For decades, analytical chemists have accepted a quiet truth about their workflows: the mass spectrometer isn't usually what fails you. It's everything upstream of it. In the latest episode of Concentrating on Chromatography, host David Oliva sat down with Mason Chilmonczyk, PhD, mechanical engineer and co-founder of Andson Biotech, to talk about why sample preparation — not instrumentation — remains the single biggest obstacle in LC-MS and native mass spectrometry workflows today, and what his team is doing to fix it at the plumbing level.
Watch the full episode on YouTube.
Mason's path into mass spectrometry is unusual. He's a mechanical engineer "all the way through" — bachelor's, master's, and PhD, all from Georgia Tech — which he argues is precisely why he was able to see the field's biggest inefficiency clearly. As he put it on the podcast, outsiders to a field often bring exactly the kind of fresh perspective that unlocks stalled progress.
Ask most analytical chemists what slows down their nanoelectrospray ionization (nano-ESI) workflows, and the answer usually isn't the chemistry. It's mechanical failure. Mason summed up his entire graduate research experience in three words: bubbles, clogs, and leaks.
Nano-ESI is prized for its low sample consumption, increased sensitivity, and enhanced ionization efficiency compared to conventional electrospray ionization — advantages that make it a preferred technique for analyzing large biomolecules. But those same ultra-low flow rates and micron-scale capillaries that give nano-ESI its sensitivity also make it notoriously fragile. A single air bubble, a partial clog, or a hairline leak can derail hours of work, and troubleshooting requires simultaneous expertise across the mass spec, the LC system, the column, and the sample itself.
"LC-MS stacks are complicated technologies," Mason explained. "Relatively simple once you get it going, but if you dive in and look at all the things going on, it's relatively complicated... and then you have a freaking leak somewhere and you can't find it." That single line captures why so many labs quietly accept plumbing failures as a fixed cost of doing science, rather than a solvable engineering problem.
The stakes rise considerably in native mass spectrometry, where the goal isn't just to detect a molecule but to preserve its structure well enough to study its secondary, tertiary, and even quaternary architecture in as close to a physiological state as possible. Native MS has become an increasingly important complement to structural biology techniques like X-ray crystallography, NMR, and cryo-EM — particularly for large, heterogeneous protein complexes that are difficult to study using those older methods.
The problem, according to Mason, is that traditional sample cleanup for native MS — offline buffer exchange, two-times spin columns — treats sample prep and ionization as separate, static problems rather than an integrated challenge. Static spray-based ionization for native MS can itself introduce structural artifacts, meaning a lab can spend hours cleaning a sample only to distort it at the exact moment it enters the instrument. As Mason put it, "there really hasn't been that many compelling methods that simultaneously clean up the sample while preserving native structure to the level of nativeness that these researchers want."
This is a genuine, actively studied limitation in the field. Reviews of native MS methodology point to sample preparation, instrumentation, and activation methods as ongoing areas needing improvement even as the underlying technique matures.
Mason's answer to this problem grew out of his PhD research in Andrei Fedorov's mechanical engineering lab at Georgia Tech, where he worked on miniaturizing a decades-old bioprocessing technique called tangential flow filtration (TFF).
In conventional TFF, a feed stream carrying the sample flows parallel — tangentially — across a membrane, while a separate buffer stream flows on the other side. This tangential geometry is what makes TFF valuable in downstream bioprocessing in the first place: because the fluid sweeps across the membrane rather than driving straight into it, TFF avoids the "filter cake" buildup and clogging that plagues normal flow filtration, while remaining gentle on shear-sensitive biomolecules.
Mason's innovation was shrinking that architecture down to the microfluidic scale. In Andson Biotech's DynaChip, the sample channel is just five microns tall — twenty times narrower than a human hair — separated from a buffer channel by a membrane only microns thick. At that scale, mass transfer approaches a pure diffusion regime, which the team leveraged to strip contaminants like salt out of a sample almost instantaneously. In Mason's words, the DynaChip can remove up to five molar sodium chloride from a sample in about one second, compressing a process that traditionally takes minutes to hours into a near-instant step.
The result, now marketed as the DynaChip X1 platform, functions as an inline, front-end sample prep module that sits between an injection and the mass spec itself — cleaning up samples in real time rather than requiring a separate offline prep step. Andson has demonstrated the platform at recent American Society for Mass Spectrometry (ASMS) conferences, where it's been positioned as a way to accelerate workflows that traditionally take hours down to just minutes.
Mason is candid about the trade-off: the DynaChip doesn't perform chromatographic separation, since everything is injected together. For complex mixtures requiring fine separation, that's a real limitation. But for the many workflows where the bigger risk is a fragile sample degrading during hours of manual prep, an instant, repeatable inline cleanup can outperform "perfect" separation that arrives too slowly or too inconsistently.
One of the more striking claims in the conversation was about accessibility. Mason described a recent intern at Andson Biotech who had never even seen a mass spectrometer before joining the company. Six weeks later, she was running the instrument independently, generating quality assurance data on glycoform resolution for the company's consumables. By contrast, Mason noted it took him roughly eighteen months into his own PhD to generate his first usable mass spec data.
That gap is the practical definition of what Mason means by "democratizing" mass spectrometry: engineering a system with embedded sensors and guided troubleshooting routines so that operators don't need years of hands-on experience to trust their data. It's a meaningful shift for a field where instrument expertise has traditionally been a multi-year investment, and it echoes his broader thesis that innovation often stalls not because people are attached to the status quo, but because they simply haven't been shown that a faster, more accessible alternative exists.
Perhaps the most forward-looking part of the conversation focused on cell therapy manufacturing, which Mason described as "the Wild West" of biopharma analytics. Unlike monoclonal antibodies, where the manufacturing process is relatively well-characterized, cell therapies behave more like living systems — batch outcomes can vary significantly even when starting material and protocols are nominally identical.
This isn't just anecdotal. Published analyses of cell and gene therapy manufacturing confirm that a lack of reliable, standardized potency assays remains a universal challenge across the field, compounded by donor-to-donor cell variability and the difficulty of designing in vitro assays that actually represent a therapy's real mechanism of action in the body.
Mason's proposed solution, developed in part through NIH-funded research at Andson Biotech, is to couple real-time mass spectrometry directly to bioreactors — capturing the temporal evolution of biomolecules that cells secrete as they grow and differentiate. This builds on a growing body of evidence that a cell's secretome — the mix of cytokines, growth factors, and other molecules it releases — can serve as a meaningful, measurable proxy for potency and therapeutic efficacy. Research into mesenchymal stromal cell (MSC) therapies, for instance, has identified specific secreted factors like VEGF, IL-8, and IFN-γ that correlate with immunomodulatory and regenerative potency, suggesting that dynamic, untargeted secretome monitoring could eventually predict batch quality before a therapy ever reaches a patient.
The key word Mason emphasized is untargeted. Rather than requiring scientists to guess in advance which cytokine or biomarker matters, mass spec can survey the full secreted profile of a bioreactor in real time — potentially closing what he described as a communication gap between cell therapy developers who know which biological outcomes matter and analytical chemists who don't yet know what to look for in a given mixture.
Not every organization is equally enthusiastic. Mason noted that attitudes toward both native mass spec and cell therapy investment vary widely across pharma companies — some are "all in," while others have scaled back or never built out a dedicated division at all. That divide, he suggested, has less to do with the strength of the science and more to do with organizational risk tolerance and where a given company sees its near-term return on investment.
Sample prep failures in nano-ESI and native MS are largely mechanical and engineering problems — bubbles, clogs, and leaks — not chemistry problems, which means they're solvable with better hardware design.
Miniaturized tangential flow filtration can compress sample cleanup from hours to seconds while preserving the native structure of sensitive biomolecules.
Making mass spec workflows genuinely accessible to non-experts, not just faster for experts, may be the bigger unlock for scientific throughput industry-wide.
Cell therapy manufacturing's biggest gap isn't better instruments — it's a lack of reliable, real-time potency indicators that can be tracked as therapies are made, not just after the fact.
For the full conversation — including Mason's take on Graham Cooks' vision for mass spec's future and why "native" mass spec may never be fully native — watch the complete episode of Concentrating on Chromatography.