The study modeled an integrated Kuwaiti refinery-petrochemical complex using Honeywell UniSIM process simulation and previously developed membrane models. It examined two options: using recovered hydrogen and captured carbon dioxide to increase urea output, or using recovered hydrogen to replace part of the hydrogen generated by steam methane reforming while maintaining existing urea capacity.

For the first option, a polysulfone membrane purified tail gas from an atmosphere residue desulfurization unit. The simulated process recovered 665 kmol/h of hydrogen at 98 mol% purity and 99% recovery. After compression to 160 bar, the hydrogen could support production of an additional 212 metric tons of urea per day, equivalent to 30% of the plant’s original capacity.

The same option required a multistage membrane system to capture carbon dioxide from reformer flue gas. Three membrane stages were modeled to raise carbon dioxide purity from 8.4% to a target of 94 mol%. The process could recover enough carbon dioxide for the additional urea and was estimated to reduce direct carbon dioxide emissions by 34%.

However, the study estimated that electricity required by the membrane compressors would generate additional carbon dioxide at the nearby gas-fired power station. Accounting for those emissions lowered the net reduction to 10%. The combined hydrogen-recovery and carbon-capture configuration was estimated to require $30.8 million in capital expenditure and $4.6 million in annual operating expenditure, with an expected payback period of about three years, including a possible urea-plant upgrade.

The second option used the purified refinery hydrogen to replace some steam-methane-reforming hydrogen without capturing flue-gas carbon dioxide or increasing urea output. The simulation predicted that SMR hydrogen production would fall from about 1,722 to 1,057 kmol/h and reformer carbon dioxide production would decline from 430 to 265 kmol/h.

For this fixed-capacity configuration, direct carbon dioxide emissions were estimated to fall by 39%. Including carbon dioxide attributed to electricity use for the hydrogen compressor, the model estimated a net reduction of 35%. The membrane system was assigned capital expenditure of $4.5 million and annual operating expenditure of $754,000, with an estimated payback period of approximately six months and only minor process modification.

The hydrogen-recovery estimate was $0.092 per kilogram under a 20-year plant-life assumption with no interest rate. Applying a 7% interest rate increased the estimate to $0.111 per kilogram, while the authors reported that the solution’s savings changed by only 2%. The calculations also assumed an industrial electricity tariff of $0.032 per kilowatt-hour and excluded installation, insurance, site preparation and labor costs.

These findings are modeled technical and economic estimates rather than results from an operating demonstration. The analysis was based on one petrochemical plant and its simulation inputs, and the proposed membrane systems were not reported as having been tested at industrial scale.

Subject of Research: Simulation and economic assessment of refinery-hydrogen recycling for carbon dioxide reduction in a urea plant

Article Title: Innovative solutions for CO 2 utilization and reduction in urea plant

Article References: Alqaheem, Y., Alobaid, M., Alkandari, S., Alsuwaileh, F., Almutawa, A., & Nayefeh, L. (2026). Innovative solutions for CO2 utilization and reduction in urea plant. Results in Engineering, 32, Article 113634. https://doi.org/10.1016/j.rineng.2026.113634

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113634

Keywords: process simulation, carbon dioxide capture, hydrogen recovery, urea production, membrane separation, refinery integration, steam methane reforming, economic assessment, Kuwait, Innovative, solutions, utilization

The research team, working at the University of Minnesota under protocols approved by the institution’s animal care committee, isolated primary hepatocytes from Sprague–Dawley rats using a two-step perfusion method. Livers were first flushed with University of Wisconsin solution, then perfused ex vivo with Krebs–Henseleit buffer followed by collagenase to digest the tissue architecture, after which gentle mechanical disruption released the cells. Trypan blue exclusion confirmed viabilities between 95 and 98 percent, ensuring that the starting material was healthy enough for the demanding aggregation process ahead. Each experimental condition drew on hepatocytes from a single isolation, a design choice that minimized inter-isolation variability and allowed the investigators to attribute differences in spheroid morphology specifically to the mechanical and seeding parameters under test.

The formation method itself is strikingly simple. Cells were plated at either 1.0 or 1.5 million cells per milliliter in 20 milliliters of spheroid formation media in rectangular Mayo dishes, which were then placed on rocker platforms housed inside a standard humidified carbon dioxide incubator at 37 degrees Celsius. The team tested five rocking speeds—7, 10, 12, 15, and 18 cycles per minute—over seven days of culture, and to rule out instrument-specific artifacts they ran matched conditions on two different commercial rocker platforms. After modest adjustment of speed, tilt angle, and inoculation density, the two platforms produced comparable spheroid size distributions, suggesting that the underlying physics of the system, rather than any particular device, governs the outcome. Within 24 hours, media was exchanged from spheroid-forming to spheroid culture conditions, and fluorescent live-dead staining with acridine orange and propidium iodide revealed smooth-bordered spherical aggregates with nearly complete viability.

The headline finding is a clean inverse relationship between rocking speed and spheroid size. At day 7, mean spheroid diameters measured 199 micrometers at 7 cycles per minute, 173 at 10, 129 at 12, 90.9 at 15, and just 80.2 micrometers at 18 cycles per minute. Median spheroid areas told the same story, falling from roughly 29,400 square micrometers at the slowest speed to 4,300 square micrometers at the fastest. Statistical analysis using the Kruskal–Wallis test confirmed significant effects of both culture day and a speed-by-day interaction on spheroid area, with post hoc Mann–Whitney comparisons showing the inverse speed-size relationship held at every time point from day 1 through day 7. The physical explanation is intuitive: faster rocking imposes greater shear forces and reduces the contact time available for cells to find and bind one another, while slower motion permits extended cell-cell interaction and larger aggregate growth.

There is, however, a floor to the relationship. At 7 cycles per minute, consistent spheroids did not appear until day 3, and in some dishes they never formed at all, presumably because the kinetic energy available for aggregation fell below a critical threshold. The team also observed a characteristic sinusoidal pattern in spheroid size across the week: aggregates grew for the first several days, peaked around days 3 to 5, and then contracted. At 10 cycles per minute, for example, mean area rose from about 13,750 square micrometers on day 1 to a peak of 26,730 on day 4 before settling back to roughly 22,500 by day 7. The authors attribute this shrinkage not to cell loss but to compaction, as maturing intercellular junctions and cytoskeletal remodeling draw the aggregate into a tighter, denser sphere.

Inoculation density provided the second, independent control axis. Holding speed constant at 15 cycles per minute, cultures seeded at 1.5 million cells per milliliter produced significantly larger spheroids than those seeded at 1.0 million at every time point measured, with day 7 mean areas of 7,203 versus 5,627 square micrometers. A significant density-by-day interaction indicated the magnitude of the effect shifted over the culture period, but the overall temporal pattern of early expansion followed by stabilization was preserved. The mechanism again follows from simple probability: more cells per milliliter means more frequent cell-cell encounters, greater aggregate stability, and ultimately larger and more numerous spheroids. Together, speed and density give labs a two-dimensional parameter space in which a desired spheroid size can be targeted rather than left to chance.

Critically, the team showed that size tuning is biologically meaningful, not merely a morphological curiosity. Urea production normalized to DNA content peaked at intermediate rocking speeds, with a median of about 3.47 micrograms per microgram at 12 cycles per minute, while the slowest condition at 7 cycles per minute produced almost none—consistent with diffusion-limited metabolism in oversized aggregates—and production declined again at the highest speeds, possibly reflecting shear stress or insufficient cell-cell signaling. Albumin and urea output both climbed steadily over ten days of culture, with albumin rising from a median of roughly 1,087 nanograms per microgram DNA on day 1 to about 1,534 by day 10. When converted to per-cell terms, these spheroids secreted an estimated 7 to 9 micrograms of albumin per million cells per day, several-fold higher than reported values for static two-dimensional hepatocyte cultures, and urea output by day 7 exceeded that of conventional hepatocyte monolayers.

Efficiency and variability metrics rounded out the practical picture. Volumetric analysis on day 4 estimated that a median of 86 percent of seeded hepatocyte volume had been incorporated into spheroids, with four of five datasets clustering between 73 and 95 percent, the lower values likely reflecting occasional cell adherence to dish walls and the tendency of compacting spheroids to be undersized by diameter-based volume estimates. Size dispersion, quantified by the coefficient of quartile variation, was widest on day 1—as expected during initial aggregation—and converged to a stable 21 to 24 percent by day 7 across all speeds, a range the authors note is acceptable for bioartificial liver applications even though static microwell methods can achieve tighter distributions on the order of plus or minus 10 to 20 micrometers.

The authors are candid about limitations. All experiments used primary rat hepatocytes, and interspecies differences in cytochrome P450 expression, metabolic capacity, and cell-cell dynamics mean extrapolation to human cells requires caution, though the underlying physical principles should generalize. Size was quantified from two-dimensional images assuming ellipsoidal geometry, and long-term culture beyond seven days was not assessed. Absolute parameter values may also need recalibration on different rocker platforms. Still, the study’s contribution is a framework rather than a recipe: by demonstrating that two universally available variables reproducibly govern spheroid dimensions while preserving metabolic function, the work hands bioartificial liver engineers, toxicologists, and regenerative medicine researchers a scalable, inexpensive tool for standardizing one of the field’s most important building blocks—no specialized equipment required.

Subject of Research: Tunable control of hepatic spheroid size through rocking speed and cell inoculation density

Article Title: The effect of rocking speed and cell inoculation density on hepatic spheroid size: A simple method for size modulation

Article References: Megaly, M. G., Tobolt, D., Chen, A. A.-F. D., Namsrai, B.-E., Fisher, B., Rao, J. S., Ramesh, S., Chimedtseren, A., Nyberg, S. L., Clemens, M., Lee, C. Y., Bischof, J. C., & Finger, E. B. (2026). The effect of rocking speed and cell inoculation density on hepatic spheroid size: A simple method for size modulation. Bioengineering & Translational Medicine, 11(5), Article e70154. https://doi.org/10.1002/btm2.70154

Image Credits: AI Generated

DOI: 10.1002/btm2.70154

Keywords: hepatic spheroids, bioartificial liver, hepatocytes, rocking speed, inoculation density, spheroid size, drug toxicity testing, regenerative medicine, 3D cell culture, urea production, albumin secretion, shear stress