OD600 GUIDE

OD600 during bacterial growth: growth curves and log phase

How to read lag, exponential, and stationary phases from an OD600 growth curve, calculate doubling time, and know when OD stops tracking viability.

Repeated OD600 readings over time produce a growth curve. Read against the culture’s physiology, the curve tells you where the culture is, how fast it is growing, and when to act on it.

The phases

• Lag. OD600 is flat while cells adapt to the new medium and prepare to divide. The culture is metabolically active — synthesizing enzymes, repairing damage — but not yet increasing in measurable biomass. A long lag often means the inoculum came from a very different medium or a stressed state, not that “nothing is happening.”
• Exponential (log). OD600 rises approximately exponentially. On a log-transformed plot this phase is a straight line; its slope is the specific growth rate under those conditions. This is the phase in which OD-to-cell conversions calibrated in log phase are most trustworthy.
• Stationary. OD600 plateaus as growth and death balance and resources become limiting or products accumulate. Cell number is at its maximum for the batch, but viability may already be falling — the OD plateau hides that decline.
• Decline (death). OD600 falls slowly or stays nearly flat while viable counts drop. Lysis, not death, is what reduces scattering, so the OD curve systematically overstates viability in this phase.

Calculating doubling time

Within the log phase, doubling time follows from any two points:

doubling time (h) = (t₂ − t₁) × log₁₀2 ÷ (log₁₀OD₂ − log₁₀OD₁)

Worked example: OD 0.10 at 2 h and OD 0.40 at 4 h. log₁₀(0.40) − log₁₀(0.10) = 0.602; (4 − 2) × 0.301 ÷ 0.602 = 1.0 h. Points must be genuinely inside the log phase — including a lag-phase point flattens the estimate, and including a stationary-phase point deflates it. When in doubt, fit a line to several log-phase points on a log-transformed plot instead of trusting two.

Inoculating to a target starting OD

To start a culture at OD 0.05 from a dense starter, apply C₁V₁ = C₂V₂: the volume of starter = (target OD × target volume) ÷ starter OD. A 50 mL culture at OD 0.05 from a starter at OD 2.0 needs 0.05 × 50 ÷ 2.0 = 1.25 mL. The OD600 calculator includes an inoculation helper that does this arithmetic and shows the substituted values.

Where OD600 misleads

• Non-linearity at high density. Past roughly 0.8–1.0 on a standard 1 cm instrument, the response compresses. Dilute, re-blank, and apply the dilution correction before comparing points across the range.
• The stationary-to-decline transition. OD barely moves while CFU collapses. If your question is viability, plate counts from the same flask are the companion measurement.
• Morphological changes. Filamentation (common under stress or certain genetic constructs) raises OD disproportionately; clumping lowers it. Neither reflects a true change in cell number.
• Medium evaporation. Over long experiments, evaporation concentrates the culture and lifts OD without growth. Caps, controlled shaking, and a medium-only control well in plate experiments expose the drift. In a 96-well plate, edge wells evaporate fastest — either fill them with sterile water or treat them as sacrificial.

From curve to specific growth rate

Doubling time is the intuitive form; the formal quantity is the specific growth rate μ:

μ (per hour) = (ln OD₂ − ln OD₁) ÷ (t₂ − t₁)
doubling time = ln 2 ÷ μ ≈ 0.693 ÷ μ

For the example above — OD 0.10 at 2 h and OD 0.40 at 4 h: μ = (ln 0.40 − ln 0.10) ÷ 2 = 1.386 ÷ 2 = 0.693 h⁻¹, and the doubling time is 0.693 ÷ 0.693 = 1.0 h, matching the log₁₀ calculation. μ is the quantity to report when comparing growth across conditions, because it composes cleanly: in a mixture of organisms, the observed μ reflects the fastest growing population, which is why a curve alone cannot establish purity.

Include a medium-only control well or tube alongside the experiment. Its reading over time exposes evaporation drift, condensation on lids, and instrument warm-up — artefacts that otherwise appear as slow “growth” in every sample.

Practical experimental design

Blank with the medium used in the experiment, keep cuvette or plate layout consistent, and include biological replicates — technical replicates catch pipetting errors, biological replicates catch the variation that actually matters. When comparing growth rates between conditions, compute μ per replicate first and compare the replicate values, rather than fitting one averaged curve; the spread between biological replicates is the honest error bar. Sample frequently through the expected log phase (every 15–30 minutes for fast growers) and less often elsewhere; missing log phase is the most common design error. Estimate cells/mL from each reading with the OD600 calculator, using a calibration from building an OD600 to cells/mL calibration — and remember that in stationary and decline phases, OD-based estimates diverge ever further from viable counts.

Open the OD600 calculator ↗