The seven stages of chemical engineering technology from concept to industrialization (Issue 12/100) -- Scientific concept validation

Thread Content

This post was last edited by xiouxingzhe on 2026-6-23 15:23. The seven stages of chemical technology from idea to industrialization (Issue 12/100). The birth of innovation: the validation of scientific concepts. Dear friends: Hello everyone! In the previous issue, we discussed the core principle of the second phase: eliminating those approaches that are not viable at the lowest possible cost, before investing heavily. Starting from this issue, we will go into detail on how to do it. Let’s talk about the first filter first: the argumentation of scientific concepts. What is the purpose of scientific concept argumentation? In simple terms, it comes down to one thing: before entering the laboratory, work through the logic on paper first. If it doesn’t work on paper, it’s highly likely it won’t work in the laboratory either. I have more lessons than experience in this area. When working on projects in earlier years, I would rush into the laboratory without thoroughly considering certain ideas, and it wasn’t until a few months later that I realized there was something wrong with the approach. I learned my lesson later on: for any new idea, I first work through it on paper, thoroughly exploring all the possibilities before taking action. In this issue, we will discuss the three steps of scientific concept validation: problem definition, scientific hypothesis, and desktop modeling. I. Problem definition: Turning vague ideas into verifiable scientific questions. When you pick an idea from a pool of creative suggestions, the first thing to do is not to test it, but to articulate it clearly. Vague ideas cannot be verified. “\"Develop a more energy-efficient distillation scheme\" — this is not a scientific issue. “Can a heat pump be used to replace steam in order to reduce the energy consumption of distillation towers by 30%? — That’s what matters. How to make something that’s blurry become clear? I do three things. First, clarify the current situation. What is the mainstream practice in this field of technology at present? Where is its performance ceiling? What is its biggest pain point? Vague terms such as “high energy consumption” or “low yield” should not be used; specific data is required. It’s not about saying “high steam consumption,” but rather about “how many tons of steam are used per ton of product at present, what is the design value, and what is the actual value during operation?” Second, define the goals. If it is accomplished, what would the ideal outcome be? It’s not “better”; it’s “30% reduction in energy consumption” or “the residence time reduced to within X seconds”. The goal should be quantified to such a degree that one can say to themselves, “Reaching this number means success.” Third, clarify the boundaries. Which constraints must be met? Safety red line – highly toxic raw materials must not be used. Environmental protection threshold – emissions of waste gases, wastewater, and solid waste must meet specified standards. Economic red line – raw material costs must not be higher than those of the existing process. Once these three steps are completed, you will have a core question statement in one sentence: \"Can YY goal be achieved through XX method while meeting ZZ constraints?\" ”This sentence serves as the benchmark for all subsequent work. My own habit is to write this sentence on a piece of paper and stick it in front of my desk. When working on experiment design and data analysis later on, take a quick look from time to time – is what I’m doing now answering this question? If not, I’ve gone off track. II. Scientific hypothesis: Building a complete logical chain. Once the problem has been defined, it is time to formulate a scientific hypothesis. Assume it is not “I think it’s fine,” but a complete logical chain: “I believe... because...” Taking the continuous flow process for diphenylamine as an example. This hypothesis was derived step by step in this manner: the diazotization reaction proceeds very rapidly, with a residence time of only sub-seconds ； The mixing and heat transfer rates in the microchannels can be matched to this reaction rate ； The limitation of traditional batch reactors is that slow mixing and heat transfer rates lead to side reactions; therefore, replacing the batch reactor with a microchannel one should improve selectivity. Each step in this logical chain is based on established scientific principles. You can work through it step by step: where did these reaction rate data come from? Have heat transfer calculations for microchannels been done? Is this side reaction mechanism reported in the literature? Here I would like to distinguish between two assumptions. One is a weak assumption. “I think increasing the reaction temperature can improve the yield. ”Why? Because higher temperatures result in faster molecular movement, accelerating the reaction. This is a weak hypothesis—it has no clear scope of application, no falsifiable boundaries. Another is a strong assumption. “I believe raising the reaction temperature from 200 degrees to 220 degrees can increase the conversion rate from 85% to over 92%, while increasing the selectivity of by-products by no more than two percentage points. If that is not the case, it may be that at high temperatures the rate of side reactions increases faster than that of the main reaction, and further experiments are needed to verify this. ”This is a strong hypothesis—it includes specific data predictions and clarifies where the error might occur if there is one. The value of a hypothesis lies not in its being necessarily true, but in its being falsifiable. You can clearly see at which stage it is the most vulnerable, and the next experiment can focus on attacking this weakest point. A hypothesis that cannot be falsified has no value in science. III. Desktop simulation: First perform the thermodynamic calculations and material balance calculations. Once the assumptions are established, there’s no need to go to the laboratory yet. First, work through it on paper. Thermodynamics first challenge: calculate ΔG. If a reaction is thermodynamically impossible to proceed spontaneously, no matter how good the catalyst is or how clever the operating conditions, it’s all in vain. This step can be completed using paper and pen or simple thermodynamics software. A few hours of computation can perhaps avoid months of futile lab work. There is one thing that is easily overlooked when calculating ΔG: it is not only necessary to determine the value under standard conditions, but also to estimate the change in Gibbs free energy under the actual temperature and pressure conditions of operation. Sometimes, for a given reaction, ΔG is negative at room temperature, but becomes positive at higher temperatures – because the effect of the entropy change is amplified at high temperatures. If this change is not taken into account in advance, wrong judgments may be made. The second step in material balance: based on stoichiometric relationships, roughly estimate the consumption of primary materials, the theoretical yield of products, as well as the types and approximate quantities of by-products. No precise simulation is needed; Excel is sufficient. This step will give you a preliminary idea of the cost-effectiveness – if, on a rough estimate, the cost of raw materials is already higher than the market price, then this approach likely needs to be reconsidered. I still remember that in the early stages of the phenylacetic acid project, the consideration in this regard was not thorough enough. A lack of understanding of the complexity of the catalytic system and the network of side reactions led to significant costs later on. Since then, I’ve made a rule for myself: before entering the laboratory, I first work through the logic on paper. Of course, desktop simulations cannot replace experiments; they can only rule out those approaches that are clearly unsound. Whether it will work in the end still requires verification in the laboratory. But it at least gives you a rough idea of the layout before you enter the laboratory. IV. An example of inference: After discussing so many methods, they may still seem abstract. I’ll use my own simulation record as an example—not to prove how correct this example is, but to show fellow enthusiasts what a decent tabletop simulation looks like. The project background is the improvement of the synthesis route for a certain fine chemical intermediate. The current process involves two steps: reaction followed by separation, with a one-way yield of just over 60%, and high steam consumption. The idea is to combine reaction and distillation in one column. Problem definition: Can reactive distillation be used to replace the two-step process, thereby increasing the yield per pass from 65% to over 80%, while reducing the steam consumption to about half of the current level? The constraints are that the temperature must not exceed 180 degrees (otherwise the product will decompose), there must be no increase in waste, and the cost of raw materials must not be higher than that of the existing process. Scientific hypothesis: Since this reaction is a reversible reaction subject to equilibrium constraints, the conversion rate has a ceiling of around 75%. The azeotemperture of the product and water is much lower than the reaction temperature; if the product is continuously distilled out from the top of the tower using this azeotrope, Le Chatelier’s principle dictates that the reaction equilibrium will be continuously shifted in favor of the product, allowing for nearly complete conversion in theory. If this assumption is not valid, the most likely point of failure is that the content of the product in the azeotic composition is too low, and the amount carried away is insufficient to shift the equilibrium ； Or the reaction and distillation temperature ranges do not match. Thermodynamic quick assessment: Under standard conditions, ΔG is approximately -12 kJ per mole; within the operating temperature range, ΔG remains negative, indicating thermodynamic feasibility. Adequate literature on the azeotrope data of the product and water is available, with independent sources to verify it. The product is stable below 180 degrees, and it is safe to keep the reaction section at temperatures below this value. Preliminary material balance: Based on an annual production capacity of 1,000 tons and an yield of 85%, compared to the existing process, there is a reduction in raw material consumption, steam consumption, and the amount of by-products to varying degrees. Economic sensitivity analysis shows that if the yield is below 80%, the advantage is not significant ； The yield exceeds 85%, and the total variable costs have decreased significantly. Conclusion: Thermodynamically feasible and economically promising. The key variable is the yield – whether the yield can reach over 85% determines the economic viability of the project. It is recommended to conduct proof-of-concept experiments next, to verify on a small-scale device that the conversion rate and selectivity do not meet the expected levels; emphasis should be placed on assessing the thermal stability of the product under reactive distillation conditions. This analysis took half a day, mainly spent looking up azeotrope data and literature on thermal stability. But it answers a key question: is this direction worth pursuing in the laboratory? The answer is it’s worth it. Thermodynamics is not violated; economics offers advantages under assumptions of reasonable yields, and the greatest uncertainties can be verified through experiments. If calculations show that even with a yield of 90%, the total variable cost remains higher than that of the existing process, then it is not worth proceeding with experiments in the laboratory. This direction is blocked by desktop simulations outside the laboratory – and this is the value of the innovation incubation phase. V. Regarding novelty search: Assuming it has been created and the simulations have been completed, one final step cannot be skipped: the novelty search. Has a team already worked on your idea? What level have others reached? If there are already public reports, at least you won’t have to go through it all over again for nothing. If no one has done it yet, then this path might be yours. When conducting a novelty search, don’t limit yourself to Chinese content or just one database. Check patent databases, academic searches, and industry reports – look into everything that can be found. I have another habit when conducting a novelty search: I don’t just look at whether someone has already done it, but also at what level they stopped once they did it. Was it that a paper was published and nothing further happened, or was pilot production achieved but industrialization was not pursued? If it is the former, it may mean that the laboratory approach is feasible but there are issues with scaling up. If it is the latter, it may mean that the market did not need it at that time. Are the reasons why others gave up still valid for you? If the market conditions, raw material prices, and policies and regulations change, might paths that were not viable in the past now deserve to be reconsidered? VI. Preview of the Next Issue: Issue 13 – Business Concept Validation: How Large Is the Market? Is It Economically Viable? Even if something is scientifically feasible, another question arises: Is it worth pursuing from a business perspective? Next time, we will discuss how to conduct market analysis, assess the competitive landscape, and perform preliminary cost calculations at the lowest possible cost. High precision is not pursued, but the basic judgment must not be wrong. Table of Contents link: Seven-Stage Model of Chemical Technology from Idea to Industrialization (Issue 100) – Table of Contents (updated in real time)