Scaling up open-source batteries: what’s worth pursuing?
Kirk Pollard Smith, PhD
Flow Battery Research Collective
Daniel Fernandez Pinto, PhD
Flow Battery Research Collective
January 31, 2026
Why open source energy storage?
- No open source energy storage solution exists.
- The task is not trivial.
- Currently no clear path to create rechargeable energy storage with reproducible characteristics from raw materials.
- An open source alternative will improve energy independence and enhance industrial development.
- An open source alternative should improve energy storage access in developing countries.
Goals
- Open source licensing.
- Medium and large scale storage.
- Low material costs.
- High reliability.
- High material availability.
- Openly available characterization.
- Open source characterization hardware and software.
A benchtop test cell
![]()
Zinc-iodine
Kit tested with zinc-iodine chemistry achieved hundreds of hours of stable cycling.
Zn-I battery chemical costs ~80 USD/kWh
Negative Terminal (Anode): \(\ce{Zn_{(s)} -> Zn^2+ + 2e-}\)
Positive Terminal (Cathode): \(\ce{I3- + 2e- -> 3I- }\)
Iron electrolytes
Currently testing lower cost all-Fe chemistries using innovative MgCl₂ and CaCl₂ electrolytes.
Fe-CaCl₂ battery chemical costs ~30 USD/kWh
Negative Terminal (Anode): \(\ce{Fe_{(s)} -> Fe^2+ + 2e-}\)
Positive Terminal (Cathode): \(\ce{Fe^3+ + e- -> Fe^2+}\)
Development of a 175 cm² cell
![]()
Stacking the 175 cm² cell
![]()
Static batteries
- Achieved good stability and high energy density
- 30-35Wh/L, comparable to lead batteries
- Cost is 8-16 USD/kWh
Cost estimate for Cu-Mn: graphite felt
| CuSO₄·5H₂O |
5.71 kg |
$1.40/kg |
$8.00 |
| MnSO₄·H₂O |
3.86 kg |
$0.45/kg |
$1.74 |
| Methanesulfonic acid (70%) |
7.85 kg |
$1.55/kg |
$12.17 |
| Conductive carbon felt (battery grade) |
5.71 m² |
$25/m² |
$142.85 |
| Graphite current collectors (cathodes), 2 mm assumed, 2.857 m² |
10.29 kg |
$6/kg |
$61.71 |
| Copper current collectors (anodes), 0.5 mm assumed, 2.857 m² |
12.80 kg |
$5.90/kg |
$75.52 |
| Celgard separator |
2.857 m² |
$2.20/m² |
$6.29 |
- Full 1kWh battery estimated material cost: ~310 USD
- For comparison FeLiPo cost/kWh is currently 60-80 USD
Cost estimate for Cu-Mn: coke?
| CuSO₄·5H₂O |
5.71 kg |
$1.40/kg |
$8.00 |
| MnSO₄·H₂O |
3.86 kg |
$0.45/kg |
$1.74 |
| Methanesulfonic acid (70%) |
7.85 kg |
$1.55/kg |
$12.17 |
| Calcined petroleum coke (CPC) |
22.6 kg |
$0.60/kg |
$13.56 |
| Activated carbon (10 wt%) |
2.66 kg |
$2.37/kg |
$6.30 |
| Conductive carbon black (3 wt%) |
0.80 kg |
$1.20/kg |
$0.96 |
| PTFE binder (2 wt%) |
0.53 kg |
$10/kg |
$5.30 |
| Separator |
2.86 m² |
$1.70/m² |
$4.86 |
| Titanium tabs (only) |
small |
- |
<$2 |
Full 1kWh battery estimated material cost: ~55 USD
Our plan
- Continue to test and validate our large scale flow battery cell.
- Produce stable charge/discharge curves with the large flow cell.
- Implement a 1kWh stacked flow battery using Zn-I at 12V.
- Validate the Fe chemistry at a small scale so that we can pursue a large scale implementation.
- Test a small scale pretroleum coke Cu-Mn battery.
- Test a medium scale design for a Cu-Mn battery (35-100Wh).
- Continue to explore interesting chemistries at a small scale.
Acknowledgements
- Josh Hauser, Prof Sanli Faez & FAIR Battery team from Utrecht University
- Alexander Quinn from the Brushett group at MIT (@quinnale)
- All the forum members, esp. @sepi, @gus, @czahl
- All our financial contributors on our Open Collective
- NLnet’s NGI0 Entrust fund
Current cell design
Flow frame design inspired by O’Conner, Bailey et al.1
Current status: benchtop cell, “development kit”
![]()
Specified entire system: pumps, tubing, reservoirs, documentation etc. Low-cost, widely available, safe components/materials for ease of replication.
Assembled cell
![]()
Simple power electronics
![]()
In the lab
![]()
First chemistry
Our initial chemistry is zinc-iodine (architecture inspired by Xie et al.1 and electrolyte by Lee et al. 2), but we are exploring more varieties, such as: all-iron, zinc-iron, soluble iron-manganese (with chelates)
Negative Terminal (Anode): \(\ce{Zn_{(s)} -> Zn^2+ + 2e-}\)
Positive Terminal (Cathode): \(\ce{I3- + 2e- -> 3I- }\)
Overall: \(\ce{Zn_{(s)} + I3- -> Zn^2+ + 3I-}, E^\ominus = 1.3 V\)
Parasitic reaction: \(\ce{6I- + O2 + 2 H2O -> 2I3- + 4OH- }\)
Triethylene glycol is added to form soluble iodide complexes at higher SOCs
Why Zn/I?
Easy to source, low-cost reagents (vs. vanadium, chromium…)
Compatible with cheap microporous membranes, such as paper
Resistant to dendrites
No appreciable hydrogen evolution
Acceptable energy density (>20Wh/L)
No strong acids or bases needed
Lower toxicity (vs. vanadium, chromium…)
Exploring all-iron, water-in-salt electrolytes (WiSE)
Based on Liu et al1, all-iron hybrid RFB approach using highly concentrated divalent chloride salts, e.g. 4.5 M MgCl₂ or CaCl₂ alongside FeCl₂.
Hydrogen evolution greatly reduced.
Initial testing in progress, including approx. viscosity measurements: https://fbrc.nodebb.com/topic/44/only-fe-system/21
Scaling up: control of centrifugal pumps
![]()
The pumps we will used for the large-format cell, 6 L/min magnetically-driven
175 cm² flow frame CFD
![]()
Design in FreeCAD, model in OpenFOAM
175 cm² flow frame CFD
![Design in FreeCAD, model in OpenFOAM]()
## Assembling the 175 cm² cell {visibility=“uncounted”}
Leak testing the 175 cm² cell
## Assembling the 175 cm² cell {visibility=“uncounted”}