Why worry and rust are the same process, solved

Rust eats steel by a slow reaction that only runs while the surface is wet. Worry wears a body down by a slow activation that only runs while the mind keeps it open. The maths of exposure time is the same.

Heavy orange rust on the links of a chain left in salty sea spray near the Golden Gate Bridge
Photo: Marlith, CC BY-SA 3.0, via Wikimedia Commons

Worry and rust are not the same chemistry, but they follow the same damage law: a small, steady reaction that runs only while a surface stays exposed, so the harm grows with exposure time more than with intensity. Steel in coastal air can lose 80 to 200 µm a year because salt and moisture keep its surface wet for most of the year. Worry does its damage the same way, by keeping blood pressure, heart rate and cortisol raised long after the trigger has passed.

Key takeaways
  • Damage = rate × time: both rust and worry are exposure problems, not event problems.
  • 1.3 to 200 µm/yr: how fast carbon steel corrodes across ISO 9223 categories C1 to C5.
  • g = 0.45 to 0.51: rise in blood pressure while people worry or ruminate in lab studies (Ottaviani 2015, 60 studies).
  • Time of wetness is the master variable for rust; time spent worrying is the master variable for strain.
  • 40% recovery: generalised anxiety patients who postponed worry to a fixed slot (Krzikalla 2024).

This is a collision post: two things with no business in the same sentence, tested for a real link. First the corrosion science, worked through with numbers. Then the psychology of worry and rumination. Then a verdict on how far the comparison actually holds, and where it breaks.

What does rust have to do with worry?

Both are slow, low-level processes that do almost nothing in a single moment and a great deal over months. A drop of seawater on a steel rail does no visible harm in a minute. A worried thought about an exam does no visible harm in a minute either. The damage comes from the process staying switched on.

The comparison only earns its place if the same mathematics fits both. So the test in this post is strict:

  • Is there a rate that depends on conditions (salt, moisture; uncertainty, threat)?
  • Is there an exposure time that multiplies the rate?
  • Does protection work by cutting exposure rather than by fighting the reaction head-on?
The story

On 15 December 1967, during the evening rush, the Silver Bridge between Point Pleasant, West Virginia and Kanauga, Ohio fell into the Ohio River. Forty-six people died and 32 vehicles went down with it. The bridge had stood since 1928. Investigators traced the failure to a single eyebar, number 330, in the suspension chain. It carried a crack only about 2.5 mm deep, grown slowly by corrosion inside a joint that no inspector could see. Once that one link snapped, the rest of the chain followed in seconds. The disaster led to the first national bridge inspection programme in the United States, created by the Federal-Aid Highway Act of 1968.

Wikipedia: Silver Bridge (summarising the 1970 NTSB report)

The point of the story is not the collapse. It is the quiet before it: forty years of a tiny reaction running inside a joint, unnoticed because nothing dramatic ever happened on any single day.

Archive photo of the Silver Bridge over the Ohio River before its 1967 collapse
The Silver Bridge, opened in 1928. One corroded eyebar with a crack about 2.5 mm deep brought it down in 1967.Photo: Unknown, No restrictions, via Wikimedia Commons

How does iron actually rust?

Rust is an electrochemical cell that builds itself on any wet steel surface: one patch of metal gives up electrons, another patch hands them to oxygen, and the water film carries the ions between them. No water film, no circuit, no rust.

The two half-reactions

  • Anode (where metal is lost): \( \mathrm{Fe \rightarrow Fe^{2+} + 2e^-} \)
  • Cathode (where oxygen is reduced): \( \mathrm{O_2 + 2H_2O + 4e^- \rightarrow 4OH^-} \)
  • The iron and hydroxide ions meet in the water film and oxidise further into the hydrated iron oxides we call rust.

Salt matters because chloride makes the water film a far better conductor, and it keeps the surface damp at lower humidity. That is why a chain near the sea looks like the cover photo and the same chain in a dry inland garage does not.

Rust is a tiny battery inside a water dropsteelwater film + saltanode: Fe → Fe²⁺ + 2e⁻cathode at the edge:O₂ + 2H₂O + 4e⁻ → 4OH⁻O₂ enters mostwhere film is thinelectrons flow through the metal to the edges
Under a drop of salty water, steel at the centre dissolves (anode) while oxygen is reduced at the edge, where the film is thinnest. Dry the drop and the circuit stops. Drawing by the author from standard corrosion chemistry.

How fast does rust eat steel, in numbers?

From about 1.3 µm a year in very dry, clean air to 200 µm a year in the harshest marine and industrial air, according to the international corrosivity standard ISO 9223. Above that sits an "extreme" category reaching 700 µm.

ISO 9223 categoryRiskCarbon steel, first-year lossZinc, first-year loss
C1Very lowup to 1.3 µmup to 0.1 µm
C2Low1.3 to 25 µm0.1 to 0.7 µm
C3Medium25 to 50 µm0.7 to 2.1 µm
C4High50 to 80 µm2.1 to 4.2 µm
C5Very high80 to 200 µm4.2 to 8.4 µm
CXExtreme200 to 700 µm8.4 to 25 µm

The zinc column is the quiet hero of this table. Zinc corrodes 10 to 30 times more slowly than steel in the same air, which is why galvanised steel lasts. Remember that ratio; it comes back when we get to protecting a mind.

Worked derivation: from a tiny current to lost metal

Because rust is a battery, its speed can be written as an electric current, and Faraday's law of electrolysis converts current into metal lost. In plain words: every two electrons that leave carry away one iron atom.

\[ \text{penetration rate} = \frac{i\,M}{n\,F\,\rho} \]

Here \( i \) is corrosion current density, \( M = 55.85 \) g/mol for iron, \( n = 2 \) electrons per atom, \( F = 96{,}485 \) C/mol and \( \rho = 7.87 \) g/cm³.

  1. Pick a small current. \( i = 1\ \mu\text{A/cm}^2 = 1\times10^{-6}\ \text{A/cm}^2 \), far less than a phone LED draws.
  2. Mass lost per second. \( \dfrac{10^{-6}\times 55.85}{2\times 96{,}485} = 2.89\times10^{-10}\ \text{g cm}^{-2}\text{s}^{-1} \)
  3. Thickness per second. Divide by density: \( 2.89\times10^{-10}/7.87 = 3.68\times10^{-11}\ \text{cm/s} \)
  4. Per year. Multiply by \( 3.156\times10^{7} \) s: \( 1.16\times10^{-3}\ \text{cm} = 11.6\ \mu\text{m/yr} \)
  5. Read it. One microamp per square centimetre, too small to feel, removes about 12 µm of steel a year. A C5 coastal site at 150 µm/yr is running roughly 13 µA/cm² for every hour the surface is wet.

So rust is never loud. It is a whisper of current that never stops while the conditions hold.

What decides how much a surface corrodes?

Mostly how long the surface stays wet and what is dissolved in that wetness. A 2017 analysis that pooled three worldwide exposure programmes turned this into simple damage equations.

Annual Atmospheric Corrosion of Carbon Steel Worldwide. An Integration of ISOCORRAG, ICP/UNECE and MICAT DatabasesBelén Chico, Daniel de la Fuente, Iván Díaz, Joaquín Simancas, Manuel Morcillo · Materials · 2017

The authors merged first-year steel coupon data from three international programmes covering 38 countries and fitted damage functions. For marine sites (206 data points) the best fit was C = −24.50 + 0.75 Cl + 0.67 SO₂ + 77.32 TOW, in µm per year, where Cl and SO₂ are deposition rates in mg/m² per day and TOW is the fraction of the year the surface is wet. For inland sites (333 points), SO₂ alone explained most of the variance. Limit: the marine fit explains only 47% of the variance, and the authors warn against using it far outside the measured range.

Read the paper (free)

The paper defines time of wetness the way ISO does: the hours in a year when relative humidity is above 80% and the temperature is above 0°C. That single fraction carries a coefficient of 77.32 µm. Every extra tenth of the year spent wet adds about 7.7 µm of lost steel.

Sensitivity check: what moves the answer most?

Take an illustrative coastal Indian site. These inputs are my own assumptions for a monsoon coast, not measured values: chloride 60 mg/m²·day, SO₂ 20 mg/m²·day, wet 60% of the year.

\[ C = -24.50 + 0.75(60) + 0.67(20) + 77.32(0.60) = 80.3\ \mu\text{m/yr} \]
  • Halve the salt (Cl = 30): C falls to 57.8 µm/yr, a drop of 22.5.
  • Halve the wet time (TOW = 0.30): C falls to 57.1 µm/yr, a drop of 23.2.
  • Halve the SO₂ (10): C falls to 73.6 µm/yr, a drop of 6.7.

Cutting the hours of exposure does as much as cutting the aggressive chemical. That is the bridge to worry: you often cannot remove the stressor, but you can shorten how long the surface stays wet.

Use this to check what you just read. Change the salt, the pollution and the wet fraction for your city.

How fast would bare steel rust where you live?

Close-up of rust pits eaten into the surface of a riveted steel water pipe
Rust pits in a riveted steel water pipe. Each pit is a spot where the anode stayed put and kept dissolving, year after year.Photo: State Government Photographer, CC0, via Wikimedia Commons

What does worry do to the body over time?

Worry keeps the stress response running after the stressor has gone. The perseverative cognition hypothesis, proposed by Jos Brosschot, William Gerin and Julian Thayer in 2006, says that what harms health is not the stressful event itself but the time we spend mentally replaying it or rehearsing the future, because that is when the body stays activated.

Worry is the stressor kept switched on after it has left the room.The idea behind the perseverative cognition hypothesis, in my words
Physiological concomitants of perseverative cognition: A systematic review and meta-analysisCristina Ottaviani, Julian F. Thayer, Bart Verkuil, Antonia Lonigro, Bernadette Medea, Alessandro Couyoumdjian, Jos F. Brosschot · Psychological Bulletin · 2016

A meta-analysis of 60 studies of healthy people. During induced worry or rumination in experiments, systolic blood pressure rose (g = 0.45), diastolic pressure rose (g = 0.51), heart rate rose (g = 0.28) and cortisol rose (g = 0.36), while heart rate variability fell (g = 0.15). Correlational studies of everyday worry showed the same direction for heart rate (0.20), cortisol (0.32) and heart rate variability (0.27). Limit: these are short-term readings; the step from repeated activation to long-term disease is inferred, and the blood pressure results showed signs of publication bias.

Read the abstract
How much worry and rumination shift the body (effect size g)
See the data as a table
Effect size g (g)
Diastolic blood pressure (lab)0.51
Systolic blood pressure (lab)0.45
Cortisol (lab)0.36
Cortisol (daily life)0.32
Heart rate (lab)0.28
Lower heart rate variability (daily life)0.27
Heart rate (daily life)0.2
Lower heart rate variability (lab)0.15

Source: Ottaviani et al. 2016, Psychological Bulletin (60 studies) · g of 0.2 is small, 0.5 medium. Lab = worry induced in experiments; daily life = correlational studies.

A blood pressure shift of g = 0.5 for a few minutes is harmless. The concern is duration: a mind that keeps the loop open for hours a day is holding that shift for hours a day. That is time of wetness, for the cardiovascular system.

Worry also changes behaviour. A 2016 meta-analysis of 19 studies by Clancy and colleagues found rumination linked to more smoking, drinking and unhealthy eating (r = 0.122), while plain worry had a weaker link (r = 0.048). Rumination, like salt, makes the film more aggressive.

Is it really the same equation?

Yes, at the level of dose and exposure, and no at the level of mechanism. Here is the shared model, stated plainly.

Assumptions. Damage accumulates as the integral of a rate over the time the process is active. The rate depends on conditions; outside active periods it is close to zero.

\[ D = \int_0^{T} r(t)\,dt \;\approx\; \bar r \times t_{\text{active}} \]
  • For steel: \( \bar r \) is the corrosion rate while wet (set by chloride and SO₂), and \( t_{\text{active}} \) is time of wetness. Chico's 77.32 µm coefficient on TOW is this term measured.
  • For a person: \( \bar r \) is the physiological activation while worrying (Ottaviani's g values), and \( t_{\text{active}} \) is the time spent in perseverative thought.

The sensitivity result transfers directly. Halving the hours you spend in worry cuts the dose by the same fraction as halving its intensity, and hours are often easier to change than intensity.

One damage law, two surfacesSteelMind and bodyExposed surfacebare metalan open, unresolved concernActive filmwater + chloriderepetitive negative thoughtRate while active~12 µm/yr per µA/cm²BP up g ≈ 0.5, HRV downExposure timetime of wetnesshours spent worryingProtectionpaint, zinc, dryingworry window, planningdamage ≈ rate while active × time active
The shared structure. Each row is supported by a source in this post; the mapping itself is the author's analysis.

Why do both get worse once they start?

Because the product of the process feeds the process. Rust is porous and holds water, so a rusted spot stays wet longer than clean steel, raising its own time of wetness. That is how a few pits become the pocked pipe in the photo above.

Rumination has the same self-feeding shape. In the Clancy meta-analysis it was rumination, the backward-looking replay, that tracked worse health behaviour, and those behaviours (poor sleep, alcohol) leave a person more reactive the next day. A 2025 study in Stress and Health of 146 people found that high ruminators showed less systolic blood pressure adaptation when the same stressor was repeated: their bodies did not calm down on the second exposure the way low ruminators' did.

How do you protect a surface, metal or mental?

Engineers do not try to stop oxygen existing. They cut exposure: paint keeps water off, galvanising puts a slower-corroding zinc layer in front, drainage dries the surface, and weathering steel grows a dense rust skin that seals itself.

Close-up of the dense brown protective rust layer on a weathering steel sculpture after two years outdoors
Weathering (Cor-Ten) steel after two years outside. Its tight rust layer slows further corrosion, a case where controlled exposure builds protection.Photo: No machine-readable author provided. Gerardus assumed (based on copyright claims)., Public domain, via Wikimedia Commons

The best-tested mental equivalent is worry postponement: you do not try to stop the worry, you move it to a scheduled slot, which shrinks the hours the film is wet.

Worry Postponement From the Metacognitive Perspective: A Randomized Waitlist-Controlled TrialChristina Krzikalla, Ulrike Buhlmann, Jan Schug and colleagues · Clinical Psychology in Europe · 2024

Eighty-two people, 47 with generalised anxiety disorder and 35 with hypochondriasis, were randomised to a two-session worry postponement intervention or a waitlist. Over six days they practised postponing worry to a fixed later time. Worry fell significantly in the treated anxiety group, with a large pre-post effect, a 40% recovery rate and gains that held at a four-week follow-up. Limit: small samples; the hypochondriasis group showed little benefit and its randomisation was partly unbalanced.

Read the paper (free)
  1. Name the slot. Pick a 15 to 30 minute worry window at the same time daily, not near bedtime.
  2. Notice and park. When a worry arrives outside the window, write one line about it and tell yourself you will deal with it at the slot.
  3. Return to the task. Bring attention back to what you were doing; this is the "drying" step.
  4. Use the window. At the slot, look at the list. Solve what can be solved, plan what needs planning, and let the rest go.
  5. Expect fewer items. Many parked worries feel smaller by the time the slot comes, which is the point the metacognitive model makes: worry is more controllable than it feels.

Other ways of shortening exposure include slow breathing, which lowers the arousal that worry raises (see slow breathing and the vagus nerve), and sleep techniques for the night-time worrier in the CBT-I sleep guide. Mind-wandering into worry has its own brain network, covered in the default mode network. For the steel side of the story, the monsoon maintenance guide shows the same exposure logic applied to a bike.

FeatureRustWorrySame?
Damage per momentTiny (µA/cm²)Tiny (minutes of raised BP)Yes
Master variableTime of wetnessTime spent worryingYes
AccelerantChloride, SO₂Rumination, poor sleepYes, in form
Self-feedingPorous rust holds waterRumination worsens next-day reactivityYes
Reversible?NoLargely yesNo
ChemistryOxidation of ironNeural and hormonal activationNo
Verdict: the link holds as a law of exposure, not as chemistry

Rust and worry are not one process at the molecular level, but both obey damage ≈ rate × time active. For steel, Chico's damage function puts 77.32 µm per year on time of wetness, so halving wet hours at a coastal site cuts loss by about 23 µm a year, as much as halving the salt. For people, worry raises blood pressure by g ≈ 0.5 while it runs, so the hours it runs set the dose. The comparison breaks on reversibility: lost iron is gone, while a mind's strain falls once the loop closes. Protect both by cutting exposure.

What if next?

Questions people ask

Is worrying actually bad for your health?

Short bursts are normal. The concern is prolonged worry: Ottaviani's meta-analysis of 60 studies links it to higher blood pressure, heart rate and cortisol and lower heart rate variability while it lasts. Whether that adds up to disease is inferred, not proven.

What is the difference between worry and rumination?

Worry is mostly about the future ("what if this goes wrong"); rumination is mostly about the past ("why did this happen to me"). Both are perseverative cognition, but rumination showed the stronger link with unhealthy behaviours in the Clancy meta-analysis.

Why does rust spread faster near the sea?

Sea salt deposits chloride, which makes the surface water film a better electrolyte and keeps it wet at lower humidity. In Chico's marine equation, every extra 1 mg/m² per day of chloride adds 0.75 µm a year of steel loss.

Does scheduling worry time really work?

In a 2024 randomised trial, worry postponement reduced worry in people with generalised anxiety disorder, with a 40% recovery rate after two sessions. It is a self-help technique, not a substitute for treatment if worry is severe.

Sources

  1. Chico B. et al.: Annual Atmospheric Corrosion of Carbon Steel Worldwide, Materials (2017) (source for the marine damage function, the 77.32 µm TOW coefficient and the TOW definition)
  2. Šmak M. et al.: The Influence of Hot-Dip Galvanizing on the Mechanical Properties of High-Strength Steels, Materials (2021) (source for the ISO 9223 corrosivity table for steel and zinc)
  3. Ottaviani C. et al.: Physiological concomitants of perseverative cognition, Psychological Bulletin (2016) (source for the effect sizes in the chart)
  4. Clancy F. et al.: Perseverative Cognition and Health Behaviors, Frontiers in Human Neuroscience (2016) (source for r = 0.122 rumination and 0.048 worry)
  5. Krzikalla C. et al.: Worry Postponement From the Metacognitive Perspective, Clinical Psychology in Europe (2024) (source for the trial and 40% recovery)
  6. Costello A.M. et al.: Rumination and Cardiovascular Adaptation to Repeated Psychological Stress, Stress and Health (2025)
  7. Wendsche J., Weigelt O., Syrek C.J.: Unfinished work tasks and work-related thoughts during off-job time, Anxiety, Stress and Coping (2026)
  8. Wikipedia: Silver Bridge (source for the 1967 collapse, 46 deaths and the 2.5 mm crack)
  9. Wikipedia: Faraday's laws of electrolysis (the law used in the derivation; iron constants from standard tables)
  10. Newslaundry: 4 years, 170 collapses, 202 deaths: What's ailing India's bridges? (2025)
  11. Ministry of Health and Family Welfare: Tele MANAS

More on the science of the mind on the Manthan world page.

worryruminationrustcorrosionperseverative cognitionISO 9223time of wetnessworry postponement
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of Mindblog.

Keep reading in Mindblog

From the other blogs