Separable differential equations
When separation is needed
- can be integrated directly, because the right-hand side contains only .
- cannot — the right-hand side contains , and you cannot integrate with respect to without knowing what is.
- Separation of variables rearranges the equation so that each side contains only one variable.
The method
- Separate — get all the terms with on one side, and all the terms with on the other.
- Integrate both sides, each with respect to its own variable.
- Include one constant, conventionally on the side.
- Apply the initial condition to find the constant.
- Rearrange to make the subject if the question asks for it.
The standard case:
-
This is by far the most common form, and it deserves to be known by heart.
-
Separate:
- Integrate both sides:
- Exponentiate both sides to free :
- is just another positive constant. Absorbing the sign as well and calling it :
-
Each part:
- — the value of when , since
- — the rate constant; positive means growth, negative means decay
-
This result is worth memorising. Any equation of the form "rate of change is proportional to the amount present" has an exponential solution.
Handling the constant carefully
- The constant can be carried in two equivalent ways:
- Leave it as inside the logarithm stage, then exponentiate:
- Convert to a multiplicative constant immediately
- Be consistent. The 2025 report warns: "care must be taken to use the constant of integration consistently throughout a solution."
- Do not write . Exponentiating a sum gives a product, not a sum — that is the single most common error in this topic.
Separating when both sides are more complicated
- Worked through — solve with when :
- Separate:
- Integrate:
- Apply the condition: , so
- Rearrange: , so
- Note that the positive root is chosen because at the initial condition, and is continuous.
Choosing the sign of a square root
- Separation often produces , and the square root has two branches.
- Use the initial condition to decide which branch applies, and say why.
- The solution cannot jump between branches, because it would have to pass through , where the original equation is usually undefined.
Two conditions: finding as well
- Applied questions often give two pieces of information, because both and are unknown:
- The first (usually at ) gives .
- The second (a later measurement) gives , and finding it needs logarithms.
- The 2026 specification notes that "solving differential equations may require manipulation of exponents and logarithms".
Worked ExampleSeparating and solving with two conditions
A quantity satisfies , where is in hours. Initially . Find in terms of , and find how long it takes for to fall to 40.
Step 1 — Check whether direct integration works
The right-hand side contains , so we cannot integrate with respect to directly. The variables must be separated first.
Step 2 — Separate the variables
Divide both sides by and multiply by , so each side has only one variable:
Step 3 — Integrate both sides
Each side is integrated with respect to its own variable:
Step 4 — Exponentiate to free
Raise to the power of both sides, remembering that a sum in the exponent becomes a product:
Replace by a single constant , and drop the modulus since is a positive physical quantity:
Step 5 — Apply the initial condition
"Initially " means when . Since :
Step 6 — Set up the second part
We want the time at which :
Step 7 — Isolate the exponential
Divide both sides by 250:
Step 8 — Take natural logarithms
Apply to both sides, using :
Step 9 — Solve for
Step 10 — Check
Substituting back:
And differentiating the solution:
which is the original differential equation.