Please Select A Language
简体中文
English
BLOG

Solar Panel Performance | Monocrystalline vs. Polycrystalline vs. Thin-film

Monocrystalline silicon has an efficiency of 22%-24% (highest), polycrystalline silicon 18%-20% (excellent cost-effectiveness), and thin-film 10%-15% (excellent flexibility).

Monocrystalline or polycrystalline silicon is suitable for areas with high sunlight, while thin-film is better for low-light conditions or curved surfaces.


Monocrystalline


Monocrystalline silicon modules have a global market share exceeding 95% (PV Tech 2023), with mass-production efficiency of 19%-22% and lab efficiency reaching 26% (NREL data), achieved due to the high electron mobility in the single-crystal silicon lattice.

For the same installed capacity, they save 10%-15% land area compared to polycrystalline silicon, with first-year degradation ≤2% and 25-year linear degradation ≤0.55%. PERC/TOPCon technologies enable bifaciality over 80%.



Technical Foundation


How Silicon Material is Purified:

The starting point for monocrystalline silicon is silicon material. The raw material is quartz sand (silicon dioxide), requiring two-step purification to reach solar-grade (99.9999% purity, i.e., 6N).

The first step uses the carbothermal reduction method: Quartz sand reacts with coke in an electric arc furnace (1800-2000°C) to produce metallurgical-grade silicon (MG-Si, purity 98%-99%), consuming about 13,000 kWh per ton of silicon material (US REC Silicon data).

The second step uses the Siemens process (mainstream): Metallurgical-grade silicon reacts with hydrogen chloride gas to form trichlorosilane (SiHCl₃), purified in distillation columns (removing impurities like boron, phosphorus to <0.1ppb), then deposited on silicon rods at 1100°C under hydrogen, resulting in rod-shaped polysilicon (purity 9N-11N).

In 2023, the Siemens process production line of German company Wacker Chemie consumed about 60 kWh per kg of solar-grade silicon, with stable purity above 99.9999% (total impurities <1ppm).

Growing a Perfect Silicon Ingot:

Purified polysilicon needs to be formed into monocrystalline silicon ingots. The mainstream technology is the Czochralski method (CZ method).

The process occurs in a quartz crucible: Place polysilicon fragments into the crucible, heat to 1414°C (silicon melting point) to melt, introduce a seed crystal (a small piece of single-crystal silicon, diameter 4-6 inches), slowly pull upwards at 0.5-2mm per minute. Meanwhile, the crucible rotates counterclockwise (5-15 rpm) and the seed crystal rotates clockwise (10-20 rpm), allowing the molten silicon to solidify along the crystal orientation of the seed, forming a cylindrical ingot. Key parameters:

l Ingot diameter: 182mm (mainstream), 210mm (large size); Length 1.5-2.5m;

l Pulling time: 50-120 hours (longer for larger diameter and longer length);

l Impurity control: Crucible uses high-purity quartz (SiO₂), with argon gas flow to reduce oxygen impurities (oxygen content <15ppma, otherwise oxygen donors form affecting efficiency);

l Energy consumption: About 80-100 kWh per kg of ingot (US GT Advanced Technologies equipment data).

The Nuance of Slicing Ingots into Thin Wafers:

Silicon ingots need to be sliced into 180-200μm thick wafers (earlier used slurry cutting, now mainstream is diamond wire cutting).

The diamond wire consists of high-strength steel wire (diameter 40-50μm) coated with diamond particles (grit size 3-10μm). During cutting, under the action of coolant (water-based suspension containing silicon carbide abrasive), the wire moves rapidly back and forth (speed 10-15m/s), slicing the ingot into thin wafers. Key data:

l Cutting loss: About 45-50 wafers per kg of ingot (at 180μm thickness), loss rate <20% (slurry cutting loss rate over 30%);

l Surface roughness: After cutting, wafer surface roughness Ra <0.5μm, requiring subsequent polishing (using aluminum oxide powder) to reduce to Ra <0.2μm;

l Breakage rate: Diamond wire cutting breakage rate <0.5% (slurry cutting 1%-2%), but requires careful wire tension control (deviation ±5% can cause wire breakage).

How Solar Cells are "Activated":

Silicon wafers need to be processed into solar cells (individual power generation units), involving five steps:

1. Texturing: Corrode the wafer surface with sodium hydroxide solution (80°C) to form a pyramid texture (height 5-10μm), further reducing reflectivity to <10% (untextured silicon wafer reflectivity about 35%);

2. Diffusion to Form PN Junction: Place wafers into a tube diffusion furnace (850-900°C), introduce phosphorus oxychloride (POCl₃) gas. Phosphorus atoms diffuse into the wafer surface layer (depth 0.3-0.5μm), forming an N-type layer (forming a PN junction with the wafer's P-type substrate). Sheet resistance controlled at 40-60 Ω/□ (too low prone to short circuit, too high causes high resistance);

3. Edge Isolation: Use plasma etching to remove the PN junction at the wafer edges (to prevent cell short circuit), etch width 0.5-1mm;

4. Anti-Reflection Coating: Use PECVD equipment (400-500°C) to deposit a silicon nitride (SiNₓ, thickness 70-90nm) film on the surface, refractive index 2.0-2.1, reducing reflectivity to <5% (visible light range);

5. Screen Printing Electrodes: Use screen printing to apply silver paste (busbars + fingers). Busbar width 1-1.5mm (collects current), finger width 30-40μm (densely distributed to collect electrons).

Module Assembly:

Cells are connected in series to form a module (common 60-cell/72-cell). Assembly process:

l Stringing: Connect cell positive/negative terminals with tabbing wire (copper-based tin-plated), soldering temperature 250-300°C, pull force required >1N/mm;

l Layup: Glass (3.2mm tempered low-iron glass) + EVA encapsulant (ethylene-vinyl acetate copolymer, thickness 0.5mm) + cell strings + EVA + backsheet (TPT structure: PVF + PET + PVF, thickness 0.3mm);

l Lamination: Vacuum laminator (temperature 145°C, pressure -0.1MPa) heats for 15-20 minutes, causing EVA to cross-link and cure (cross-linking degree >75%);

l Framing: Aluminum alloy frame (wall thickness 1.2mm), apply sealant (silicone) for waterproofing, grounding resistance <0.1Ω.


Performance Parameters


Conversion Efficiency:

Among mass-produced modules, P-type monocrystalline silicon (boron-doped) efficiency is 19%-22%, while N-type monocrystalline silicon (phosphorus-doped) achieves 22%-25% efficiency due to longer minority carrier lifetime.

For comparison, polycrystalline silicon mass-production efficiency is 17%-19% (1-3 percentage points gap), amorphous silicon thin-film only 6%-12% (over 10 percentage points gap).

At lab level, the theoretical efficiency limit for monocrystalline silicon is 29.4% (Shockley-Queisser limit).

In 2023, Germany's Fraunhofer ISE broke 32.5% with perovskite-silicon tandem cells, but mass production still relies on monocrystalline silicon base.

The efficiency difference stems from crystal structure: monocrystalline silicon atoms are arranged without grain boundaries, electron mobility 1500 cm²/Vs, resulting in lower energy loss.

Temperature Coefficient:

Monocrystalline silicon temperature coefficient is -0.35% ~ -0.45%/°C (P-type average -0.40%, N-type -0.30%), meaning power drops about 0.4% per 1°C temperature rise.

Compared to thin-film (-0.20% ~ -0.30%/°C), monocrystalline silicon degradation is slightly more noticeable at high temperatures, but actual impact depends on the environment.

For example, in Arizona, USA, module temperature can reach 65°C in summer (ambient 45°C).

Monocrystalline silicon power drops 18%-22% compared to 25°C standard conditions, while thin-film drops 12%-16%.

Tests by Germany's Fraunhofer Institute show that in regions with average annual temperature of 10°C (e.g., Sweden), monocrystalline silicon's temperature coefficient disadvantage disappears, and annual energy yield exceeds thin-film by 15%-20%.

Low-Light Performance:

Low-light performance refers to power generation capability under low irradiance (e.g., 200W/m², equivalent to cloudy/rainy days). Monocrystalline silicon performs moderately well.

Using power under Standard Test Conditions (1,000W/m²) as baseline: monocrystalline silicon outputs 85%-90% at 200W/m², polycrystalline silicon 80%-85%, amorphous silicon thin-film 75%-80%.

Real-world case: In Berlin, Germany, during winter cloudy/rainy days (daily average irradiance 150-200W/m²), monocrystalline silicon modules generate 5%-7% more daily energy than polycrystalline silicon, and 10%-15% more than thin-film.

The reason is the stronger built-in electric field in the PN junction of monocrystalline silicon, resulting in higher electron-hole separation efficiency under low light.

Degradation Rate:

Monocrystalline silicon first-year degradation ≤2% (P-type 1.5%-2%, N-type 1%-1.5%, due to less boron-oxygen complex formation in N-type), and linear degradation ≤0.55%/year over 25 years (P-type 0.5%-0.55%, N-type 0.4%-0.5%).

Compared to polycrystalline silicon: First-year degradation ≤2.5%, linear degradation ≤0.7%/year.

Taking warranty data as an example, US SunPower (Maxeon) monocrystalline silicon modules promise ≥92% of initial power after 25 years (first-year degradation 1%, then 0.25% annually), while German Q Cells modules promise ≥84.8% after 25 years.

Main causes of degradation are surface passivation layer aging and metal electrode diffusion. N-type monocrystalline silicon, lacking boron-oxygen pairs, has LID at least 50% lower than P-type.

Bifaciality:

P-type monocrystalline silicon bifaciality is 75%-80% (standard PERC), N-type TOPCon bifaciality is 80%-85% (tunnel oxide layer reduces rear recombination), HJT (heterojunction) bifaciality is even higher at 90%-95%.

Rear-side power gain depends on ground reflectivity: White TPO roof (70% reflectivity) gains 15%-20%, sandy ground (40% reflectivity) gains 10%-15%, grassland (20% reflectivity) gains 5%-8%.

A SolarPower Europe 2023 report shows that a ground-mounted power plant in Texas, USA, using N-type monocrystalline silicon bifacial modules (85% bifaciality), generated 22% more annual energy than monofacial modules.

Anti-PID Capability:

Monocrystalline silicon anti-PID capability relates to doping type: P-type monocrystalline silicon PID degradation 3%-5% (over 25 years). N-type, due to better surface passivation after phosphorus doping, PID degradation <1% (SunPower test data).

Test method per IEC 62804: Module subjected to -1500V voltage at 85°C, 85% RH for 96 hours. P-type monocrystalline silicon power drops 3%-5%, N-type drops <1%.

In practical application, at a power plant in Florida, USA (high humidity coastal area), N-type monocrystalline silicon showed only 0.8% PID degradation after 5 years of operation, while P-type reached 3.2%.

Mechanical Performance:

Monocrystalline silicon modules tested per IEC 61215: Front wind load resistance 2400Pa (equivalent to 56m/s wind speed, Category 17 typhoon), rear load 5400Pa (1.8m snow load);

Hail impact test uses 25mm diameter ice balls (mass 7.53g) impacting at 23m/s (equivalent to Midwest US hail), no cracks allowed.

US UL certification data shows monocrystalline silicon module displacement <3mm under 60m/s wind speed (exceeding Category 17), aluminum frame tensile strength >160MPa (ordinary steel 120MPa).

Compared to thin-film, monocrystalline silicon is slightly more prone to micro-cracks (0.1%-0.3%) due to brittle wafers, but optimized encapsulation (e.g., double-glass structure) can reduce micro-crack rate below 0.05%.


Scenario Performance


Residential Rooftop:

A household in Sacramento, California, USA, with 220m² roof area, installed 400W monocrystalline silicon modules (efficiency 21.5%, size 1.7m×1.1m), total 54 pieces (21.6kW).

Local annual average sunshine 5.8 kWh/m²/day (NREL data), annual generation 31,000 kWh.

Compared to same area polycrystalline silicon modules (efficiency 19%, requiring 60 pieces), monocrystalline silicon generates 2,800 kWh more annually (about 9%).

This household's annual electricity consumption is 15,000 kWh, surplus sold to the grid at 0.18/kWh (California net metering rate), annual revenue 504.

Installation cost for monocrystalline silicon is 4,320 higher than polycrystalline silicon (0.25/W vs $0.21/W), but the extra revenue from higher generation recoups the premium in 2.8 years.

Visually, monocrystalline silicon's pure black mirror finish is closer to roof tile color, more harmonious than polycrystalline silicon's deep blue granular appearance. Local installers report such orders exceed 70%.

Commercial Building:

A 5-story office building in Berlin, Germany, uses monocrystalline silicon double-glass modules on its south facade (transparency 30%, power 320W/piece, efficiency 22%), total 120 pieces (38.4kW).

Local annual average sunshine 3.2 kWh/m²/day, modules installed vertically (conforming to facade), annual generation 38,000 kWh.

The building's annual air conditioning electricity consumption is 19,000 kWh, module generation offsets 20% of the load, saving €3,800 annually (German commercial electricity rate €0.20/kWh).

Compared to traditional glass curtain wall (cost €600/m²), double-glass module curtain wall cost is €850/m², but power generation revenue covers the price difference in 5 years.

Module surface hardness reaches Mohs scale 6 (tempered glass standard), wind pressure resistance 2400Pa (Berlin max wind speed 22m/s), 5 years of operation with no breakage records.

Off-Grid Farm:

A cattle farm in Queensland, Australia, installed a 10kW monocrystalline silicon off-grid system (paired with Tesla Powerwall 2 storage, 13.5kWh).

Modules use bifacial double-glass (efficiency 22.5%, 210mm large size), ground covered with gravel (reflectivity 35%), rear-side generation gain 18%, total annual generation 16,000 kWh.

Local average annual cloudy/rainy days 30, fully charged storage can supply power for 48 hours (2kW load: water pump, lighting, refrigerator).

Compared to thin-film modules (efficiency 10%, requiring 20kW), the monocrystalline silicon system is 40% smaller in volume (racking length 30% less), 35% lighter (transportation cost lower by $800).

Maintenance over 5 years: only cleaned twice (low dust), failure rate 0.2 times/year.

Vehicle-Mounted Foldable Panel:

US RV enthusiasts use 200W monocrystalline silicon foldable panels (efficiency 23%, weight 1kg, size 1.2m×0.5m), specific power 200W/kg (thin-film only 80W/kg).

Deployed when parked (tilt 30°), daily generation 1.2 kWh (sunny day), powering RV refrigerator (0.5 kWh/day), LED lights (0.1 kWh), phone charging (0.05 kWh), surplus 0.55 kWh stored in cell.

Compared to thin-film panels, monocrystalline silicon generates 25% more in weak morning/evening light (irradiance 300W/m²) and 18% more on cloudy days (200W/m²).

Field test at Arches National Park, Utah (high altitude, strong UV), continuous use for 3 years with no efficiency degradation (annual degradation 0.3%).

Ground-Mounted Power Plant:

A 100MW ground-mounted plant in Abilene, Texas, USA, uses N-type monocrystalline silicon bifacial modules (efficiency 24.5%, bifaciality 85%), ground covered with white TPO reflective film (reflectivity 70%).

Local annual average sunshine 6.5 kWh/m²/day, module tilt 25, total annual generation 185 million kWh (monofacial modules 152 million kWh, bifacial gain 22%).

Compared to polycrystalline silicon monofacial modules, monocrystalline silicon generates 33 million kWh more annually, LCOE lower by $0.015/kWh (calculated over 25 years).

O&M data shows that after dust shading, monocrystalline silicon module power recovery rate is 95% (washing once/month), 5% higher than thin-film (90%), due to better hydrophobicity of surface silicon nitride film (contact angle >90), making dust easier to slide off.

Cold Region:

A ski resort in Kiruna, Sweden, installed 50kW monocrystalline silicon modules (efficiency 22%, temperature coefficient -0.30%/°C), mounted on racks beside ski slopes (tilt 45°).

Local annual average temperature -2°C, winter -20°C, module temperature often below 0°C (about -5°C when ambient is -15°C).

At low temperatures, monocrystalline silicon efficiency actually increases 2%-3% (negative temperature coefficient characteristic), January daily average generation is 15% higher than at 25°C standard conditions.

After snowfall, the module's black surface absorbs heat quickly, melting snow within 2 hours (polycrystalline dark blue needs 4 hours), reducing manual cleaning cost.

Annual total generation 52,000 kWh, supplying power for resort cable cars and lighting, covering 30% of annual electricity consumption.


Polycrystalline


Polycrystalline (Polycrystalline) solar panels are composed of randomly arranged tiny silicon grains, produced by casting (molten silicon poured into a mold and cooled), costing 10%-20% less than monocrystalline silicon.

Mass production efficiency is 15%-17% (NREL 2023), low-light generation is 3%-5% higher than monocrystalline silicon, temperature coefficient -0.40%/°C.

Mainly used in large ground-mounted power plants in the USA, Europe, and industrial/commercial rooftops in Australia, holding a 40% share of global large-scale projects (IEA 2023), competing based on levelized cost of electricity (LCOE) advantage.


Manufacturing Process


Silicon Purification and Charge Preparation

Quartz sand and coke are reduced in an electric arc furnace to produce 98% pure crude silicon, further refined via the Siemens Process:

l Trichlorosilane (TCS) gas decomposes in a 1000°C reactor, depositing high-purity polysilicon rods (purity 99.9999%);

l Energy consumption about 60 kWh per kg of polysilicon (IEA 2023), 15% lower than monocrystalline Czochralski method;

l Before charging, silicon material needs crushing to 5-50mm particles, mixed with boron/phosphorus dopants to control conductivity type (primarily p-type).


Directional Solidification Casting Furnace Operation Details


Thermal Field Design Determines Grain Quality:

l Uses graphite crucible to hold molten silicon, water-cooled copper plate at bottom creates unidirectional temperature gradient (5-10°C/cm);

l Heating zone temperature maintained at 1414°C (silicon melting point), top cooling zone gradually lowered to 1350°C;

l Solidification process takes 20-30 hours, grains grow upward from crucible bottom, size about 0.5-2mm (NREL 2022 microscopy analysis).

Key Control Parameters:

Parameter

Typical Value

Impact

Cooling Rate

0.5-2°C/minute

Rate too fast leads to overly small grains (0.5mm)

Argon Flow Rate

50-100 L/minute

Prevents silicon oxidation, impurity concentration <0.1ppba

Crucible Coating Material

Si₃N₄

Reduces adhesion between molten silicon and crucible

Ingot Formation and Post-Processing

After solidification, a square silicon ingot is formed (typical size 840×840×400mm):

l After removal, cut off the top impurity-rich zone (approx. 20mm thick) and irregular edges;

l Surface grinding to roughness Ra≤1μm, ensuring slicing accuracy;

l Infrared flaw detection scans for internal cracks (defect rate <0.5%).


Wire Sawing Process and Kerf Loss Control


Multi-wire Sawing Machine Workflow:

1. Nickel-coated steel wire (diameter 180μm) moves reciprocally at 10 m/s;

2. Silicon carbide abrasive slurry (grit #800) injected into cutting zone;

3. Single cut produces 250μm thick silicon wafers (monocrystalline typically 180μm).

Material Loss Comparison:

l Ingot-to-wafer yield 92% (monocrystalline 88%), as square ingot avoids circular edge waste;

l Silicon wafer area per kg of ingot approx. 48 m² (monocrystalline only 42 m²) (PV Magazine 2023 test).

Grain Boundary Formation Mechanism and Electrical Impact

Microscopy reveals random grain boundary network (Figure 1):

l Grain boundary width approx. 0.1-1μm, contains high dislocation and impurity concentrations (oxygen/carbon concentration up to 1e16 atoms/cm³);

l Electron recombination occurs when crossing grain boundaries, carrier lifetime 30% lower than monocrystalline silicon (NREL carrier testing);

l Passivation treatment (silicon nitride coating) can reduce surface recombination velocity below 100 cm/s.

Overseas Production Line Example: US SunPower Casting Facility

l Uses GT Solar DSS™450 casting furnaces, single furnace capacity 450kg;

l Thermal field zone temperature control (±1°C precision), grain size standard deviation <0.2mm;

l Wafer yield 97.3% (2023 Q2 data), kerf loss 0.08mm/wafer.

Technology Evolution: Continuous Casting and Magnetic Field Application

New Process Breakthroughs:

l Norway REC continuous casting technology: Molten silicon pulled continuously through narrow mold, capacity increased 50%;

l Japan Panasonic electromagnetic stirring casting: Applies 0.1T transverse magnetic field to suppress convection, oxygen content reduced from 15ppma to 8ppma;

l Germany Fraunhofer ISE laser etching grain boundaries: Selectively ablates high-defect regions, efficiency increased 0.8% (lab data).


Process Defect Tracing and Improvement


Common Defect Types:

l Micro-cracks (35% of rejected wafers): Caused by uneven cooling thermal stress, solution is optimizing insulation layer design;

l Hidden cracks (25%): Caused by excessive wire sawing tension, switching to ceramic guide rollers reduces occurrence;

l Surface scratches (20%): Caused by uneven abrasive slurry particle size, requires increasing filtration frequency.


Cost Advantage


How are manufacturing costs saved?

Polycrystalline silicon uses the casting process, saving costs at each stage compared to monocrystalline silicon's Czochralski (CZ) method.

l Purification energy consumption 15% lower: Monocrystalline CZ method requires pulling silicon into a single-crystal ingot, consuming 75 kWh per kg (NREL 2023); polycrystalline silicon uses Siemens process refinement followed by direct casting, consuming 60 kWh per kg, saving 15,000 kWh per MW of modules (enough for an average household for 13 years).

l Equipment investment 30% less: CZ method requires high-precision crystal pullers (each 3 million), casting method uses ordinary ingot casting furnaces (each 2 million), a 1GW production line equipment cost difference is $100 million (PV Tech 2023).

l Labor cost 20% lower: Casting process has higher automation, requiring 15 fewer workers per line compared to CZ method.

What's the end-user price, how much cheaper than monocrystalline silicon?

2023 US market data:

l Polycrystalline silicon module average price 0.28/W, monocrystalline silicon PERC module 0.32/W, 0.04 cheaper per watt (equivalent to 40 savings per kW).

l For a 10kW residential system, polycrystalline silicon saves $400 compared to monocrystalline silicon (excluding installation).

l Price gap is narrowing: In 2018, polycrystalline silicon was 30% cheaper than monocrystalline (0.22 vs 0.31/W), shrank to 12.5% in 2023 (BloombergNEF).


Uses less silicon material, can also be recycled


l Silicon material utilization 8% higher: Monocrystalline silicon uses round wafers, kerf loss during cutting accounts for 12%; polycrystalline silicon uses square wafers, kerf loss only 4% (IEA 2022). 1GW polycrystalline silicon modules use 2800 tons of silicon material, monocrystalline silicon requires 3050 tons, saving 250 tons per GW (enough for 5,000 modules).

l Kerf scrap recycling rate 90%: Silicon dust from cutting melted and recast into ingots, each kg of recycled material yields 0.8 kg of new wafers (German Wacker chemical recycling data).

How cost-effective is the LCOE actually?

Levelized Cost of Electricity (LCOE) is the biggest selling point for polycrystalline silicon. Measured data from overseas power plant projects:

Project Type

Polycrystalline LCOE

Monocrystalline LCOE

Difference

Case Source

US California 500MW ground-mount

$0.028/kWh

$0.031/kWh

Saves $0.003/kWh

Berkeley Lab 2023

Australia Queensland farm

AUD 0.029/kWh

AUD 0.033/kWh

Saves AUD 0.004/kWh

Australia Clean Energy Regulator 2022

India Rajasthan 200MW

2.1 INR/kWh

2.4 INR/kWh

Saves 0.3 INR/kWh

India Ministry of New & Renewable Energy 2023 assessment

Overall calculation: California 500MW plant using polycrystalline silicon earns $120 million more over 25 years than using monocrystalline silicon (based on annual generation 800 million kWh).


Cost analysis for different country markets


l US large-scale plants: Land cost proportion <10%, priority is reducing initial investment. Polycrystalline silicon modules account for 45% of procurement (SEIA 2023), saving $120k per MW compared to monocrystalline silicon.

l European industrial/commercial rooftops: A 10MW factory project in Germany, polycrystalline silicon saves 15% on equipment costs vs monocrystalline silicon (approx. €1.5 million), but 25-year total revenue difference only 5% (due to 2% lower efficiency).

l Australian off-grid systems: Farm using polycrystalline silicon + storage, initial investment AUD 8k less than monocrystalline silicon (enough to buy 2 cell sets), sufficient for irrigation and basic electricity.


Financing costs are also lower


Lower initial investment, smaller loan pressure. US 10MW project:

l Polycrystalline silicon plan loan 8 million (interest rate 4%), monocrystalline silicon loan 9 million (same rate);

l Over 25 years, polycrystalline silicon total interest payment 3.2 million, monocrystalline silicon 3.6 million, saving $400k (BNEF 2023 calculation).


Mature supply chain, shorter delivery time


Polycrystalline silicon production line construction cycle 18 months, monocrystalline silicon requires 24 months (PV InfoLink 2023).

During Europe's 2022 energy crisis, polycrystalline silicon module delivery time was 2 months shorter than monocrystalline silicon, avoiding loss of feed-in tariff subsidies due to delayed grid connection.


Suitable Scenarios


Large-scale ground-mounted power plants:

Large ground-mounted plants cover extensive land, where land cost typically accounts for <10% of total cost (SEIA 2023). The primary goal is to minimize initial investment. Polycrystalline silicon's advantage here is saving 120k-150k per MW (compared to monocrystalline silicon).

l US California Mojave Desert 500MW project (commissioned 2022): Used polycrystalline silicon modules, initial investment $12 million less than monocrystalline silicon plan (Berkeley Lab data). Due to strong local irradiation (annual radiation 2,800 kWh/m²), polycrystalline silicon's low-light performance (+3%) generates more electricity in morning/evening hours, total annual generation 1.8% higher than projected.

l Spain Andalusia 300MW project (2023): Paired with single-axis trackers, polycrystalline silicon modules power loss at high temperatures (summer 40°C) is 2% greater than monocrystalline silicon (-0.40%/°C vs -0.35%/°C), but initial savings cover this loss, LCOE still €0.002/kWh lower (IEA assessment).

l Data comparison: Globally, polycrystalline silicon holds a 40% share in ground-mounted plants (IEA 2023), mainly because after the monocrystalline silicon price gap narrowed from 30% to 10%, polycrystalline silicon maintained market share with its "¥0.8 per watt cost red line" (BloombergNEF).

Industrial & commercial rooftops:

Industrial & commercial rooftop projects are often budget-constrained, and have ample installation area (e.g., factories, warehouses).

Polycrystalline silicon's square wafers adapt well to large flat surfaces, installation efficiency is 5% higher than monocrystalline silicon's round wafers (less inter-cell spacing).

l Germany Bavaria 10MW factory rooftop (2022): Budget capped at €5 million, polycrystalline silicon saved €750k (15%) compared to monocrystalline silicon. 25-year total revenue difference only 5% (due to 2% lower efficiency), but initial savings were enough to install a basic monitoring system (cost approx. €100k).

l Australia Queensland farm warehouse (2023): Rooftop area 20,000 m², polycrystalline silicon modules covering the roof provide 250kW, 30kW less than monocrystalline silicon, but saved AUD 12k initial cost.

l Note: If the roof has partial shading (e.g., ventilation ducts), polycrystalline silicon's low-light advantage is more pronounced, generating 4% more than monocrystalline silicon under shading (NREL shading test).

Off-grid systems:

Remote off-grid projects (farms, communication base stations) have minimal budgets. Polycrystalline silicon's low initial investment is the biggest attraction, and paired with storage it meets basic electricity needs.

l India Rajasthan 200MW off-grid project (2023): Powers desert villages, polycrystalline silicon modules account for 70% share. Each 5kW system (including storage) using polycrystalline silicon saves ₹8,000 (approx. $100) compared to monocrystalline silicon, enough to buy 2 sets of lead-acid batteries. Annual generation 6,000 kWh, meeting electricity needs for 10 households.

l Brazil Amazon rainforest communication base station (2022): Uses polycrystalline silicon + diesel generator hybrid power. Polycrystalline silicon initial investment R12k less than monocrystalline silicon (approx. 2,400). Due to the base station's low power load (daily average 20kWh), polycrystalline silicon's slightly lower efficiency does not affect usage, 25-year total O&M cost saved 15%.


Thin-film


CdTe (First Solar) reaches 22% mass production efficiency; CIGS labs hit 23.4% (NREL).

They weigh 1/10 of silicon, perform 5-10% better in low light, and have a lower temperature coefficient (0.25%/°C vs silicon's 0.45%/℃).


Materials and Structure


Amorphous Silicon (a-Si):

Its active layer thickness is 1-3μm, 50 times thinner than a human hair, relying on hydrogen (H) to "stabilize" the structure—hydrogen atoms account for 5-10% (atomic ratio), filling dangling bond defects in silicon atoms; otherwise electrons would collide and lose energy during transport.

The actual structure is three layers: top p-type a-Si: H (boron-doped), middle intrinsic i-a-Si: H (power generation layer), bottom n-type a-Si: H (phosphorus-doped), like a sandwich.

Manufacturing uses Plasma-Enhanced Chemical Vapor Deposition (PECVD), "spraying" a mixture of silane (SiH₄) and hydrogen (H₂) gasses onto a glass or stainless steel substrate, vacuum chamber pressure 10⁻³ Torr, temperature 200-300°C.

Kaneka (Japan) mass production line achieves a deposition rate of 0.5nm/s, taking 4 hours to coat three layers for a module (1.1m×1.4m).

But its flaw is "light-induced degradation": first-year light-induced degradation (Staebler-Wronski effect) can cause a 10% efficiency drop, followed by 0.5% annual degradation thereafter.

NREL measurements show 80% of initial efficiency remains after 10 years.

However, it wins on cost, module cost $0.3/W (2023 data), 20% lower than silicon panels.

Cadmium Telluride (CdTe):

Cadmium telluride (CdTe) is the "efficiency champion" among thin-film, with First Solar (USA) achieving 22.3% mass production efficiency (2023).

Structurally simpler: On a glass substrate, first coat a transparent conductive oxide (TCO, e.g., Fluorine-doped tin oxide FTO), then a cadmium sulfide (CdS) buffer layer (50-100nm thick, reduces interface defects), finally the CdTe absorber layer (3-4μm thick), back contact formed with copper doping for ohmic contact.

Heat CdTe powder to 600°C, letting vapor deposit onto a moving glass ribbon, cooling and crystallizing into a film.

First Solar's Arizona factory uses this method, production line speed 6 meters/minute, outputting 12,000 modules per day.

Its PN junction is a "heterojunction" (CdS and CdTe have different bandgaps), open-circuit voltage can reach 0.85V, 0.3V higher than amorphous silicon.

The drawback is cadmium content (toxic), so modules are encapsulated with double-layer tempered glass, edges sealed with epoxy resin.

EU REACH regulations require over 98% cadmium extraction during recycling.

Copper Indium Gallium Selenide (CIGS):

Structurally, five layers: plastic or metal foil substrate → molybdenum (Mo) back electrode (0.5μm thick, conductive and flexible) → CIGS absorber layer (2-3μm thick, core power generation layer) → buffer layer (CdS or Zn(O, S), 50nm thick, protects absorber) → window layer (i-ZnO + AZO, 1μm thick, transparent conductive).

The CIGS absorber layer composition is precise: copper (Cu), indium (In), gallium (Ga), selenium (Se) mixed in ratios Cu/(In+Ga)=0.8-1.0, Ga/(In+Ga)=0.2-0.3.

This allows bandgap tuning from 1.0eV to 1.7eV, adapting to different sunlight spectra.

Solar Frontier (Japan) uses co-evaporation for this layer: separately heating Cu, In, Ga, Se to 1200°C, 500°C, 900°C, 200°C, letting atoms co-deposit onto the substrate to "grow" the film, uniformity error <5%.

Flexible versions use polyimide (PI) substrate, thermal expansion coefficient matches CIGS (approx. 4×10⁻⁶/°C).

Efficiency remains after 100,000 bending cycles with bending radius <10cm.

Roll-to-Roll Manufacturing:

The process resembles a printing press: Unwind roll (glass/plastic roll) → cleaning (ultrasonic dust removal, water 60°C) → coating (spray layers sequentially) → laser scribing (create circuitry, precision ±50μm) → rewind (roll into 1.5km long module strips).

For example, First Solar's CdTe line, glass roll width 1.1 meters, speed 6 meters/minute, runs 24/7, annual output 1.2GW modules.

Equipment uses magnetron sputtering for TCO layer, power 10kW, target material tin oxide ceramic, sputtering rate 50nm/s.

Compared to silicon panels: Silicon wafer kerf loss 40% (from ingot to thin wafer), thin-film material utilization above 95%. NREL calculated CdTe material consumption 0.02 kg per watt, silicon panels require 0.5 kg (25x difference).


Hidden Details in Structural Design


Glass substrate (3.2mm thick) used for ground-mounted plants, resistant to weathering and pressure; stainless steel substrate (0.1mm thick) used for industrial roofs, fire-resistant;

Plastic substrate (PI or PET, 0.05-0.2mm thick) for flexible modules, can conform to curved surfaces.

Ground-mounted plants use double glass + EVA encapsulant (0.5mm thick), water vapor transmission rate <0.1g/m²/day;

Flexible modules use polymer films (PVF or PVDF), light and UV resistant.

First Solar's CdTe module frames use anodized aluminum, thickness 1.5mm, withstands 55m/s winds (equivalent to Category 16 typhoon), no rust after 2000-hour salt spray test.



Efficiency and Power Generation Capability


Mass Production Efficiency:

l Amorphous Silicon (a-Si): Most mature but lowest efficiency. Sharp (Japan) 2023 module mass production efficiency 7.2%, Kaneka (Japan) slightly higher at 8.1%. Lab maximum 10% (NREL 2024), but commercial products generally stuck at 6-8%, suitable for portable devices with low efficiency requirements.

l Cadmium Telluride (CdTe): First Solar (USA) leads, 2023 Series 7 module efficiency 22.3% (area 2.08m²), 1.2 percentage points higher than 2020's 21.1%. Arizona factory yield 98%, each module power 410W (silicon panel of same area about 460W).

l Copper Indium Gallium Selenide (CIGS): Solar Frontier (Japan) "Galaxy" series mass production efficiency 19.2% (2024), flexible version 17.5%. Lab maximum 23.4% (German HZB Institute), 1.1 percentage points higher than 2022, targeting 20% mass production by 2025.

Compared to monocrystalline silicon: TOPCon modules (US SunPower) mass production 25.8%, HJT (Japan Panasonic) 25.5%.

Thin-film efficiency generally 3-8 percentage points lower, but material usage is only 1/25 of silicon panels (NREL data).

Cloudy/Low-Light Conditions:

l Test Conditions: NREL (US National Renewable Energy Laboratory) 2023 simulation of cloudy days (200W/m², equivalent to dawn/dusk), overcast (100W/m²) and indoor lighting (50W/m²).

l Data Comparison (using same 100W module as example):

Irradiance (W/m²)

CdTe Output (W)

CIGS Output (W)

Monocrystalline TOPCon Output (W)

Thin-film Advantage (%)

200

85

83

70

+21.4

100

52

50

38

+36.8

50

28

27

19

+47.4

Hot Weather:

l Coefficient Comparison: CdTe 0.25%/°C (First Solar), CIGS 0.30%/°C (Solar Frontier), amorphous silicon 0.35%/°C (Sharp); monocrystalline silicon TOPCon 0.45%/°C (SunPower), HJT 0.38%/°C (Panasonic).

l Measured at 50°C: Module surface temperature 50°C (ambient 35°C + direct sun), CdTe power degrades 8% (100W module becomes 92W), CIGS degrades 10% (90W), silicon panel degrades 15% (85W) (NREL outdoor testing).

l Extreme Case: Phoenix, Arizona, USA (summer often 45°C), a CdTe plant annual energy yield 4.8% higher than silicon panels, due to less high-temperature degradation (First Solar O&M data).

Actual Energy Generation:

l Formula: Annual Generation (kWh) = Module Power (W) × Peak Sun Hours (h) × System Efficiency (0.85) × 365 / 1000.

l Case Comparison (Davis, California, USA, peak sun hours 5.5h/day):

l First Solar CdTe (410W module, area 2.08m²): 410 × 5.5 × 0.85 × 365 / 1000 ≈ 702 kWh/year/module → 337 kWh/m²/year.

l SunPower TOPCon (460W module, area 2.07m²): 460 × 5.5 × 0.85 × 365 / 1000 ≈ 787 kWh/year/module → 380 kWh/m²/year.

l Thin-film disadvantage: 11% less generation per unit area, but 80% lighter weight, suitable for old buildings with insufficient roof load-bearing (no need to reinforce racking).

Degradation Rate:

l Amorphous Silicon: First-year degradation 10% (Staebler-Wronski effect), then 0.5% annually, remaining 80% of initial efficiency after 10 years (Kaneka 25-year tracking data).

l CdTe: First Solar warranty 25 years, first-year degradation ≤0.5%, then ≤0.25% annually, remaining 91.25% of initial efficiency after 25 years (2023 warranty document). Actual Arizona plant operation 15 years, efficiency only dropped 5.8%.

l CIGS: Solar Frontier warranty 25 years, first-year degradation ≤1%, then ≤0.4% annually. Lab accelerated aging test (85°C/85% humidity) 2000 hours, efficiency dropped 3% (German TÜV certification).

l Compared to Silicon Panels: Monocrystalline silicon TOPCon first-year degradation ≤1%, then ≤0.4% annually.

Power Generation Capability Extension:

Thin-film can also be made bifacial, but gain is lower than silicon panels (silicon bifacial gain 10-25%, thin-film 5-15%).

l Bifacial CdTe: First Solar pilot line modules use transparent backsheet, gain 8% under grass-reflected light (Arizona test).

l Angle Adaptation: Thin-film's better low-light performance yields 3-5% more generation than silicon panels when laid flat (e.g., parking lot canopies) (Fraunhofer Institute Germany), due to higher proportion of diffuse light.