(2)). The nickel-plated brass compression fittings were used to ensure a secure and leak-tight connection between the 6 mm o.d. Teflon lines (with tubing in Figure 1c) and the 4 mm o.d. flexible tubing used to deliver the gas to the NMR tube. The nickel-plated brass connectors (Figure 1a(2)) were also utilized in the experiments reported below at magnetic fields of 9.4 T, displaying no observable magnetic behavior. We employed a portable H₂ gas detector to monitor possible leaks in the apparatus (see Appendix A RS COMPONENTS - RS GD38).

To minimize field inhomogeneities in the NMR detection region, a 250 µm i.d. quartz capillary (Molex, Part Number: 1068150026) was glued to a 1/16″o.d. flexible Teflon tube, which was then inserted and bonded to the inner wall of the 4 mm o.d. tube using a two-component epoxy adhesive.

The core of the distribution system is represented by the control unit in Figure 1c(4) (expanded in Figure 1e,f). It operates from a standard AC 110/220 V, 50 Hz mains input, converted to 24 V DC (250 W) through a regulated power supply as detailed in the schematic diagram in Figure 2. The DC output feeds a four-pole double-throw (4PDT) MAN/AUTO switch, allowing the operator to select between manual and automatic operating modes. In manual mode, four illuminated pushbuttons (red, green, blue, and white — see also Figure 1f) provide direct activation of each output channel. Each button lights up when pressed, indicating that the corresponding circuit is energized. In automatic mode, control is transferred to a four-channel relay board (24 V DC, opto-isolated), equipped with electromechanical relays rated for up to 10 A at 250 V AC or 30 V DC. Each relay drives a 24 V DC solenoid valve that regulates gas flow within the system. The optocoupler interface ensures electrical isolation between the control logic and the load lines. When the system operates automatically, as in the 9.4 T experiments reported below, a series of LED switch on to indicate which valves are active, thus providing consistent visual feedback across both operating modes.

All components are housed within an aluminum enclosure (approximately 200 × 150 × 80 mm) with front-panel indicators, MAN/AUTO switch and connectors for external wiring. The design enables straightforward transition between manual testing and automated operation, ensuring robust and reproducible control of the solenoid valves under low-voltage (24 V DC) conditions. Although the solenoid valves heat up if kept active for extended periods (e.g., bubbling times longer than 1 min), potentially leading to malfunctions in their opening and closing mechanisms, no such issues were observed so far during our experiments, even for bubbling durations up to 10 minutes and continuous operation over several hours. In automatic mode, the solenoid valves are automatically deactivated at the end of each bubbling cycle.

2.3. The Temperature Control unit and Mu-Metal for SABRE-SHEATH Experiments

The temperature control unit consists of a Wilmad suprasil VT dewar (see Figure 3a). This insert (outer diameter 10.5 mm; lower section length 242 mm; upper section length 72 mm) is constructed from high-purity Suprasil® fused silica (quartz glass). It is designed as a non-silvered, unslotted insert, offering excellent optical purity, structural stability, and very low thermal expansion over a broad temperature range (rating from ~120 K up to ~600 K in standard use). In our experimental setup (see Scheme in Figure 3b), this Suprasil VT Dewar is connected to a liquid-nitrogen reservoir via an evaporator. At the bottom of the suprasil dewar a thermocouple measures the temperature (T probe in Figure 3a). The temperature probe is mounted to be in contact with the outer wall of the NMR tube to monitor the temperature of the gas directly. That sensor is interfaced with an external control unit Bruker BVT2000 which dynamically adjusts the flow of N₂ gas from the evaporator to regulate the temperature. Through this feedback scheme, we have observed temperature stability better than ± 0.1 K over periods of several hours. The Wilmad dewar is placed at the centre of the µ-metal shield (Sas Ateliers Soudupin, Zero gauss chamber, model: ZG-100-300-3-0 with 3 layers μ-metal thickness 1mm internal diameter : 100 mm length : 300 mm) which contains a solenoidal coil (364 turns of 1 mm copper wire wound in two layers for a total length of 18.5 cm) coupled via a 10 kΩ resistance to DC power supply (model Keysight E36231A) (Figure 3a) to generate the requested magnetic field for hyperpolarization experiments. Wilmad suprasil VT dewar is connected to a liquid nitrogen tank via an evaporator line. The sub microtesla magnetic field necessary for SABRE-SHEATH experiments was measured via a flux gate magnetometer (Bartington – MAG-03MC 1000 and Bartington MAGMETER 2). The calibration curve in Figure 3c was performed returning a good linearity between the applied voltage and the measured magnetic field.

2.4. How to Cable the for Automatic Parahydrogen Experiments at High Magnetic Fields

To perform high-field NMR experiments (i.e. 9.4 T for us), it is advantageous to employ a controlled bubbling system that can be directly operated and regulated by the pulse program initiating the NMR experiment itself. Here, we describe how the TTL output lines of the Bruker NMR console were used to control the state of the solenoid valves via the custom-built switch box shown in Figure 1c(4). In our setup, the Bruker Ascend 400 MHz spectrometer equipped with an Avance III NanoBay console provides four TTL output channels that can serve as trigger signals to control the solenoid valves. Since three valves were required for our system, only three of these lines were utilized. According to the Bruker manual (IPSO AQS Unit Technical Manual), several TTL lines—depending on the specific console configuration—can be assigned to transmit or receive trigger signals.

The implementation simply requires connecting the corresponding console pins to the switch box via BNC cables and editing the Avance.incl file to define the logical state (HIGH or LOW) of each valve during the pulse sequence execution. Once this configuration is established, the valve state can be controlled directly within the TopSpin pulse program by inserting the instruction TTL1 HIGH or TTL1 LOW for the valve associated with trigger 1 (and similarly for the trigger 2 and 3). Example of a pulse sequence used for bubbling and detection is provided in the Appendix B.

For Bruker NMR users, the first step is to consult the table listing the pin assignments for the RCP and Control signals of the T-controller on the IPSO AQS (see Table 4 in the Avance III NanoBay IPSO AQS Unit Technical Manual). In our setup, the console allows sending an output signal through pin V6. Table 3 of the same manual identifies the V bundle and the corresponding V6 pin. The signal from this pin can be routed through a BNC connector cable and coupled to the output of the switch box illustrated in Figure 1c,e, and Figure 2. To correctly use the V6 pin as a trigger output, the Avance.incl file must be modified. First, locate the Avance.incl file; then, refer to the appropriate Table in the Avance III NanoBay IPSO AQS Unit Technical Manual to find the setnmr number corresponding to the desired pin, and define the TTL line accordingly, as described in the Bruker manual. For instance, in our case the Avance.incl is located at: C:\Bruker\TopSpin3.7.0\exp\stan\nmr\lists\pp, and the TTL line for pin V6 can be defined as follows:

/*trigger outputs 1*/

#define TTL1_LOW setnmr4|14

#define TTL1_HIGH setnmr4^14

A similar syntax has been used for the other TTL lines.

3. Results and Discussion

As a demonstration of the versatility of the setup described above, we present a series of characterization experiments on the SABRE catalyst Ir-SIPr and gold standard Ir-IMes, synthesized in our laboratory. The Ir-SIPr complex features an NHC ligand bearing 2,6-diisopropylphenyl substituents and lacks the backbone unsaturation typical of its analog IPr. Together, these catalysts proved to be alternatives to the benchmark Ir-IMes, which remains, however, the best performing NHC catalyst at the moment [46]. The sample was prepared as detailed in the Materials and Methods section.

Figure 4. Pyruvate hyperpolarization by SABRE. Free pyruvate (4) is obtained as a reversible process via [Ir(H)₂(κ²-pyruvate)(DMSO)(NHC)] (3), whereas hydrogen exchange is mediated by ([IrCl(H)₂(DMSO)₂(NHC)]) (2) as detailed by Duckett et al. in ref.^47,48.

Figure 4. Pyruvate hyperpolarization by SABRE. Free pyruvate (4) is obtained as a reversible process via [Ir(H)₂(κ²-pyruvate)(DMSO)(NHC)] (3), whereas hydrogen exchange is mediated by ([IrCl(H)₂(DMSO)₂(NHC)]) (2) as detailed by Duckett et al. in ref.^47,48.

The SABRE mechanism for pyruvate follows the description by Duckett and co-workers [47,48]. The NHC precatalyst (either Ir-SIPr or Ir-IMes) (1) is activated in methanol-d₄ under a continuous parahydrogen flow (≈30 sccm) at 6 bar, with the activation times estimated below for Ir-IMes at 265 K (see Figure 5c). The co-ligand DMSO, at an optimal 5 equivalent concentration with respect to the catalyst, is instrumental to the reversible pyruvate to catalyst binding. In absence of co-ligand no pyruvate hyperpolarization is observed. Upon binding, the resulting sulfoxide complexes [Ir(H)₂(κ²-pyruvate)(DMSO)(NHC)] (3) (see Figure 4) act as the active polarization-transfer species for ¹³C-pyruvate, while [IrCl(H)₂(DMSO)₂(NHC)] (2) (see Figure 4) mediates hydrogen exchange. Two coordination geometries can be identified for pyruvate binding: 3a (axial) and 3b (equatorial).

3.1. Experiments at 1.4 T with a Benchtop NMR and at 9.4 T

Here we report SABRE-SHEATH experiments conducted at 1.4 T and complementary ¹H and ¹³C measurements at 9.4 T using Ir–SIPr and Ir–IMes catalysts with the setup described above.

Differences between NHC–Ir catalysts can be rationalized in terms of their electronic and steric properties [49,50]. The Tolman Electronic Parameter (TEP) [51], derived from ν(CO) stretching frequency in [Ni(CO)₃L], and the Huynh Electronic Parameter (HEP) [52], obtained from the ¹³C chemical shift of the carbene carbon in [PdBr₂(NHC)(pyridine)], respectively report on the σ-donor/π-acceptor balance and σ-donor strength of a given NHC ligand. Additional probes such as the ⁷⁷Se NMR shift of NHC–Se adducts or the ¹³C shift of the free carbene carbon provide complementary information on the electronic density at the metal center.

Steric effects are usually quantified by the percent buried volume (%Vbur), which describes the fraction of the coordination sphere occupied by the ligand, or, less commonly, by the Tolman cone angle, which represents the solid angle subtended by the ligand at the metal center. Taken together, these descriptors explain the distinct behavior of Ir–IMes and Ir–SIPr complexes, the latter being both a stronger σ-donor and more sterically demanding.

In SABRE-SHEATH, when bubbling is performed at the level-anticrossing (LAC) magnetic field (~0.33 μT), polarization is spontaneously transferred from parahydrogen to the J-coupled ¹³C nuclei. As a result, two main classes of signals are observed: one corresponding to the free substrate (4 in Figure 4) and up to two associated with substrate molecules bound to the catalyst (3a and 3b in Figure 4).

This behavior reflects the sterically constrained dynamics imposed by the bulky diisopropylphenyl groups of the SIPr ligand, which hinder formation or stabilization of the 3a isomer. A similar trend was found for the unsaturated analogue Ir–IPr, consistent with earlier observations linking ligand rigidity and donor strength to SABRE efficiency [48].

Figure 5a shows the temperature dependence of the ¹³C hyperpolarization for Ir–SIPr between 240 K and 305 K. Each sample was bubbled with parahydrogen for 25 s at 6 bar and 80 sccm, then manually transferred in 3 s to the 60 MHz benchtop NMR system for a 90° ¹³C pulse and signal acquisition. A 10 min delay between experiments allowed full thermal equilibration. The temperature at which each species reaches its maximum signal intensity—260 K for the bound 3b and 272.5 K for free pyruvate—highlights the distinct exchange dynamics governing carbon hyperpolarization.

Precise timing of the bubbling process, enabled by the switch box, allows accurate estimation of the build-up time for ¹³C polarization. Figure 5b reports experiments at 280 K, 6 bar, and 80 sccm with bubbling times from 1 s to 180 s. Integration of the free pyruvate (left peak in sub panel 5b) and equatorially bound 3b (right peak in sub panel 5b) peaks yields build-up times of 33.8 ± 3.8 s and 16.8 ± 3.3 s, respectively.

In addition, the effect of parahydrogen pressure at different temperatures was investigated for Ir-IMes, as shown in Figure 5c,d. The concentration of H₂ in methanol-d₄ solutions is expected to increase linearly with pressure within the explored temperature range [53]. Consistently, the signals corresponding to both free pyruvate (4) and the equatorially bound species (3b) increase with pressure, although the overall enhancement in signal intensity from 2 bar to 6 bar is only about 20%. However, the pressure dependence of the hyperpolarized signal can be significantly steeper for ¹H species, underscoring the importance of accurately controlling this parameter.

The switch box (see Figure 1c,e,f and Figure 2) can be either operated manually as for the experiments at 1.4 T or automatically via Topspin pulse program as for the experiments at 9.4 T. It enables precisely controlled bubbling directly inside the NMR probe at high magnetic field. Again, we tested the system using Ir-SIPr and the Ir-IMes catalysts. We first monitored the ¹H hydride signal as a function of bubbling time to assess catalyst activation at 265 K, and once the catalyst activation was achieved, we investigated the polarization transfer from hydrides to carbon via the SEPP-INEPT pulse sequence [54,55,56].

When the spectra at 9.4 T are normalized to the most intense resonance in the 0–10 ppm range, two conclusions emerge. (i) Ir-SIPr exclusively exhibits 3b hydrides (hydride protons corresponding to 3b at around -27.2 ppm and -29.1 ppm) consistently with what we found in Figure 5a,b, while Ir-IMes shows both 3a (hydride protons corresponding to 3a at around -24.0 ppm and -14.9 ppm) and 3b species (hydride protons corresponding to 3b at around -27.2 ppm and -29.1 ppm) together with signals from 2 (hydride protons corresponding to 2 at around -15.5 ppm and -21.5 ppm), consistently with what already reported in literature [46]. (ii) The integrated intensities of the 3b hydrides indicate a ratio of approximately 6.7 between Ir-IMes and Ir-SIPr (Figure 6a,b), with an activation time for Ir-IMes of ~1 min at 265 K and 6 bar pressure (Figure 6c).

Overall, these results show that stronger σ-donation and steric bulk in Ir-SIPr stabilize hydride species but limit exchange dynamics, reducing pyruvate hyperpolarization efficiency. In contrast, the less hindered IMes ligand promotes faster exchange between 3a, 3b, and 2, leading to higher polarization transfer at comparable conditions.

This mechanistic understanding also underscores the value of precisely controlled and automated experimental conditions, as subtle differences in ligand dynamics can only be probed reliably when gas flow, timing, and field exposure are accurately synchronized.

Building on these findings, the experimental setup allows polarization-transfer experiments at high magnetic field (9.4 T for us) to be conducted in a fully automated mode, offering the possibility to explore pulse sequences designed to convert hydride-derived two-spin order into heteronuclear magnetization. In particular, polarization can be transferred between the longitudinal two-spin order I_1zI_2z (where I₁ and I₂ denote the nuclear hydrides’ spin) and the heteronuclear spin S_x, the transverse spin operator associated to ¹³C in [1-¹³C]pyruvate) (Figure 6e).

One such experiment is the SEPP-INEPT sequence (Figure 6e), recently analyzed by Assaf et al. [54], which enables controlled transfer of hydride polarization to ¹³C. Without delving into technical details, the sequence converts hydride magnetization into ¹³C coherence, and the intensity of the resulting bound-form ¹³C signal depends on the hydride–¹³C J-coupling and the evolution delay τ₂. Using the automated switch box, we implemented this sequence to acquire ¹³C spectra. For Ir-IMes, the polarization of the 3b species was sufficiently high to detect the bound ¹³C resonance at 168.5 ppm (τ₁ = 20 ms, τ₂ = 60 ms), which can be compared to the one-pulse-detection ¹³C spectrum showing the free pyruvate signal at ~170 ppm (Figure 6e). In contrast, for Ir-SIPr, no detectable ¹³C signal was observed, consistent with the lower polarization level of the 3b form evident in Figure 6a,b. The 2S_yF_z operator, antiphase magnetization on ¹³C spin, present before the 90° pulses on ¹H and ¹³C is modulated by the J coupling between the two hydrides during τ₁, and by the J coupling between one of the hydrides and ¹³C during τ₂. By fixing τ₁ at 20 ms and varying τ₂ between 10 ms and 600 ms, we can, as introduced by Assaf et al. [54], estimate the heteronuclear J coupling. The signal is overall modulated according to the function Sin[2π J_SF τ₂] × Exp[–2 τ₂ R], where R represents the relaxation rate. The blue fitting in Figure 6e employing the above function, yields a J coupling between the hydride at –29.1 ppm and the ¹³C nucleus of bound (3b) pyruvate of approximately ~1.3 Hz. This value should be regarded as an estimate, whose accuracy can be improved by increasing the number of τ₂ points used in the experiment.

These results confirm that efficient hydride–substrate exchange dynamics are a prerequisite for detectable heteronuclear polarization transfer, highlighting how the interplay between electronic donation and steric hindrance dictates both catalytic activity and polarization efficiency.

4. Materials and Methods

A solution containing 6 mM Ir-NHC catalyst, 20 mM sodium [1-¹³C]pyruvate (Merck SRL, CAS: 87976-71-4) and 30 mM DMSO is prepared in methanol-d₄. In the case of 1.4 T experiments the sample was bubbled with 50% parahydrogen at 6 bars for 20 minutes at ambient temperature for catalyst activation. In the case of 9.4 T experiments the sample was prepared in an identical way but the activation was followed by repeated bubbling for 20 seconds for Ir-IMes. The Ir-SIPr and Ir-IMes were synthesized as described in the SI of reference [46]

5. Conclusions

We have presented a cost-effective and versatile setup for performing parahydrogen-based hyperpolarization experiments at 50% parahydrogen enrichment under precisely controlled pressure, temperature, and flow conditions. The system integrates a custom-built switch box that can be operated either manually or automatically through TTL triggers directly synchronized with the NMR console, thus enabling fully automated experiments at high magnetic field. The design is compatible with both benchtop and superconducting NMR instruments, extending the applicability of PHIP and SABRE techniques across a broad range of operating conditions.

The functionality of the setup has been demonstrated by investigating the hyperpolarization of sodium [1-¹³C]pyruvate in methanol-d₄ using Ir-SIPr and Ir-IMes as catalysts. At 1.4 T, we performed variable-temperature experiments in manual mode using both Ir-SIPr and Ir-IMes, while precisely controlling the bubbling time and maintaining temperature stability within ±0.1 K. For Ir-IMes, we further explored the dependence of the hyperpolarized signal not only on temperature but also on bubbling pressure, revealing the expected direct correlation between pressure and signal intensity.

At 9.4 T the activation of Ir-IMes catalyst was investigated at 265 K where we estimated to about 1 min the build-up time of the polarization via the hydrides signals. In addition, polarization transfer experiments from hydrides to ¹³C via the SEPP-INEPT sequence were demonstrated. Through the SEPP-INEPT sequence we have estimated the heteronuclear J interaction between one of the hydride and carbon 13 in 3b form to be in the 1 Hz range. Beyond demonstrating the feasibility of heteronuclear polarization transfer under SABRE conditions, the ability to run automated SEPP-INEPT experiments enables systematic optimization and quantitative analysis of spin transfer pathways, offering direct insight into the nature and dynamics of the active catalytic species.

Overall, the proposed system provides an accessible, reproducible, and scalable platform for PHIP and SABRE studies. By combining mechanical simplicity with electronic automation, it facilitates systematic investigations of catalyst performance and polarization dynamics—thereby contributing to the broader adoption and standardization of parahydrogen hyperpolarization methods in both research and teaching laboratories.