REACTION ENGINEERING · WAITLIST
Stop at ten.
Conversion is not yield.
A turns into B, and B keeps turning into C. Conversion only counts the A you used up. Drag the time cursor and watch the product peak, then drain away, while conversion keeps climbing.
Run the series reaction
Teaching modelk₁ = ln 2 / 5.0 min = 0.1386 per min · k₂ = ln 2 / 10.0 min = 0.0693 per min · t_opt = ln(k₂ / k₁) / (k₂ − k₁) = 10.0 min
Every curve is a fraction of the A fed. Conversion is A consumed, 1 − a. Yield of B is b, the share of the feed that is B right now. The dashed line marks the optimum time; the solid line is your cursor.
TWO HALF-LIVES AND A CLOCK
Run to 40 minutes, then stop at the peak. Same reactor, same feed: how much product did the extra half hour cost?
At 40.0 min conversion reads 99.6 percent, yet B is only 11.7 percent of the feed (about 12 percent). The peak was 50.0 percent at 10.0 min; 76.6 percent of that peak, about three quarters of it, has since been destroyed and is now C. Stopping at 10.0 min would have kept 4.3 times as much product.
Assumptions and limits
Both steps are first order and irreversible, the reactor is isothermal, and the model is an ideal batch reactor or an ideal plug-flow reactor read in residence time. A continuous stirred tank with the same rates peaks lower, because it holds every molecule for a different time. Conversion is A consumed; yield is B per A fed. The 50 percent peak holds only when k₂ = k₁ / 2, as in the video; other rate ratios give other peaks. When the two half-lives coincide the page uses the k₁ = k₂ limit, b = k₁t e^(−k₁t), which peaks at t = 1 / k₁. At large scale uneven mixing and slow heating spread the residence time, so the real peak shifts and flattens. That is what a pilot run measures.
WHAT THE NUMBERS MEANT
More conversion made less product.
The reactor did what you asked.
Conversion is the share of A consumed. Yield is the share of A that became B and is still B. Between them sits selectivity, and in a series reaction selectivity falls with time, because the only way to lose B is to leave it in the reactor.
At the video's rates the peak is 50 percent of the feed at 10 minutes. Run on to 40 minutes and conversion reads 99.6 percent while B has fallen to 11.7 percent: three quarters of the best case destroyed by a number that looked like success. The fix is not a better catalyst. It is a clock.
At scale the clock gets harder to read. Uneven mixing and slow heating spread the time each molecule spends inside, so the peak shifts and flattens. That is one of the things a pilot run exists to measure.
FROM THE FLASK TO THE PILOT
Taking a reaction
beyond the lab?
Jon and I are building RARE Labs toward an autonomous pilot plant builder with AI and robotics. Join the waitlist and tell us what you are trying to scale.
FOLLOW THE EVIDENCE
Sources and model limits
- Fogler, Elements of Chemical Reaction Engineering, 7eChapter 8: derivation for a series reaction ↗
Integrates dC_B/dt + k₂C_B = k₁C_A0 e^(−k₁t) to the C_B expression this page plots.
- Fogler, 7e, Chapter 8Derivation for optimum yield ↗
Sets dC_B/dt = 0 to get t_opt = ln(k₂/k₁) / (k₂ − k₁). The peak B_max = (k₁/k₂)^(k₂/(k₂ − k₁)) follows by substituting t_opt back into C_B.
- MIT OpenCourseWare 5.60, Spring 2008Lecture 33: series reactions and their limits ↗
The A to B to C mechanism solved exactly, then the k₁ ≫ k₂ and k₁ ≪ k₂ limits in the buckets-and-pipes picture.