Aluminium (Aluminum) Powder for Solar Cell Paste: Particle Size, Purity and Grade Selection 2026

A paste that prints clean can still come out of the furnace with a dull, patchy rear face. Or the wafer bows. Or adhesion fails at the busbar after encapsulation. By the time you see any of it, the printing parameters have been blamed, then the firing profile, then the batch of paste.

The aluminium powder is usually the last thing anyone suspects. That is odd, because it is 70 to 80 per cent of the paste by weight.

What the aluminium layer does

The rear aluminium paste is screen-printed, dried, and co-fired with the front silver. Somewhere above 577 °C, the aluminium-silicon eutectic temperature, the aluminium in contact with the wafer turns liquid, dissolves a few micrometres of silicon, and on cooling that silicon regrows doped with aluminium. The result is the p+ back surface field: it keeps electrons away from the rear surface, gives the cell an ohmic back contact, and the aluminium left behind reflects long-wavelength light back into the absorber.

Aluminium melts at 660 °C, so the peak of a co-firing profile is normally quoted somewhere between 750 and 850 °C, held for a few seconds. Short window, high temperature, and a lot of the paste’s behaviour in that window is decided before it ever reaches the printer.

Four numbers a formulator reads first

NumberWhat it drivesGet it wrong and
D50 and the top cutHow much silicon the melt dissolves, and how evenlyPatchy BSF, blisters, pinholes from a coarse tail
Active aluminiumMetal available to alloy, not locked in oxideHigh contact resistance, weak back field
Tap densitySolids loading you can reach at a printable viscosityThin, grainy print, or a paste that cannot be printed thin enough
Fe, Cu, moistureCarrier lifetime and paste stabilityEfficiency loss that no furnace tweak recovers

D50 sits at the centre of it. A coarser powder dissolves more silicon and leaves a thicker alloyed layer, which helps contact and hurts warp. A finer powder sinters earlier and more completely, which can blister the rear if the organics have not burned out yet. The top cut matters as much as the median: a 6 µm median with a coarse tail still shows up as speckles and pinholes on a 180 µm wafer, so ask for both figures on the same certificate.

Active aluminium is the fraction that can actually react. The rest is the oxide skin, which every aluminium particle carries and which thickens with time and moisture. Some oxide is unavoidable and useful, since it keeps the powder from being pyrophoric. Too much of it, and you are paying for metal that will not alloy.

Why paste makers moved to spherical

A spherical particle rolls. It meters through a feeder without bridging, packs to a higher tap density, and lets you raise solids at the same viscosity. That last point is the whole game in rear paste: more metal per wet micrometre means a thinner print, less vehicle to burn out, and less shrinkage stress on a wafer that is getting thinner every generation.

An irregular or flake particle has more surface area per gram, which sounds like an advantage until the viscosity climbs and the printed edge definition goes. Flakes also orient under the squeegee, so shrinkage stops being isotropic. Flake aluminium earns its place in pigments and coatings. In rear paste, the market settled on spherical for reasons that hold up.

What we can supply, and the limit

Our fine aluminium powder range runs from FLQT0 at 43 µm down to FLQT5 at 2 µm, produced and tested to GJB 1738A-2015, with tap density up to 1.6 g/cm³ on the FLQT1 to FLQT4 cuts.

GradeD50 (µm)Tap density (g/cm³)Where it fits
FLQT52 ± 1—Fine rear paste, thin prints
FLQT46 ± 1.5≥ 1.4Mainstream rear paste
FLQT313 ± 2≥ 1.5Coarser pastes, thicker alloyed layer
FLQT1–FLQT229 ± 3 / 24 ± 3≥ 1.6Beyond paste work

Typical chemical limits across the range: active aluminium ≥ 98% (97% on FLQT5), Fe ≤ 0.2%, Cu ≤ 0.015%, Si ≤ 0.2%, moisture ≤ 0.10%. Anything between the standard cuts can be produced to order.

The limit is worth stating plainly. If your paste needs a median below 2 µm, or a surface treatment we do not run, we will tell you instead of shipping something that fails on your line. Two micrometres is where our standard range stops, and a sample trial is the only honest way to find out whether that is fine enough for your print thickness and firing profile.

Running a sample through your line

Ask for one to two kilograms. Print it at the viscosity you actually run, not at the viscosity in the data sheet. Then measure the same things you measure for your current supplier:

  • Viscosity at two shear rates, before and after 24 hours of rest
  • Print weight and wet thickness on your own screen
  • Cross-section of the alloyed layer, and whether it is continuous
  • Adhesion after encapsulation, and warp after co-fire
  • I-V on the same wafer batch, so the comparison is not confounded

Hold the furnace profile fixed while you swap the powder. If you change both at once you will learn nothing, and most sample trials fail for exactly that reason. Our particle size guide explains how D50 maps to behaviour if you need to justify a grade change internally.

Freight and paperwork

Uncoated fine aluminium powder is UN 1396, Class 4.3: it releases flammable hydrogen on contact with water. Coated powder moves as UN 1309, Class 4.1. The HS code follows the particle structure, 7603.10 for non-lamellar powder and 7603.20 for flake, so spherical paste grades fall under 7603.10. We issue an MSDS and a dangerous goods declaration with every shipment, and air freight is often refused for fine cuts, so plan on sea freight and a few weeks of transit.

Storage is the other half of it. Keep the powder dry, keep acids and oxidisers out of the same area, and never put water on burning aluminium powder. If you are still comparing suppliers, our supplier selection guide covers MOQ, samples and the documents worth asking for.

FAQ

Q: What aluminium powder is used in solar cell paste?

A: Fine spherical aluminium powder, usually with a single-digit micrometre median, high active aluminium and low iron and copper. The sphere packs well enough to print thin, and the fine median lets the alloyed layer form in a few seconds.

Q: What D50 should I specify for rear paste?

A: It follows your print thickness and firing profile rather than a fixed number. Most rear pastes sit in the single-digit micrometre range; our FLQT4 at 6 µm and FLQT5 at 2 µm cover that ground.

Q: Spherical or irregular powder for paste?

A: Spherical, for packing and print definition. Irregular and flake particles raise viscosity at the same solids loading and shrink anisotropically.

Q: How do I qualify a sample before ordering?

A: One to two kilograms, printed at production viscosity, furnace profile unchanged, and measured on the same wafer batch as your current powder: alloyed layer continuity, adhesion, warp, and I-V.

Q: What HS code and UN number apply?

A: HS 7603.10 for non-lamellar (spherical) powder, 7603.20 for flake. Uncoated powder is UN 1396, Class 4.3; coated powder is UN 1309, Class 4.1.

Get a grade recommendation

    Tell us what you need, and an engineer replies within one working day.

    Grade or D50, application and monthly quantity are enough to start. Samples from 1 kg.

    Prefer email? contact@nmgjinhui.cn, or call +86 139 4071 2462.

    Manufacturing since 2019Exports to 50+ countriesQuality, environment, OHS and energy systems certifiedCertificates and plant photos →

    By submitting this form, you agree to our Privacy Policy.