Technology
Residence time distribution
Residence time distribution describes how long individual elements of fluid spend inside a vessel. It is the difference between the residence time a reactor is designed for and the residence time the material actually experiences — and that difference is what determines conversion, selectivity and impurity profile.
The definition
Danckwerts introduced the formalism in 1953. The exit age distribution, written E(t), is defined so that E(t)dt is the fraction of material leaving the vessel that spent between t and t + dt inside it. By construction the integral of E(t) over all time is one.
Two ideal limits bracket every real vessel. In plug flow every element spends exactly the same time inside, so E(t) is a spike at the nominal residence time. In a perfectly mixed tank the distribution is exponential, with the largest fraction leaving almost immediately. Real vessels sit between the two, and the RTD is the measurement that says where.
What the moments mean
Mean residence time
The first moment of E(t). Compared with volume divided by flow rate, a shorter measured mean means part of the vessel volume is not participating — dead zones.
Variance
The second central moment. It quantifies the spread and feeds both standard models: tanks-in-series N equals mean squared over variance, and the axial dispersion number.
Shape
An early spike means bypass. A long tail means stagnant volume exchanging slowly. Two peaks mean two parallel flow paths — a mechanical fault, not a kinetic one.
How it works
How the measurement is made
-
01
Inject a tracer
A pulse gives E(t) directly; a step gives the cumulative F(t), whose derivative is E(t). The tracer must be inert to the chemistry and hydrodynamically similar to the process fluid, or you measure the tracer's path rather than the process's.
-
02
Detect at the outlet
Concentration against time. Time resolution of the detector sets the resolution of the leading edge, which is where bypass shows up.
-
03
Record the full tail
Run to at least three nominal residence times. Truncating the tail biases both the mean and the variance low and makes the vessel look better than it is.
-
04
Fit and interpret
Normalise to unit area, compute the moments, then fit tanks-in-series or axial dispersion to get a parameter you can carry into a reactor model.
What an RTD lets you do
- Predict conversion for a known kinetic scheme instead of assuming plug flow and hoping.
- Locate dead volume and bypass, then fix the internals that cause them.
- Scale up on a measured hydrodynamic target rather than on a geometric rule of thumb.
- Demonstrate to a regulator that material spends the time in the vessel the process description claims — a requirement for continuous pharmaceutical manufacturing.
- Diagnose why two nominally identical units perform differently.
The raw trace, and the curve fitted to it
On the left of a real run you see the tracer arriving: concentration against clock time, noisy at the baseline and then climbing as the slug reaches the detector. That raw trace is not yet an RTD. Normalising it to unit area and fitting gives the exit age distribution E(t) — the smooth curve — from which the mean residence time and the variance are computed. The software does this at the end of the run and reports the residence time directly.
Watching it happen
Live tracer concentration during the experiment. The flat stretch is baseline before the slug arrives; the rise is the leading edge, which is where bypass would show up as an early spike; the tail is where dead volume lives, and it is the part most often truncated by stopping the run too early.
Measuring RTD on your own reactor
The RTD module runs the experiment from a USB device and generates the plot at the end of the run. Tell us the vessel, the flow rate and the tracer you can use.