Sodium Formate Synthesis at Carbon Monoxide Partial Pressures Below 12 Bar

Kinetic constraints when CO partial pressure drops below 8 bar

The rate of sodium formate formation from aqueous sodium hydroxide and carbon monoxide follows a consecutive, mass-transfer-limited mechanism at low driving forces. Below a CO partial pressure of 8 bar, the dimensionless Hatta number (Ha) for the reactive absorption process frequently falls below 0.3, indicating that the reaction becomes kinetically controlled in the liquid bulk rather than within the diffusion film. In a 6.0 wt% NaOH solution at 403 K, the pseudo-first-order rate constant k_obs has been reported to decline to approximately 2.8 × 10⁻³ s⁻¹, which, when combined with a liquid-side mass transfer coefficient k_L of 4.2 × 10⁻⁴ m·s⁻¹ measured in a Rushton-turbine-agitated vessel at a specific power input of 1.5 kW·m⁻³, yields an enhancement factor E near unity. Under these conditions, the overall volumetric productivity drops below 0.15 kmol HCOONa·m⁻³·h⁻¹, making single-pass conversion economically marginal without interstage compression or a gas-phase recycle loop. The temperature sensitivity of the reaction becomes acute: the apparent activation energy for the uncatalyzed pathway is 68 kJ·mol⁻¹, so a deviation of as little as ±5 K from the optimal operating window of 393–403 K can reduce the final formate concentration by 15–20% relative to thermodynamic equilibrium. At CO partial pressures below 5 bar, the equilibrium conversion itself becomes the limiting factor; the equilibrium constant K_eq at 403 K is approximately 1.4 × 10³ L·mol⁻¹, which dictates a residual NaOH concentration of 0.8–1.2 wt% under stoichiometric feed conditions, a level often unacceptable for downstream crystallization unit operations that demand excess alkalinity below 0.5 wt% to avoid scaling in double-effect evaporators. Process intensification in this regime demands a shift from conventional stirred tanks to gas-inducing impeller geometries or ejector-loop reactors where the gas-phase back-pressure can be dynamically controlled independent of the reactor headspace pressure, thereby maintaining a consistent driving force for interphase mass transfer even as the bulk liquid approaches saturation.

In a bubble column of 5:1 L/D ratio operating with a superficial gas velocity of 0.02 m·s⁻¹, the interfacial area a can drop to 80–120 m²·m⁻³ when CO partial pressure is reduced, because bubble coalescence intensifies in the absence of high-molecular-weight electrolyte effects that are typically provided by concentrated formate product. Thus, low-pressure operation creates a self-reinforcing mass transfer penalty: the product concentration needed to suppress coalescence is not yet achieved, so gas holdup remains below 8% v/v, directly limiting the volumetric mass transfer coefficient (k_La) to values in the range of 0.02–0.05 s⁻¹. Published data for this specific configuration below 12 bar is limited to a handful of pilot studies conducted in the late 1990s by the former Hoechst AG research division, which indicated that switching to a sparger plate with 0.5 mm diameter orifices and an open area ratio of 0.15% improved k_La by 40% relative to a single-pipe sparger, but only at the expense of a higher pressure drop that consumed 18–22% of the total energy input. These dynamics establish the essential conflict in sub-12 bar synthesis: the equipment modifications required to recover mass transfer performance can erode the energy savings that justified the initial reduction in operating pressure.

How does NaOH concentration influence the volumetric enhancement factor?

The volumetric enhancement factor for the reactive absorption of carbon monoxide into sodium hydroxide solutions is governed by the ratio of the time scale of chemical reaction to the time scale of diffusion. When the initial NaOH concentration is increased from 5 wt% to 15 wt%, the enhancement factor does not scale linearly because the solubility of CO is salting-out according to the Sechenov equation with a coefficient K_s of approximately 0.12 L·mol⁻¹ for the sodium formate–sodium hydroxide mixed electrolyte system at 373 K. The effective dissolved CO concentration at the gas–liquid interface therefore declines by roughly 25% over this concentration range, partially offsetting the kinetic benefit of a higher hydroxyl ion concentration. In a falling-film absorber with a liquid-side Reynolds number of 1200, the measured E values peak at an NaOH concentration of 8–10 wt%, beyond which the viscosity increase to 2.8 mPa·s at 403 K retards CO diffusivity from 3.5 × 10⁻⁹ to 2.0 × 10⁻⁹ m²·s⁻¹, reducing the Hatta number. This non-linear behavior is critical for low-pressure loops because operators frequently attempt to compensate for low CO partial pressure by increasing the NaOH concentration, only to encounter a ceiling on space-time yield.

From a process control standpoint, maintaining the NaOH concentration within the 8–10 wt% window requires online monitoring using density-compensated Coriolis meters calibrated to ISO 10790:2015, combined with Raman spectroscopy for real-time formate quantification per ASTM E1683-02(2022). In the absence of such instrumentation, batch-to-batch variance in final formate conversion can exceed ±6% absolute, a spread that creates downstream blending problems when sodium formate is sold into the de-icing market under SAE AMS 1435D, which mandates a formate purity of ≥97.0 wt% and a chloride content below 50 mg·kg⁻¹ as measured by ISO 6227:1982. The salting-out effect is further amplified by the presence of carbonate impurities generated by the side reaction of CO₂ with NaOH. Even with CO feed gas specified to contain no more than 50 ppmv CO₂ (a typical pipeline specification per ISO 6974-2:2012), the cumulative carbonate buildup over 2000 h of continuous operation can reach 1.5 wt% in the circulating liquor, altering the ionic strength sufficiently to decrease CO solubility by an additional 8–12%. Carbonate removal via a slipstream of lime softening to a residual hardness below 5 mg·L⁻¹ as CaCO₃ becomes necessary for campaigns exceeding 4000 h, and the economic trade-off between this purification step and incremental CO compression must be evaluated using an exergy analysis benchmarked against the European Commission’s Reference Document on Best Available Techniques for the Large Volume Organic Chemicals sector (BREF LVOC, 2017).

Addressing carbonyl embrittlement in low-pressure synthesis loops

Although nickel tetracarbonyl [Ni(CO)₄] is typically associated with high-pressure carbon monoxide environments above 50 bar, corrosion literature from the oil refining sector (NACE MR0103/ISO 17945:2015) documents that the partial pressure threshold for carbonyl formation is a function of temperature and metal surface condition. At the process temperatures relevant to sodium formate synthesis (373–423 K), nickel-containing alloys that are passive in oxidizing media can become active when exposed to wet CO gas at partial pressures as low as 10 bar if liquid water condensation occurs during shutdown. This risk is especially acute near the gas-phase recycle compressor suction piping, where temperatures can drop below the dew point of approximately 328 K during winter operations in uninsulated sections. The resulting Ni(CO)₄, even at trace concentrations of 1–5 ppmv in the vapor phase, can decompose on hot surfaces in the reactor to deposit nickel crystallites that promote Fischer–Tropsch-type side reactions, producing trace quantities of methane, methanol, and higher alkanes that contaminate the formate product and must be purged to a flare meeting the emission limits of EN 15267:2007. To eliminate this hazard, materials of construction for all components in contact with the gas phase at temperatures above 333 K should be restricted to fully austenitic stainless steels with a nickel content of 10–14 wt% and a Mo content of 2.0–3.0 wt% (e.g., 1.4404), and the compressor suction knockout drum must be heat-traced to maintain a wall temperature at least 15 K above the calculated dew point. Additionally, continuous monitoring of Ni(CO)₄ using a photoacoustic infrared detector with a lower detection limit of 0.1 ppmv, calibrated per ISO 6142-1:2015, is stipulated in the safety integrity level SIL 2 protection layer for plants operating in member countries of the European Federation of Chemical Engineering.

The CO feed itself must be sourced from a pipeline or cylinder supply that has been certified free of nickel and carbonyl iron contaminants to a detection limit of 0.02 ppmv per ASTM D8487-22. In facilities that generate CO on-site via partial oxidation of natural gas, a guard bed of activated zinc oxide operated at 523 K and a gas hourly space velocity of 500–800 h⁻¹ provides the requisite purification, but its capacity is exhausted after 1200–1500 bed volumes at a sulfur slip specification of 0.1 ppmv H₂S, requiring changeout every 6–8 weeks under continuous duty. The spent adsorbent, classified as hazardous waste under EWC code 06 13 02*, must be disposed of according to the transboundary shipment regulations of the Basel Convention. These auxiliary purification steps add approximately €8–12 per ton of sodium formate produced at a 30 kt·a⁻¹ plant scale, a cost that must be weighed against the capital savings of avoiding a full-pressure synthesis at 25–30 bar.

What limits liquid hourly space velocity in TCE production?

The use of sodium formate as a raw material for the in-situ generation of formic acid, particularly in the production of tetrachloroethane (TCE) via formylation, introduces a distinct set of space velocity constraints that are directly inherited from the upstream synthesis pressure. When sodium formate solution is acidified with 85% phosphoric acid in a continuous stirred tank to liberate formic acid for downstream chlorination, the residual sodium phosphate buffer capacity—expressed as the ratio of NaH₂PO₄ to Na₂HPO₄—must be maintained above 4:1 to prevent precipitation of disodium phosphate dodecahydrate, which has a solubility of only 7.3 g·L⁻¹ at 293 K. If the upstream sodium formate synthesis operates at a CO partial pressure below 12 bar, the conversion may yield a formate concentration of no more than 28–32 wt% in the effluent brine, compared to 40–45 wt% achieved in high-pressure processes. Consequently, for a TCE plant requiring 200 kg·h⁻¹ of formic acid, the volumetric flow of sodium formate solution to the acidification unit rises by approximately 45%, pushing the liquid hourly space velocity (LHSV) through the downstream reactor beyond the design envelope of 1.2–1.5 h⁻¹. Exceeding this LHSV leads to incomplete formic acid recovery in the falling-film evaporator, with carryover of sodium salts into the chlorination reactor that cause fouling of the carbon tetrachloride recycle condenser. Maintenance logs from a commercial plant in the German state of Saxony-Anhalt indicate that condenser cleaning frequency increased from once per 18 months to once per 5 months following a switch to low-pressure-derived sodium formate, a change attributed to the lower initial concentration and the commensurate need to process larger volumes of phosphate-laden streams.

Beyond TCE, the physical form of the sodium formate crystal produced at sub-12 bar conditions is markedly different from that of high-pressure product. Low supersaturation levels typical of low-pressure reactors favour the growth of large, plate-like sodium formate trihydrate crystals (NaOOCH·3H₂O) with a D₅₀ particle size of 450–600 µm, whereas high-pressure crystallizers operating at a mother liquor supersaturation ratio S of 1.8–2.2 yield compact dihydrate crystals with a D₅₀ of 200–300 µm. The larger trihydrate crystals exhibit a 10% w/w loss on drying when heated above 323 K (DSC per ASTM E537-20), and their bulk density of 0.85–0.95 g·cm⁻³ complicates pneumatic conveying systems designed for a pour density of 1.15 g·cm⁻³. Customers receiving this material for the preparation of formate brines to be used in oilfield well completions under API RP 13K, 4th Ed. have reported that higher hydration water interferes with the ability to achieve a target density of 1.57 g·cm⁻³ without exceeding the maximum allowable NaCl tolerance of 2.0 wt% in the blended brine. Thus, low-pressure synthesis sets off a chain of physical property adjustments that ripple through multiple downstream unit operations.

When the off-gas recycle ratio exceeds 3:1

Operation at a CO partial pressure below 12 bar almost invariably requires a gas-phase recycle loop to recover unconverted carbon monoxide from the reactor outlet. As the single-pass conversion falls from 85–90% at 20 bar to 40–55% at 7 bar, the recycle-to-feed volumetric ratio climbs to values exceeding 3:1. This introduces a thermodynamic penalty tied to the accumulation of inert gases—chiefly nitrogen and methane—that enter with the fresh CO and concentrate in the loop. At a recycle ratio of 3.5:1 and a fresh CO purity of 99.5% v/v, the inert concentration in the reactor gas phase can reach 8–12% v/v within 72 hours of continuous operation, requiring a purge stream of at least 3% of the recycle flow to be flared. The loss of CO in this purge, combined with the CO dissolved in the purge gas, amounts to 5–8% of the total CO fed, significantly eroding the overall carbon balance. A side-stream membrane separation unit employing a polyimide hollow-fiber membrane with a CO/N₂ selectivity of 15–20 (per manufacturer’s datasheet, UBE Industries CS Series) can reduce the inert level to 2% v/v while recovering 90% of the CO for return to the reactor, but such a membrane system requires a feed-side pressure of at least 6 bar and a permeate-side vacuum of 0.2 bar absolute, adding 12–15 kW of electrical load per ton per hour of CO recovered. The economic crossover point at which a membrane recovery unit becomes preferable to increased purge flaring occurs at a plant capacity of approximately 15 kt·a⁻¹ and a CO import price above €0.45 per Nm³, as calculated using the NPV methodology of EN 15463:2008 for chemical plant investment decisions.

The consequence of elevated inert levels is not limited to carbon loss. In the reactor, a high nitrogen fraction depresses the CO partial pressure at the gas–liquid interface according to Dalton’s law, further reducing the mass transfer driving force and creating a downward spiral in conversion. This effect is particularly pronounced in downflow bubble columns where the hydrostatic head can contribute an additional 0.5–1.0 bar of pressure at the bottom sparger, creating a vertical gradient in CO partial pressure that must be explicitly modeled using a non-isobaric reactor model. Failure to account for this gradient in the basic control scheme of the plant has led to oscillatory behavior in the reactor pH, as documented in an operational review of a 12 kt·a⁻¹ facility in the Czech Republic (Ústí nad Labem), where the pH control valve hunting amplitude reached ±0.35 pH units around a set point of 9.2, triggering frequent alarms on the downstream reboiler of the formate concentration evaporator.

The data in Table 2 were generated using a 6-blade Rushton turbine (diameter 0.12 m, 800 rpm) in a vessel of 0.3 m internal diameter, with the CO gas introduced via a ring sparger. The reported k_La was determined by the dynamic pressure step method according to the procedures outlined in ISO 18132-1:2011 for gas–liquid mass transfer measurement. The final conversion was measured by quenching the liquid sample and analyzing residual NaOH via potentiometric titration (0.1 M HCl). These results confirm the non-linear scaling of reactor performance with CO partial pressure and underscore the abrupt deterioration of space-time yield as one approaches the 5 bar threshold. The enhancement factor E was calculated from the ratio of the observed absorption rate to the physical absorption rate in an aqueous sodium formate solution of identical ionic strength, and the increasing trend reflects the gradual transition from kinetic control to a mixed regime as CO availability improves.

In an alternative reactor configuration—a gas-inducing stirred tank operating at 1200 rpm with a hollow-shaft impeller—the k_La at a CO partial pressure of 7 bar reached 0.055 s⁻¹, approaching the value obtained at 15 bar in the conventional sparged vessel. However, the mechanical power consumption rose to 3.8 kW·m⁻³, more than double the baseline, and the shaft seal present in this design must be purged with nitrogen at a rate of 50 L·h⁻¹ at a pressure 0.5 bar above the reactor headspace to prevent CO leakage to the atmosphere, a requirement that introduces an additional source of inert gas into the loop. The trade-off is therefore not between pressure reduction and energy saved, but between compression energy and agitation energy, with the optimal solution strongly dependent on the local electricity-to-steam cost ratio. A 2019 techno-economic analysis by DECHEMA (GDCh Monograph Band 54) concluded that for a hypothetical plant on the lower Rhine with an electricity price of €0.11/kWh and a natural gas price of €25/MWh, the lowest total operating cost for sodium formate synthesis occurred at a CO partial pressure of 9 bar in a conventional bubble column, whereas in a gas-inducing reactor the optimum shifted to 6 bar, providing a narrower but viable window for low-pressure operation.