Doc #46 — Lighting

---

RECOMMENDED ACTIONS (BY ACTUAL URGENCY)

No urgency — happens automatically (days to weeks):

1. Electric lighting continues working in all grid-connected areas. No government intervention needed for general lighting.

First months:

2. Include all lighting stock — LED bulbs, fluorescent tubes, CFLs, incandescent bulbs, batteries, torches, headlamps — in the national consumable inventory (Doc #1, Doc #8). Establish actual stock levels.
3. Issue public guidance on lamp conservation: avoid leaving lights on unnecessarily, use task lighting rather than room lighting where possible, protect LED bulbs from voltage spikes (use surge protectors where available).
4. Requisition wholesale and distributor stocks of LED bulbs, fluorescent tubes, and electronic components (ballasts, drivers) into the national reserve. Allocate to essential facilities first.
5. Inventory all battery-powered portable lighting (torches, headlamps, lanterns, camping lights) and battery stocks. Allocate to off-grid communities, emergency services, and essential outdoor work.

First year:

6. Ensure off-grid communities have small-scale power to keep existing LED lighting functional: micro-hydro where streams are available (Doc #72), retained solar panels and charge controllers, vehicle batteries rotated to grid-connected charging points, or small wind generators. The goal is that no community relies on candles as a primary light source.
7. Begin training electricians in lamp repair and driver replacement — extending the life of existing LED and fluorescent fixtures.
8. Begin experimental program on carbon-filament incandescent bulb production using NZ glass and carbonised plant fibre.
9. Assess NZ tungsten sources: scrap tungsten in existing tool tips, electrical contacts, welding electrodes, and lamp filaments; geological survey of any known tungsten-bearing mineralisation.

Phase 2–3 (Years 1–7):

10. Develop glass envelope production capability for lighting at NZ glass workshops (coordination with Doc #98).
11. Pilot carbon-filament incandescent bulbs — target: a functional, reproducible bulb design using NZ materials.
12. Develop vacuum pump capability (mechanical rotary pumps can be fabricated from NZ-produced metal; see Doc #91) for bulb evacuation.
13. If trans-Tasman trade develops, prioritise import of tungsten wire, LED components, and fluorescent phosphors as high-value, low-volume trade goods.

Phase 4+ (Years 7–15):

14. Scale up incandescent bulb production — whether carbon-filament or tungsten-filament depending on materials availability.
15. Explore arc lamp technology for high-intensity applications (street lighting, industrial, surgical).

---

ECONOMIC JUSTIFICATION

The value of electric lighting

Lighting is essential infrastructure. The loss of adequate lighting compresses the productive day to daylight hours, reduces the quality and safety of indoor work, increases accident rates, and imposes a severe psychological burden. Pre-electric societies went to bed at or shortly after dark for a reason — there was little productive activity possible by candlelight.⁵

Quantifying the gap

The difference between electric and pre-electric lighting is not marginal — it is roughly two orders of magnitude:

Light source	Approximate lumens	Approximate efficacy (lumens/watt)
Single tallow candle	12–13	N/A (chemical energy)
Oil lamp (flat wick, chimney)	60–100	N/A
Carbon-filament incandescent (if produced)	100–200	2–4
Tungsten incandescent (60W)	700–900	12–15
Compact fluorescent (15W)	750–900	50–60
LED bulb (10W)	750–900	75–100

A household that loses electric lighting and substitutes tallow candles loses approximately 98% of its lighting output.⁶ Even a crude carbon-filament incandescent bulb, if NZ can produce them, provides 10–15 times the light of a candle while consuming grid electricity rather than scarce tallow.

Person-years for domestic bulb production

Establishing a small-scale incandescent bulb workshop — glass envelope production, filament preparation, vacuum system, assembly — requires an estimated 5–15 person-years of development effort over Phase 2–3, depending on whether carbon or tungsten filaments are used.⁷ This is a modest investment relative to the value delivered. A single workshop producing even 10,000–50,000 bulbs per year could supply essential facilities (hospitals, schools, workshops, community halls) and gradually extend to household use.

The cost of failure — not producing bulbs before the LED stock is exhausted — is severe. NZ would transition to candle and oil lamp lighting, increasing tallow demand by an estimated 30,000–40,000 tonnes per year (roughly 30–80% of estimated tallow production under nuclear winter conditions), competing directly with soap, lubricant, biodiesel, and food uses (Doc #34, Doc #37, Doc #57), while simultaneously introducing open-flame fire hazards into every household.⁸ Even crude incandescent bulbs running on NZ’s renewable grid represent a dramatically more efficient use of national resources than burning tallow for light.

Breakeven

If the bulb workshop avoids even half of the tallow that would otherwise be burned for lighting — roughly 15,000–20,000 tonnes per year — and that tallow is redirected to soap and lubricant production, the economic value vastly exceeds the workshop’s labour cost within the first year of production. Grid electricity for lighting is effectively free at the margin (NZ has surplus renewable generation capacity under reduced industrial demand); tallow is a constrained national resource with multiple competing uses.

---

1. NZ’S CURRENT LIGHTING STOCK

1.1 Installed lighting

NZ has approximately 1.9 million occupied dwellings and a substantial stock of commercial, industrial, institutional, and public buildings.⁹ The average NZ dwelling contains an estimated 15–30 individual light fittings (ceiling, wall, under-cabinet, outdoor), with commercial and industrial buildings adding significantly to the total.

Estimated total installed lamps (all types): 40–80 million individual lamp units nationally. This is a rough estimate based on dwelling counts and average fitting density; the actual number would be established through the national asset census (Doc #8).¹⁰

Technology mix (approximate, as of mid-2020s):

• LED bulbs and integrated LED fittings: 40–60% of installed stock (NZ has transitioned rapidly to LED since approximately 2015)¹¹
• Fluorescent tubes (linear T8 and T5): 20–30% (dominant in commercial, institutional, and some residential)
• Compact fluorescent lamps (CFLs): 10–15% (declining as LEDs replace them)
• Halogen and incandescent: 5–10% (largely phased out by NZ regulation and market shift)¹²

1.2 Warehouse and retail stocks

At any given time, NZ distributor warehouses and retail outlets hold stocks representing several months of normal replacement demand. NZ’s normal lamp replacement rate is estimated at 10–15 million units per year across all types.¹³ If in-country commercial stocks represent 2–4 months of normal supply, this suggests roughly 2–5 million replacement lamps in the distribution chain.

These stocks should be requisitioned under the national consumable inventory (Doc #1) and allocated by priority.

1.3 Battery-powered portable lighting

NZ households collectively hold a large stock of battery-powered torches, headlamps, and lanterns. No published data exists on total portable lighting stocks, but a reasonable assumption is 1–3 devices per household, suggesting 2–6 million portable lighting devices nationally.¹⁴ LED torches and headlamps are extremely efficient — a modern LED headlamp drawing 1–3 watts provides adequate task lighting for 20–100 hours on a set of AA batteries. Battery stocks (Doc #35) determine how long these devices remain useful.

Rechargeable devices (particularly USB-rechargeable LED lanterns and headlamps) are especially valuable — they can be charged from grid power, solar panels, or vehicle batteries indefinitely.

1.4 Vehicle lighting

NZ’s approximately 4.4 million registered vehicles contain headlights, tail lights, interior lights, and dashboard lighting.¹⁵ As vehicles are mothballed (Doc #6, Doc #33), their lighting components — particularly LED headlight units and halogen bulbs — become a potential parts source for fixed lighting applications. An H4 halogen headlight bulb running on a 12V battery or transformer produces roughly 1,000–1,500 lumens — useful room lighting. Vehicle headlight units could be repurposed for workshop, surgical, or agricultural lighting.

---

2. DEPLETION TIMELINE

2.1 LED bulbs

LED bulbs are the best news in NZ’s lighting picture. They are solid-state devices with no filament to burn out, no glass envelope under vacuum stress, and no gas fill to leak. Rated lifespans of 15,000–50,000 hours are well-supported by testing data, though real-world performance varies with operating temperature, voltage stability, and duty cycle.¹⁶

At 6 hours per day average use:

• 15,000-hour rated LED: approximately 5–9 years (varying with actual daily use of 4–8 hours)
• 25,000-hour rated LED: approximately 8–17 years
• 50,000-hour rated LED: approximately 17–34 years

Under recovery conditions, average daily use will likely be lower than pre-event (more outdoor work, less evening screen time, conservation-conscious behaviour), extending these lifespans further.

The weak point: LED drivers. The LED chip itself is extremely durable. The electronics that convert mains AC power to the low-voltage DC the LED requires — the driver circuit — is the typical failure point. Driver circuits contain electrolytic capacitors that degrade with heat and age (typical rated life: 5,000–20,000 hours at rated temperature), along with semiconductor switching components.¹⁷ A well-designed LED bulb’s driver may last as long as the LED itself; a poorly designed one may fail in 2–5 years.

Implication: NZ should expect a gradual increase in LED bulb failures beginning roughly 5–10 years post-event, accelerating over years 10–20. Some bulbs will last 20+ years; others will fail within 5. The driver failure mode is typically that the bulb goes dark or flickers — it does not usually produce a hazard. In some cases, a failed driver can be replaced if the LED module itself is intact and replacement drivers or simple DC power supplies are available.

2.2 Fluorescent tubes

Linear fluorescent tubes (T8 and T5) have rated lives of 10,000–20,000 hours — roughly 5–10 years at 6 hours per day.¹⁸ However, they also depend on ballasts (magnetic or electronic) to regulate the electrical discharge. Electronic ballasts have similar failure modes to LED drivers (capacitor ageing). Magnetic ballasts are more robust — simple wound copper-and-iron transformers with no electronic components — and can last decades.

The phosphor problem: Fluorescent tubes contain a phosphor coating on the inside of the glass tube that converts ultraviolet light from the mercury discharge into visible light. Over time, the phosphor degrades and the tube dims. Eventually the tube fails entirely (usually when the electrode coatings at each end are depleted). The phosphor compounds are specialty chemicals (typically rare earth phosphors for modern tubes) that NZ cannot produce.¹⁹

Mercury: Each fluorescent tube contains a small amount of mercury (~3–5 mg for modern tubes, more for older ones).²⁰ Mercury is essential for the fluorescent discharge and NZ has no mercury production. Mercury can potentially be recovered from failed tubes by careful crushing and distillation, but this is a small quantity relative to the total mercury inventory in NZ (thermometers, dental amalgam, industrial instruments — Doc #8 would establish the total).

2.3 Halogen and incandescent bulbs

The small remaining stock of halogen and incandescent bulbs has the shortest lifespan — typically 1,000–3,000 hours (halogen) or 750–1,000 hours (incandescent), meaning 6 months to 2 years at 6 hours per day.²¹ These will be the first to deplete. Their value lies not in longevity but in simplicity — they require no electronic driver, run directly on mains voltage (via a step-down transformer for low-voltage halogen), and can potentially be reverse-engineered for domestic production more readily than LED or fluorescent technology.

2.4 Summary depletion projection

Lighting type	Estimated stock	Typical lifespan (at 6 hrs/day)	Approximate depletion window
LED bulbs	20–40 million	7–25 years	Years 7–25 (gradual, driver-dependent)
Fluorescent tubes	10–20 million	5–10 years	Years 5–15 (ballast and phosphor dependent)
CFLs	5–10 million	4–8 years	Years 4–12
Halogen/incandescent	3–8 million	0.5–2 years	Years 1–3
Battery-powered LED	2–6 million devices	Battery-limited	Years 1–10 (battery dependent)

The key insight: NZ has approximately a decade before electric lighting becomes scarce for most grid-connected households. This is enough time to develop domestic incandescent production capability — if the work begins early (Phase 1–2 experimental program, Phase 2–3 pilot production). If development is delayed until the LED stock is visibly failing, the lead time for production capability (3–7 years) means a gap during which NZ falls back on candles and oil lamps — a significant and avoidable failure.

---

3. EXTENDING THE EXISTING STOCK

3.1 Conservation measures

Simple conservation extends the lighting stock substantially:

• Reduce daily operating hours. Organise schedules around daylight. Use electric lighting only when natural light is insufficient. Under nuclear winter with reduced daylight hours, this is harder — but even a reduction from 6 to 4 hours per day extends LED lifespans by 50%.
• Task lighting over room lighting. One LED desk lamp where someone is working uses one bulb; lighting an entire room for one person’s benefit wastes several. Encourage portable and adjustable task lighting.
• Temperature management. LED driver life is strongly affected by operating temperature. Ensure adequate ventilation around enclosed fittings. Remove recessed LED downlights from poorly ventilated ceiling cavities if possible, or replace with pendant or surface-mount fittings that dissipate heat better.²²
• Voltage protection. Voltage spikes from grid instability accelerate electronic component failure. Simple surge suppressors (MOVs — metal oxide varistors) at switchboards protect connected lighting. NZ has existing stocks of surge protectors; MOVs can potentially be fabricated from zinc oxide (Doc #41 discusses zinc oxide sources).²³

3.2 Repair and component reuse

When an LED bulb fails, the failure is usually in the driver, not the LED. An electrician with basic electronics skills can:

• Open the bulb housing
• Test the LED module with a DC power supply to confirm it still functions
• Replace the failed driver with a salvaged working driver from another failed bulb, or with a simple external DC power supply
• Reconnect and reseal

This is not consumer-level repair — it requires tools, test equipment, and knowledge — but it is well within the capability of NZ’s electricians and electronics technicians. A training program in LED bulb repair and component salvage should be part of the accelerated trade training program (Doc #157).

Similarly, fluorescent fixtures with failed electronic ballasts can be rewired to use surviving magnetic ballasts, extending fixture life substantially.

3.3 Vehicle lighting repurposing

As the vehicle fleet is mothballed, lighting components can be harvested:

• Halogen headlight bulbs (H1, H4, H7, etc.): Run on 12V DC. A 230V-to-12V transformer (available from existing stocks or fabricable from laminated iron cores and copper wire — a standard electrical workshop task, though winding a transformer to handle 55–60W continuous load requires care to avoid overheating) powers them from the grid. A single H4 bulb produces 1,000–1,500 lumens — excellent room lighting, though at high power consumption (~55–60W) and relatively short life (~500–1,000 hours).²⁴
• LED headlight modules: More efficient but typically require specific DC voltage (often 9–16V) and current regulation.
• 12V interior and marker lights: Lower output but long-lived.
• Wiring harness and connectors: Useful for lighting circuit fabrication.

Vehicle lighting repurposing is a pragmatic short-term measure — it converts an otherwise idle asset into functional lighting.

---

4. CONTINGENCY: PRE-ELECTRIC LIGHTING

This section covers tallow candles and oil lamps as a fallback if domestic incandescent production does not come online before the LED stock is exhausted. This is not the plan — it is what happens if the plan fails. A widespread reversion to pre-electric lighting would represent one of the more significant quality-of-life failures in NZ’s recovery.

4.1 Why this outcome should be avoided

Tallow candles and oil lamps are drastically inferior to electric lighting. A single tallow candle produces 12–13 lumens. A modern LED bulb produces 750–900 lumens. Three or four candles burning simultaneously in a room provide enough light to move safely, eat a meal, and perform gross motor tasks — but not enough for comfortable reading, detailed craft work, or precision tasks requiring visual acuity.²⁵ Oil lamps with glass chimneys are better (60–100 lumens) but still represent an order-of-magnitude regression from electric light.

The consequences of widespread candle and oil lamp use include:

• Fire hazard: Open flames in every household — a major increase in structural fire risk at a time when fire services are constrained (Doc #166, Section 8.3)
• Tallow competition: At scale, lighting demand for tallow (potentially 30,000–40,000 tonnes per year if a significant fraction of households must rely on candles)²⁶ competes severely with soap (Doc #37), lubricant (Doc #34), biodiesel (Doc #57), and food uses
• Shortened productive day: Retirement to bed earlier, reduced capacity for evening education, reading, and skilled work
• Indoor air quality: Smoke and soot from tallow combustion

This is why domestic incandescent production (Section 5–6) — even crude, dim, short-lived carbon-filament bulbs — is worth the development effort. Even a poor electric bulb eliminates the fire hazard, eliminates tallow consumption, and provides instant on/off control. Grid electricity is effectively free at the margin; tallow is a constrained national resource.

4.2 When candles and oil lamps are appropriate

Pre-electric lighting has a limited appropriate role:

• Emergency backup during localised grid outages (candle stocks in every household and institution)
• Portable field use where battery-powered lighting is unavailable and the task does not justify carrying a generator
• Transitional use in any community where small-scale power has not yet been provisioned

The number of households that should rely on candles as their primary light source should be as close to zero as possible. If off-grid communities receive small-scale power (Section 1 of the executive summary), and if domestic incandescent production develops on schedule (Section 5–6), candle lighting is reduced to a niche emergency role — not a way of life.

4.3 The chimney lamp as best pre-electric option

If pre-electric lighting is needed, the chimney oil lamp — a wick-fed reservoir lamp with a glass chimney — produces the best light output (60–100 lumens from a flat wick design, up to 150+ lumens from a circular-wick Argand-type design).²⁷ Glass chimneys can be produced domestically once NZ glass production capability is established (Doc #98). Metal lamp bodies are fabricable in any sheet metal workshop (Doc #91). The fuel is tallow (liquid above ~45°C) or canola oil where available.

For the limited situations where pre-electric lighting is needed, chimney lamps are a significant improvement over open candles and should be the standard.

---

5. INCANDESCENT BULB PRODUCTION: DEPENDENCY CHAIN

The incandescent light bulb is the simplest form of electric lighting to manufacture — a resistive filament heated to incandescence inside a glass envelope. Edison’s first commercially practical bulb (1879) used a carbonised bamboo filament in a hand-blown glass envelope evacuated with a hand-operated pump.²⁸ The basic principle is within NZ’s reach. The difficulty lies in the details.

5.1 Glass envelopes

Requirement: A thin-walled glass bulb, approximately 50–80 mm diameter, sealed to a metal base. The glass must withstand the thermal cycling of operation (ambient to approximately 200–300°C at the envelope surface for a standard incandescent bulb) without cracking.

NZ capability: Doc #98 describes NZ’s glass production capability. Soda-lime glass bulb envelopes are well within the capability of trained glassblowers using small pot furnaces. The bulb shape is one of the simpler forms in glassblowing — a gather of molten glass blown into a spherical or pear-shaped bubble. Historically, light bulb envelopes were hand-blown before machine production began in the early 1900s.²⁹

Dependency: Requires operating glass furnace (Doc #98), soda ash or potash flux (Doc #98, Section 3), trained glassblower, and annealing capability. All achievable in Phase 2–3.

Production rate: Historical lamp glass workers produced several hundred bulb envelopes per day under factory conditions with practiced technique and dedicated equipment.³⁰ A NZ workshop with newly trained glassblowers, unfamiliar with the specific form, would likely achieve lower rates initially — perhaps 30–80 envelopes per day per worker — increasing with practice. This is modest but sufficient for initial production runs. Machine-assisted production (using a simple mould and compressed air) could increase rates significantly, though building the machinery requires additional engineering effort.

5.2 Filament material

This is the hardest problem.

Tungsten filaments (modern standard):

Tungsten wire is the standard incandescent filament material because of its extremely high melting point (3,422°C — the highest of any metal), reasonable ductility when properly processed, and acceptable evaporation rate at operating temperatures (~2,500–2,700°C for a standard bulb).³¹

Producing tungsten filament wire requires:

1. Tungsten ore or concentrate. NZ has no known economically significant tungsten deposits. The nearest known deposits are in Australia (King Island, Tasmania; Kara mine, Tasmania) and various locations in China, which produces roughly 80–85% of the world’s tungsten.³² Without trade, NZ has no tungsten source beyond scrap recovery.
2. Chemical processing. Tungsten ore (wolframite or scheelite) must be chemically processed to ammonium paratungstate, then reduced to tungsten metal powder — a multi-step chemical process requiring acids, alkalis, and high-temperature reduction in hydrogen atmosphere.³³
3. Powder metallurgy. Tungsten cannot be melted in conventional furnaces (its melting point exceeds the capability of any refractory material NZ can produce). Instead, tungsten powder is pressed into bars and sintered at approximately 2,500–3,000°C by direct resistance heating — passing electrical current through the pressed bar in a hydrogen atmosphere. This requires specialised equipment and extreme temperatures.³⁴
4. Wire drawing. The sintered tungsten bar is swaged (hammered into a smaller cross-section) while hot, then drawn through progressively finer diamond or tungsten carbide dies to produce wire as thin as 20–50 micrometres diameter for lamp filaments. This is a precision metallurgical process requiring specialised dies, lubricants, controlled heating, and experience.³⁵

Honest assessment: Tungsten filament wire production is not achievable in NZ in the near term. It requires a raw material NZ does not have, a chemical processing chain NZ has not built, powder metallurgy equipment NZ does not possess, and wire-drawing capability for refractory metals that does not exist in NZ. This is a Phase 5+ capability at best, and only if tungsten can be imported via trade.

Carbon filaments (original Edison technology):

Edison’s first successful bulbs used carbonised organic material — initially cotton thread, later bamboo fibre, and eventually artificial cellulose (squirted cellulose) filaments.³⁶ Carbon filaments operate at lower temperatures than tungsten (~1,700–2,000°C), producing dimmer, redder light at lower efficiency, and have shorter lifespans (typically 100–1,000 hours depending on the quality of carbonisation and vacuum).

NZ can produce carbon filaments:

1. Source material: Cotton thread (from existing stocks), harakeke muka fibre (Doc #100), bamboo (grows in NZ — several species are established in parks and gardens, particularly in warmer North Island regions), or other cellulose fibres.
2. Carbonisation: The organic fibre is heated in the absence of air (in a sealed container or under charcoal cover) to approximately 600–1,000°C, converting it to nearly pure carbon while retaining the fibre’s shape. This is the same basic process as charcoal making (Doc #102) applied to thin fibres.³⁷
3. Selection and testing: Not all carbonised fibres are suitable — they must be continuous, free of weak points, and of consistent cross-section. Selection and quality control require careful visual inspection and electrical testing.

Performance gap: A carbon-filament bulb produces approximately 1.5–4 lumens per watt — roughly one-third to one-fifth the efficiency of a tungsten filament, and one-twentieth to one-fiftieth that of an LED.³⁸ A 60-watt carbon-filament bulb would produce approximately 100–200 lumens — comparable to a good oil lamp but with the advantage of no smoke, no fire risk, no tallow consumption, and instant on/off. Lifespan would be 100–500 hours under reasonable vacuum conditions — requiring frequent replacement but still providing genuine electric illumination.

Assessment: Carbon-filament incandescent production is achievable with NZ materials and represents a realistic Phase 2–3 development target. The bulbs will be dim, short-lived, and energy-inefficient by modern standards. They are still dramatically better than candles and oil lamps for any community with grid access.

5.3 Vacuum and gas filling

An incandescent filament operating in air burns up almost instantly — the oxygen reacts with the hot carbon or tungsten. The envelope must be either evacuated (vacuum) or filled with an inert gas.

Vacuum:

Early incandescent bulbs were evacuated using hand-operated mercury diffusion pumps. A mechanical rotary vacuum pump — fabricable from NZ-produced metal (cast iron or steel housing, machined rotor, rubber or leather vanes) — can achieve the rough vacuum (~0.01–1 pascal) needed for carbon-filament operation.³⁹ NZ machine shops (Doc #91) can build rotary vacuum pumps; the engineering is well-documented and the precision requirements, while real, are within NZ’s machining capability.

A mercury diffusion pump (using mercury heated in a glass apparatus to create a high-vacuum jet) can achieve the higher vacuum (~0.001 pascal) that improves filament life. Glass diffusion pumps are standard laboratory equipment that NZ scientific glassblowers (Doc #98) can produce. Mercury is available from NZ sources (existing mercury stocks in instruments, thermometers, dental amalgam — see Doc #8).⁴⁰

Inert gas filling:

Modern tungsten-filament bulbs are filled with argon or a nitrogen-argon mixture (typically 93% argon, 7% nitrogen) at near-atmospheric pressure, which suppresses tungsten evaporation and allows higher filament temperatures (brighter light, longer life).⁴¹ For carbon filaments, vacuum is adequate — gas filling is less critical because carbon operates at lower temperatures.

Argon can be separated from air by fractional distillation of liquid air — a process NZ could develop but that requires cryogenic equipment (the liquefaction temperature of air is approximately -195°C). This is a Phase 4+ capability. For initial production, vacuum-filled carbon-filament bulbs are the realistic starting point.⁴²

5.4 Base and seal assembly

The filament must be mounted inside the glass envelope with electrical connections passing through the glass to an external base (typically an Edison screw or bayonet base that fits standard lamp sockets).

Lead-in wires: Thin copper or nickel wires carry current from the base pins through the glass seal to the filament. The wires must be sealed into the glass envelope in a vacuum-tight joint. This is achieved by using metals with thermal expansion coefficients closely matching the glass — traditional lead-in wire for soda-lime glass is “Dumet wire” (a copper-clad nickel-iron alloy) or platinum (expensive but perfectly matched).⁴³ NZ can produce copper and nickel wire; matching the expansion coefficients requires experimentation but is a solvable engineering problem. Historical lamp makers used platinum wire for their first lamps before developing cheaper alternatives.

Base fabrication: Edison screw (E27 is the NZ standard for general lighting) and bayonet (B22) bases are stamped from brass or steel sheet. NZ can produce both metals. Base fabrication requires stamping dies (Doc #91) and a cement or mechanical crimp to attach the base to the glass envelope.

Assessment: Base and seal assembly is achievable with NZ materials and skills. It requires development and testing but involves no materials NZ cannot produce.

---

6. CARBON-FILAMENT BULB: A REALISTIC NZ PRODUCT

6.1 Specification for a Phase 2–3 NZ-produced bulb

Based on the dependency analysis above, the most realistic first-generation NZ-produced electric light bulb would have these characteristics:

Parameter	Specification	Notes
Filament	Carbonised harakeke muka or bamboo fibre	NZ-sourced plant fibre, carbonised in sealed retort
Envelope	Hand-blown soda-lime glass, approximately 60–80 mm diameter	NZ glass from Parengarenga sand (Doc #98)
Fill	Vacuum (mechanical pump + mercury diffusion pump)	NZ-fabricated pumps
Operating voltage	230V AC (NZ mains) or 12V DC (battery systems)	Different filament configurations for each
Power consumption	40–60W (230V) or 15–25W (12V)	Depending on filament resistance
Light output	100–200 lumens (estimated)	Roughly equivalent to a good oil lamp
Colour temperature	1,800–2,200K (warm amber)	Characteristic of carbon incandescence; varies with filament temperature
Expected life	100–500 hours	Highly variable; dependent on vacuum quality and filament consistency
Base	E27 Edison screw (NZ standard)	Stamped from NZ brass or steel

6.2 Production process

1. Filament preparation: Harvest and process harakeke muka (inner fibre) into uniform strands, or prepare bamboo slivers of consistent cross-section. Carbonise in a sealed steel retort at 600–1,000°C for several hours. Select carbonised fibres by visual inspection and electrical resistance testing. Trim to length.
2. Glass envelope blowing: Glassblower produces bulb envelope from soda-lime glass, incorporating a stem (a narrow glass tube through which the envelope will be evacuated). Mount lead-in wires through the stem glass using flame sealing.
3. Filament mounting: Attach carbonised fibre filament to the lead-in wires inside the glass stem assembly. This is delicate work requiring fine motor skill and a clean workspace.
4. Sealing and evacuation: Connect the stem tube to the vacuum pump system. Evacuate the envelope to approximately 0.001–0.01 pascal. While maintaining vacuum, seal the stem by melting the glass tube closed with a torch.
5. Base attachment: Solder or cement the evacuated glass assembly into the metal base, connecting the lead-in wires to the base contacts.
6. Testing: Apply rated voltage and measure light output, current draw, and run time. Reject units that fail within the first 10 hours (indicative of poor vacuum or filament defect).

6.3 Honest performance assessment

These bulbs will be objectively poor by modern standards. An estimated 100–200 lumens from 40–60 watts is roughly 2–4 lumens per watt — less than one-tenth the efficiency of the LED bulbs NZ currently uses, and less than one-third the efficiency of the tungsten incandescent bulbs that LEDs replaced. The warm amber light is pleasant but dim. A 100-hour lifespan means replacement every 2–3 weeks at 6 hours per day of use.

They are still worth producing. A dim electric bulb with no fire risk, no smoke, no tallow consumption, and instant control is a meaningful quality-of-life improvement over candles and oil lamps. And grid electricity in NZ is effectively unlimited — the inefficiency costs watts from NZ’s surplus renewable capacity, not scarce material resources.

---

7. ARC LAMPS: HIGH-INTENSITY ALTERNATIVE (Phase 2–3; Feasibility [B])

Arc lamps — an electric discharge between two carbon electrodes — predate incandescent bulbs historically and produce much brighter light (hundreds to thousands of lumens).⁴⁴ They are unsuitable for domestic lighting (noisy, UV-producing, requiring frequent electrode adjustment) but potentially valuable for:

• Surgical lighting in hospitals where bright, focused light is critical
• Street lighting at key intersections and public areas
• Industrial lighting in workshops and factories
• Searchlights for port and maritime use

NZ feasibility: Carbon electrodes can be produced from carbonised wood or, if available, NZ graphite deposits (NZ has limited known graphite occurrences, primarily in Otago and Westland schist belts, though none have been commercially mined; the skills census and geological survey would establish usable quantities).⁴⁵ Carbonised wood is the more readily available option (similar to electrode production for electric arc furnaces — Doc #106). The arc lamp mechanism is mechanically simple — two carbon rods held in adjustable clamps with a gap between them, connected to a DC power supply. Automatic feed mechanisms that advance the rods as they consume require clockwork or motorised drive but are well within NZ’s mechanical engineering capability.⁴⁶

Assessment: Arc lamps are a Phase 2–3 niche product for high-intensity applications. They do not solve the general lighting problem but they address specific high-value needs that candles and oil lamps cannot serve.

---

8. LONGER-TERM POSSIBILITIES

8.1 Tungsten-filament bulbs via trade (Phase 3–5; Feasibility [C] — dependent on trade)

If trans-Tasman trade develops (Doc #151, Doc #98), tungsten wire is an ideal trade good: extremely high value per kilogram, low volume, and durable in transport. Australia has known tungsten deposits (King Island and Kara in Tasmania).⁴⁷ Even modest quantities of tungsten wire — tens of kilograms — would supply NZ bulb production for years (a single bulb filament weighs approximately 0.05–0.5 grams).⁴⁸

If tungsten wire becomes available: NZ’s carbon-filament bulb workshop can transition to tungsten filaments with relatively modest modification. The glass envelopes, vacuum systems, base fabrication, and assembly processes are the same. Tungsten filaments would immediately improve bulb performance — brighter light (700–900 lumens from 60W), longer life (1,000–2,000 hours), and whiter light.

8.2 Gas mantle technology (Phase 4+; Feasibility [C] — requires thorium extraction and rare earth processing)

Gas mantles — fabric meshes impregnated with thorium and cerium oxides that glow brilliantly when heated by a gas or oil flame — were the dominant lighting technology from the 1890s to the 1930s. A gas mantle lamp produces 300–600 lumens, dramatically brighter than a plain oil lamp.⁴⁹

NZ feasibility: The mantle itself requires thorium oxide (mildly radioactive; NZ has thorium-bearing mineral sands in some beach deposits, particularly monazite)⁵⁰ and cerium oxide (a rare earth — NZ has no rare earth processing). Without cerium, mantles produce less light. Thorium extraction from monazite is achievable chemistry but raises radiation handling concerns. This is a speculative possibility, not a near-term solution.

8.3 Fluorescent tube production (Phase 5+; Feasibility [D] — requires rare earth phosphors and tungsten)

Fluorescent lighting is more efficient than incandescent but requires mercury, phosphor compounds (rare earth oxides for modern tri-phosphor tubes), and precision glass tube drawing. The phosphor compounds are specialty chemicals NZ cannot produce. Mercury is available in small quantities from existing NZ stocks. Electrode fabrication requires tungsten wire (same constraint as incandescent filaments).

Assessment: Domestic fluorescent tube production is not realistic in the near term. NZ’s existing fluorescent stock should be conserved and maintained as long as possible.

8.4 LED production (Phase 7+; Feasibility [D] — requires semiconductor fabrication)

LED manufacturing requires semiconductor fabrication capability — epitaxial growth of gallium nitride on sapphire or silicon carbide substrates, precision doping, photolithographic patterning, wire bonding, and encapsulation. This is the most advanced manufacturing technology in the world. NZ cannot produce LEDs and will not be able to for decades, if ever (Doc #115).⁵¹

NZ’s LED bulb stock is a one-time endowment. Conserve it.

---

9. CRITICAL UNCERTAINTIES / KEY RISKS

Uncertainty	Why it matters	How to resolve
Total NZ lighting stock (all types)	Determines depletion timeline and allocation strategy	National asset census (Doc #8)
LED driver failure rate under NZ grid conditions	Determines whether LEDs last 7 years or 25 years	Monitor installed base; establish tracking program at essential facilities
NZ glass production timeline (Doc #98)	Glass envelopes are a prerequisite for bulb production	Track glass workshop development
Carbon filament quality achievable from NZ plant fibres	Determines whether NZ-produced bulbs are viable	Experimental program — begin Phase 1–2
Vacuum pump achievable quality	Poor vacuum = short filament life	Fabrication and testing of NZ-built pumps
Tungsten availability via trade	Transforms bulb quality from marginal to good	Depends on trans-Tasman and Pacific trade development
Grid stability in the long term (Doc #67, Doc #65)	If the grid degrades, all electric lighting is affected — not bulb-dependent but grid-dependent	Grid maintenance and monitoring
Incandescent production timeline vs. LED depletion	If bulb production is not online before LEDs fail, NZ falls back on candles — a major fire hazard and resource drain	Begin experimental program Phase 1; pilot production Phase 2–3
Off-grid power provisioning	Communities without small-scale power revert to candle lighting unnecessarily	Prioritise micro-hydro, solar, and battery rotation for off-grid settlements

---

10. CROSS-REFERENCES

Document	Relationship
Doc #1 — National Emergency Stockpile Strategy	Lamp stocks, batteries, and components in national inventory
Doc #156 — Skills Census	Lighting stock inventory; electricians and electronics technicians; scientific glassblowers
Doc #33 — Tires	Vehicle fleet mothballing releases lighting components for reuse
Doc #34 — Lubricant Production	Competing demand for tallow
Doc #35 — Battery Management	Battery supply for portable lighting; lead-acid production for 12V lighting systems
Doc #37 — Soap Production	Competing demand for tallow
Doc #166 — Firefighting	Open flame fire risks from candles and oil lamps
Doc #57 — Biodiesel	Competing demand for tallow
Doc #6 — Vehicle Fleet Management	Mothballed vehicles provide batteries for off-grid lighting power
Doc #67 — Transpower Grid Operations	Grid availability is the fundamental prerequisite for electric lighting
Doc #69 — Transformer Maintenance	Grid equipment longevity determines long-term electric lighting availability
Doc #72 — Micro-Hydro	Off-grid power source for communities beyond distribution line reach
Doc #73 — Solar Panel Maintenance	Retained solar panels provide off-grid power for LED lighting
Doc #74 — Pastoral Farming	Livestock numbers determine tallow production volume
Doc #89 — NZ Steel Glenbrook	Steel and copper for lamp components; potential tungsten scrap recovery
Doc #91 — Machine Shop Operations	Vacuum pump fabrication; base stamping dies; lamp production tooling
Doc #98 — Glass Production	Glass envelopes for bulb production; lamp chimneys for oil lamps
Doc #100 — Harakeke Fiber Processing	Filament material (muka fibre for carbonisation)
Doc #102 — Charcoal Production	Carbonisation process applicable to filament preparation
Doc #106 — Electric Arc Furnace	Carbon electrode production relevant to arc lamps
Doc #106 — Computing Self-Sufficiency	Semiconductor manufacturing — context for why LED production is not feasible
Doc #134 — Trans-Tasman and Pacific Trade	Tungsten wire as priority import item
Doc #157 — NZ–Australia Relations	Australian tungsten deposits as potential supply source
Doc #157 — Accelerated Trade Training	Electrician training in lamp repair and component salvage

---

FOOTNOTES

---

1. NZ lighting stock estimate: Based on approximately 1.9 million occupied dwellings (Stats NZ, https://www.stats.govt.nz/) with an estimated 15–30 light fittings per dwelling, plus commercial, industrial, and institutional buildings. The range of 40–80 million individual lamp units is rough and should be verified through the national asset census. No published data exists on NZ’s total installed lamp count.↩︎

2. LED bulb rated lifespan: LED manufacturers typically rate bulbs at L70 — the point at which light output has declined to 70% of initial output. Rated lifespans of 15,000–50,000 hours (at L70) are standard for quality LED bulbs. Actual performance varies with operating temperature, driver quality, and electrical conditions. See: US Department of Energy, “LED Lighting Facts,” https://www.energy.gov/; IES (Illuminating Engineering Society) LM-80 and TM-21 testing standards.↩︎

3. LED driver failure: Electrolytic capacitors in LED drivers have rated lifespans typically in the range of 5,000–20,000 hours at their rated temperature (usually 85°C or 105°C). Actual life depends on ambient temperature — every 10°C reduction in operating temperature roughly doubles capacitor life (Arrhenius relationship). In NZ’s temperate climate, capacitor life generally exceeds the rated minimum. See: Zheludev, N., “The life and times of the LED,” Nature Photonics, 2007; various LED driver reliability studies from IEEE publications.↩︎

4. NZ grid-connected population: The 85–95% estimate is based on NZ’s electricity connection rate. Stats NZ and the Electricity Authority report approximately 2.1 million electricity connections (residential and commercial) against approximately 2 million households. Off-grid households are concentrated in remote rural areas, lifestyle blocks, and some conservation-area dwellings. No precise figure for the off-grid population share is publicly available; the range reflects this uncertainty. See: Electricity Authority, https://www.ea.govt.nz/; Stats NZ dwelling statistics.↩︎

5. Historical impact of lighting on productivity: The economic history of lighting is comprehensively analysed in Nordhaus, W.D., “Do Real-Output and Real-Wage Measures Capture Reality? The History of Lighting Suggests Not,” in Bresnahan, T.F. and Gordon, R.J. (eds.), “The Economics of New Goods,” University of Chicago Press, 1997. Nordhaus documents a roughly 100-fold improvement in lighting efficiency from tallow candles to incandescent bulbs, and another 10-fold from incandescent to fluorescent/LED.↩︎

6. Candle vs. electric lighting output: A standard tallow candle produces approximately 12–13 lumens (one candela, by historical definition). A modern LED bulb producing 800 lumens is approximately 60 times brighter. The “98% loss” figure reflects going from multiple electric light sources in a room (2,000–4,000 lumens combined) to a few tallow candles (40–50 lumens combined). See: Nordhaus (1997), footnote 4; historical candlepower measurement standards.↩︎

7. Person-years estimate for bulb workshop: Based on analogous estimates for small-scale manufacturing development in the Recovery Library. The range of 5–15 person-years covers: glass envelope production setup (1–3 person-years, overlapping with Doc #98 glass development), filament carbonisation process development (1–3 person-years), vacuum system fabrication and testing (1–2 person-years), base fabrication tooling (0.5–1 person-year), assembly process development and training (1–3 person-years), and ongoing testing and quality improvement (1–3 person-years). This is an estimate; actual effort depends on available skills and how many problems are encountered during development.↩︎

8. Tallow demand for full-population candle lighting: If all ~2 million NZ households required candle lighting at an average of 2–4 candles per evening (approximately 0.5–1.0 kg tallow per household per day), annual demand would be approximately 365,000–730,000 tonnes — clearly unsustainable. At the more realistic estimate of partial population (those losing grid access over time), demand of 30,000–40,000 tonnes per year assumes roughly 200,000–400,000 households using candles regularly. The calculation assumes average consumption of approximately 0.5–1.0 kg tallow per household per day for 2–4 candles per evening.↩︎

9. NZ dwelling count: Stats NZ records approximately 1.9 million occupied private dwellings as of the most recent census. https://www.stats.govt.nz/↩︎

10. NZ lighting stock estimate: Based on approximately 1.9 million occupied dwellings (Stats NZ, https://www.stats.govt.nz/) with an estimated 15–30 light fittings per dwelling, plus commercial, industrial, and institutional buildings. The range of 40–80 million individual lamp units is rough and should be verified through the national asset census. No published data exists on NZ’s total installed lamp count.↩︎

11. NZ LED adoption: NZ has experienced rapid LED adoption since approximately 2015, driven by energy efficiency regulations, declining LED prices, and utility-sponsored distribution programs (e.g., EECA — Energy Efficiency and Conservation Authority programs). By the mid-2020s, LED is the dominant lighting technology in NZ retail sales and a large fraction of installed stock. Exact installed penetration figures are estimates. See: EECA lighting publications, https://www.eeca.govt.nz/↩︎

12. NZ lighting regulation: NZ phased out inefficient incandescent bulbs through Minimum Energy Performance Standards (MEPS) under the Energy Efficiency (Energy Using Products) Regulations. Standard incandescent bulbs (non-halogen, non-reflector) have been effectively removed from sale. Halogen bulbs remain available but are being displaced by LED. See: EECA and MBIE energy efficiency regulations.↩︎

13. NZ lamp replacement rate: Estimated from NZ dwelling count, average fittings per dwelling, and average lamp replacement frequency. At approximately 50 million installed lamps with average life of 3–5 years (blended across types), the replacement rate is roughly 10–15 million units per year. This is a rough estimate.↩︎

14. NZ portable lighting stock: No published data. The estimate of 1–3 devices per household is based on general household survey patterns in comparable countries and the prevalence of camping and outdoor recreation in NZ, where portable lighting (torches, headlamps) is common household equipment.↩︎

15. NZ vehicle fleet: Approximately 4.4 million registered vehicles as of 2023–2024. See NZ Transport Agency (Waka Kotahi), https://www.transport.govt.nz/statistics-and-insights/fle...↩︎

16. LED bulb rated lifespan: LED manufacturers typically rate bulbs at L70 — the point at which light output has declined to 70% of initial output. Rated lifespans of 15,000–50,000 hours (at L70) are standard for quality LED bulbs. Actual performance varies with operating temperature, driver quality, and electrical conditions. See: US Department of Energy, “LED Lighting Facts,” https://www.energy.gov/; IES (Illuminating Engineering Society) LM-80 and TM-21 testing standards.↩︎

17. LED driver failure: Electrolytic capacitors in LED drivers have rated lifespans typically in the range of 5,000–20,000 hours at their rated temperature (usually 85°C or 105°C). Actual life depends on ambient temperature — every 10°C reduction in operating temperature roughly doubles capacitor life (Arrhenius relationship). In NZ’s temperate climate, capacitor life generally exceeds the rated minimum. See: Zheludev, N., “The life and times of the LED,” Nature Photonics, 2007; various LED driver reliability studies from IEEE publications.↩︎

18. Fluorescent tube lifespan: Rated at 10,000–20,000 hours for standard T8 tubes and up to 30,000 hours for high-quality T5 tubes, depending on starting frequency (frequent switching shortens life). See: lamp manufacturer technical data (Philips, Osram, GE); IES Lighting Handbook.↩︎

19. Fluorescent phosphors: Modern tri-phosphor fluorescent tubes use rare earth oxide phosphors — typically europium oxide (red), terbium oxide (green), and barium magnesium aluminate doped with europium (blue). These are specialty chemicals produced from rare earth mineral processing, primarily in China. NZ has no rare earth processing capability. See: Ronda, C.R. (ed.), “Luminescence: From Theory to Applications,” Wiley-VCH, 2007.↩︎

20. Mercury content of fluorescent tubes: Modern fluorescent tubes contain approximately 3–5 mg of mercury each; older tubes may contain more (up to 50 mg). Mercury is essential for the fluorescent discharge mechanism — there is no mercury-free substitute for fluorescent lighting. See: US EPA mercury in lighting publications; NZ Ministry for the Environment waste management guidance.↩︎

21. Halogen and incandescent lifespan: Standard incandescent bulbs were typically rated at 750–1,000 hours. Halogen bulbs are rated at 2,000–3,000 hours (the halogen cycle regenerates the filament, extending life). Long-life incandescent variants (thick filament, lower temperature, dimmer light) could last 2,500+ hours but at reduced light output. See: IES Lighting Handbook; lamp manufacturer technical specifications.↩︎

22. LED thermal management: LED efficiency and lifespan are strongly temperature-dependent. Every 10°C increase in LED junction temperature above its rated operating point reduces lifespan significantly. Enclosed and recessed fittings that trap heat accelerate failure. Open or well-ventilated fittings preserve LED life. See: US DOE LED thermal management technical guidance; Cree, Philips, and other LED manufacturer application notes.↩︎

23. Surge protection for lighting: Metal oxide varistors (MOVs) — typically zinc oxide-based — clamp voltage spikes that can damage electronic components in LED drivers and electronic ballasts. Standard household surge protectors contain MOVs. Zinc oxide is producible from NZ zinc sources (limited domestic deposits; primarily from imported zinc stocks or potentially from zinc-bearing galvanised steel recycling). See: electronics protection engineering references.↩︎

24. Halogen headlight specifications: Standard H4 halogen bulbs operate at 12V, drawing approximately 55W (low beam) or 60W (high beam), producing approximately 1,000 lumens (low beam) to 1,500 lumens (high beam). Rated life approximately 500–1,000 hours. See: automotive lighting specifications; ECE R37 regulations.↩︎

25. Candle vs. electric lighting output: A standard tallow candle produces approximately 12–13 lumens (one candela, by historical definition). A modern LED bulb producing 800 lumens is approximately 60 times brighter. The “98% loss” figure reflects going from multiple electric light sources in a room (2,000–4,000 lumens combined) to a few tallow candles (40–50 lumens combined). See: Nordhaus (1997), footnote 4; historical candlepower measurement standards.↩︎

26. Tallow demand for candle lighting (contingency estimate): If a significant fraction of households must rely on candle lighting — because off-grid power provisioning fails or LED stock depletes before incandescent production comes online — demand could reach 30,000–40,000 tonnes per year for 400,000+ households. With successful off-grid power provisioning and timely incandescent production, candle demand should be negligible (emergency backup stocks only).↩︎

27. Argand lamp output: The Argand lamp, invented by Aimé Argand in 1780, uses a hollow cylindrical wick with air flowing through the centre, dramatically improving combustion efficiency and light output compared to flat-wick lamps. With a glass chimney, Argand-type lamps produce 100–150+ lumens — roughly 10 times a single candle. See: O’Dea, W.T., “The Social History of Lighting,” Routledge, 1958; Schivelbusch, W., “Disenchanted Night: The Industrialization of Light in the Nineteenth Century,” University of California Press, 1988.↩︎

28. Edison’s incandescent lamp: Thomas Edison demonstrated a practical incandescent lamp in October 1879 using a carbonised cotton thread filament in an evacuated hand-blown glass bulb. Subsequent development used carbonised bamboo fibres (harvested from specific Japanese bamboo species) which proved more durable, achieving lifespans of approximately 1,200 hours. The carbon-filament era lasted from 1879 to approximately 1910, when tungsten filaments became commercially dominant. See: Friedel, R. and Israel, P., “Edison’s Electric Light: The Art of Invention,” Johns Hopkins University Press, 2010.↩︎

29. Hand-blown lamp envelopes: Until the Corning ribbon machine was invented in 1926 (automating bulb envelope production at rates of thousands per minute), all lamp envelopes were hand-blown by skilled glass workers. A skilled lamp-glass blower could produce several hundred bulb envelopes per day. See: Friedel and Israel (note 23); Corning Glass history publications.↩︎

30. Hand-blown lamp envelopes: Until the Corning ribbon machine was invented in 1926 (automating bulb envelope production at rates of thousands per minute), all lamp envelopes were hand-blown by skilled glass workers. A skilled lamp-glass blower could produce several hundred bulb envelopes per day. See: Friedel and Israel (note 23); Corning Glass history publications.↩︎

31. Tungsten properties: Tungsten has the highest melting point of any metal (3,422°C), high density (19.25 g/cm³), and excellent high-temperature strength. Its use in lamp filaments dates from approximately 1904 (pressed tungsten, by Sándor Just and Franjo Hanaman) and was commercially established by 1910. See: Lassner, E. and Schubert, W.-D., “Tungsten: Properties, Chemistry, Technology of the Element, Alloys, and Chemical Compounds,” Springer, 1999.↩︎

32. Tungsten deposits: Global tungsten resources are concentrated in China (~60% of world mine production), Vietnam, Russia, Bolivia, and others. Australian deposits include King Island and Kara (Tasmania) — both with historically significant scheelite and wolframite deposits. NZ has no known commercially significant tungsten mineralisation, though minor tungsten occurrences have been reported in gold-bearing quartz veins in the South Island. See: USGS Mineral Commodity Summaries — Tungsten; Geoscience Australia mineral resource assessments.↩︎

33. Tungsten chemical processing: The standard route from ore to metal involves: (1) ore concentration and decomposition with alkali or acid; (2) purification to ammonium paratungstate (APT); (3) calcination to tungsten trioxide (WO₃); (4) hydrogen reduction to tungsten metal powder at 700–1,000°C. See: Lassner and Schubert (note 25).↩︎

34. Tungsten powder metallurgy and wire drawing: Tungsten is consolidated from powder by pressing into bars, then sintering at 2,500–3,000°C by direct electrical resistance heating in hydrogen atmosphere. The sintered bars are swaged while hot into rods, then drawn through diamond dies into fine wire. Typical lamp filament wire diameter is 20–50 μm. The entire process from powder to fine wire involves dozens of processing steps and requires specialised equipment at each stage. See: Lassner and Schubert (note 25); Yih, S.W.H. and Wang, C.T., “Tungsten: Sources, Metallurgy, Properties, and Applications,” Plenum Press, 1979.↩︎

35. Tungsten powder metallurgy and wire drawing: Tungsten is consolidated from powder by pressing into bars, then sintering at 2,500–3,000°C by direct electrical resistance heating in hydrogen atmosphere. The sintered bars are swaged while hot into rods, then drawn through diamond dies into fine wire. Typical lamp filament wire diameter is 20–50 μm. The entire process from powder to fine wire involves dozens of processing steps and requires specialised equipment at each stage. See: Lassner and Schubert (note 25); Yih, S.W.H. and Wang, C.T., “Tungsten: Sources, Metallurgy, Properties, and Applications,” Plenum Press, 1979.↩︎

36. Edison’s incandescent lamp: Thomas Edison demonstrated a practical incandescent lamp in October 1879 using a carbonised cotton thread filament in an evacuated hand-blown glass bulb. Subsequent development used carbonised bamboo fibres (harvested from specific Japanese bamboo species) which proved more durable, achieving lifespans of approximately 1,200 hours. The carbon-filament era lasted from 1879 to approximately 1910, when tungsten filaments became commercially dominant. See: Friedel, R. and Israel, P., “Edison’s Electric Light: The Art of Invention,” Johns Hopkins University Press, 2010.↩︎

37. Carbon filament carbonisation: The process of carbonising organic fibres for lamp filaments was developed empirically by Edison and his team at Menlo Park. The key requirements are: uniform starting material (consistent fibre cross-section), controlled heating in an oxygen-free environment (initially in a nickel or iron mould sealed in charcoal), temperatures of 600–1,000°C, and slow heating/cooling rates to prevent cracking. See: Friedel and Israel (note 23).↩︎

38. Carbon filament efficacy: Carbon-filament incandescent lamps operate at filament temperatures of approximately 1,700–2,000°C, producing light at approximately 1.5–4 lumens per watt depending on operating temperature and filament quality. By comparison, tungsten filaments at approximately 2,500–2,800°C produce 12–15 lumens per watt, and modern LEDs produce 80–150 lumens per watt. See: Nordhaus (note 4); IES Lighting Handbook historical data.↩︎

39. Mechanical vacuum pumps: Rotary vane vacuum pumps — the most common type for rough vacuum — use a rotating vane in a cylindrical chamber to compress and expel gas. They can achieve pressures of approximately 1–10 pascal (rough vacuum, adequate for early carbon-filament bulbs). Construction requires precision machining of the cylinder bore and rotor, with vanes of carbon, fibre, or rubber. The technology dates from the early 20th century and is well within NZ’s machining capability. See: Harris, N.S., “Modern Vacuum Practice,” McGraw-Hill, 3rd ed., 2005.↩︎

40. Mercury diffusion pumps: Glass mercury diffusion pumps (Sprengel pump and descendants) use a stream of falling mercury drops to trap and remove gas molecules, achieving vacuum levels of approximately 0.001 pascal or better. They were the standard high-vacuum technology from the 1860s through the early 20th century and were used for all early incandescent lamp production. Mercury consumption is modest (the same mercury is recycled). Glass diffusion pumps are standard scientific glassblowing products. See: Dushman, S., “Scientific Foundations of Vacuum Technique,” Wiley, 2nd ed., 1962.↩︎

41. Incandescent bulb gas fill: The standard gas fill for tungsten incandescent bulbs is approximately 93% argon and 7% nitrogen at roughly 80–85% of atmospheric pressure. The argon suppresses tungsten evaporation from the filament (extending life), while the nitrogen prevents arcing at the filament supports. Some speciality bulbs use krypton or xenon for improved performance. See: Waymouth, J.F., “Electric Discharge Lamps,” MIT Press, 1971; Coaton, J.R. and Marsden, A.M. (eds.), “Lamps and Lighting,” Arnold, 4th ed., 1997.↩︎

42. Argon from air separation: Argon constitutes approximately 0.93% of Earth’s atmosphere by volume. It is industrially separated by fractional distillation of liquid air — air is compressed, cooled to liquefaction (~-195°C), and then distilled to separate nitrogen, oxygen, and argon. This requires cryogenic equipment (compressors, heat exchangers, distillation columns) that NZ would need to build. BOC/Linde operates an air separation plant in NZ for industrial gas supply — if this plant can be maintained, argon is available. See: Kerry, F.G., “Industrial Gas Handbook: Gas Separation and Purification,” CRC Press, 2007.↩︎

43. Lead-in wire and glass-to-metal seals: The wire passing through the glass envelope must form a vacuum-tight seal. This requires the wire’s thermal expansion coefficient to closely match the glass. For soda-lime glass (expansion ~8.5 × 10⁻⁶/°C), suitable lead-in materials include platinum (expansion ~9.0 × 10⁻⁶/°C — excellent match but expensive) and Dumet wire (a copper-clad nickel-iron alloy, expansion ~9.2 × 10⁻⁶/°C — the standard lamp industry material since approximately 1912). NZ can produce nickel-iron wire, though matching the exact alloy composition for Dumet requires experimentation. See: Pask, J.A. and Fulrath, R.M., “Fundamentals of Glass-to-Metal Bonding,” Journal of the American Ceramic Society, 1962; lamp manufacturing references.↩︎

44. Arc lamp history and output: The carbon arc lamp, first demonstrated by Humphry Davy in the early 1800s and commercially developed from the 1870s, produces light by an electric arc between two carbon electrodes. Light output ranges from a few hundred lumens (small enclosed arc) to tens of thousands of lumens (open arc searchlight). Arc lamps dominated street lighting from the 1880s to the 1930s before being displaced by incandescent and fluorescent lighting. See: Bowers, B., “Lengthening the Day: A History of Lighting Technology,” Oxford University Press, 1998.↩︎

45. NZ graphite deposits: NZ has minor graphite occurrences reported in metamorphic schist formations in Otago (particularly the Otago Schist belt) and Westland. These have not been commercially mined and quantities are uncertain. For arc lamp electrodes, carbonised wood (charcoal processed at high temperature to increase graphitisation) is a more practical NZ-sourced feedstock. See: GNS Science mineral occurrence database; Christie, A.B. and Brathwaite, R.L., “Mineral Commodity Report 15 — Graphite,” NZ Institute of Geological and Nuclear Sciences, 2003.↩︎

46. Arc lamp electrode consumption: Carbon electrodes in arc lamps are consumed during operation — the carbon sublimes and is deposited on the inner surface of the glass enclosure (for enclosed arcs) or dispersed as smoke (for open arcs). Electrode consumption rate depends on current and gap — a typical street lighting arc lamp consumes approximately 25–50 mm of electrode per hour of operation. Automatic feed mechanisms using clockwork or motor-driven lead screws advance the electrodes to maintain the arc gap. See: Bowers (note 35).↩︎

47. Tungsten deposits: Global tungsten resources are concentrated in China (~60% of world mine production), Vietnam, Russia, Bolivia, and others. Australian deposits include King Island and Kara (Tasmania) — both with historically significant scheelite and wolframite deposits. NZ has no known commercially significant tungsten mineralisation, though minor tungsten occurrences have been reported in gold-bearing quartz veins in the South Island. See: USGS Mineral Commodity Summaries — Tungsten; Geoscience Australia mineral resource assessments.↩︎

48. Tungsten filament weight: A standard 60W incandescent bulb filament (coiled-coil design, approximately 50 cm of wire at ~45 μm diameter) weighs approximately 0.05–0.1 grams. Simpler single-coil designs for NZ production might use thicker wire (requiring more tungsten per filament) but still measured in fractions of a gram. One kilogram of tungsten wire would supply filaments for roughly 5,000–20,000 bulbs depending on design. See: Lassner and Schubert (note 25); lamp engineering references.↩︎

49. Gas mantle technology: The incandescent gas mantle (invented by Carl Auer von Welsbach in 1885) uses a mesh of thorium oxide (99%) and cerium oxide (1%) that glows brilliantly when heated by a gas flame. A mantle lamp produces 300–600 lumens — roughly 5–10 times a flat-wick oil lamp and comparable to a moderate incandescent bulb. Gas mantles were the dominant non-electric lighting technology from the 1890s through the 1930s. See: Schivelbusch (note 22); Bowers (note 35).↩︎

50. NZ monazite and thorium: Monazite (a thorium- and rare-earth-bearing phosphate mineral) is found in heavy mineral sand concentrations on some NZ west coast beaches, particularly in Westland. These deposits were investigated for their potential rare earth content in the mid-20th century but were never commercially developed. Thorium content is typically 5–7% of monazite by weight. See: Christie, A.B. and Brathwaite, R.L., “Mineral Commodity Report 17 — Rare Earths and Related Elements,” NZ Institute of Geological and Nuclear Sciences, 1998; Williams, G.J. (ed.), “Economic Geology of New Zealand,” Australasian Institute of Mining and Metallurgy, 1974.↩︎

51. LED manufacturing requirements: LED chips are produced by metal-organic chemical vapour deposition (MOCVD) of gallium nitride (GaN) and related III-V semiconductor compounds on sapphire or silicon carbide substrates, with precision doping, photolithographic patterning at micrometre scale, and encapsulation. This represents the frontier of semiconductor manufacturing — comparable in complexity to microprocessor fabrication. See: Schubert, E.F., “Light-Emitting Diodes,” Cambridge University Press, 3rd ed., 2018.↩︎