advances.sciencemag.org/cgi/content/full/4/11/eaau8131/DC1

Supplementary Materials for

Solid electrolyte interphases for high-energy aqueous aluminum electrochemical

cells

Qing Zhao, Michael J. Zachman, Wajdi I. Al Sadat, Jingxu Zheng, Lena F. Kourkoutis, Lynden Archer*

*Corresponding author. Email: laa25@cornell.edu

Published 30 November 2018, Sci. Adv. 4, eaau8131 (2018)

DOI: 10.1126/sciadv.aau8131

This PDF file includes:

Fig. S1. The SEM-EDX mapping spectra on the top view of TAL. Fig. S2. Rate performance of symmetric Al batteries using TAl and electrolyte of 1 m Al(CF3SO3)3 in water. Fig. S3. Symmetric Al battery performance using Al or TAl coupled with organic electrolyte. Fig. S4. The CV diagrams of Al-carbon fiber paper batteries. Fig. S5. ATR-FTIR spectra of different electrolytes. Fig. S6. Cross-sectional SEM image of TAl anode and corresponding EDX mapping after cycling in symmetric batteries. Fig. S7. SEM characterizations of MnO2 nanorod. Fig. S8. Galvanostatic discharge/charge curves of aqueous Al batteries using common Al anode. Fig. S9. Electrochemical properties of Al-MnO2 batteries using TAl anode–, Al(CF3SO3)3-, and Al2(SO4)3-based aqueous electrolyte. Fig. S10. GITT profiles of Al-MnO2 batteries. Fig. S11. Cycling performance comparisons with or without Mn(CF3SO3)2 addition at current density of 200 mA/g. Fig. S12. Galvanostatic discharge/charge curves at different current densities. Fig. S13. XRD patterns of MnO2 electrodes under different situations. Fig. S14. SEM images of MnO2 electrode. Fig. S15. SEM images and selected positions for EDX studies. Fig. S16. TEM images of MnO2 electrodes. Fig. S17. XPS Mn2p3/2 spectra of pristine MnO2, fully discharged MnO2 cathode, and fully charged MnO2 cathode. Table S1. EDX analysis of points in fig. S15. References (32–36)

Fig. S1. The SEM-EDX mapping spectra on the top view of TAL. Al, N and Cl are mainly

element on the surface of Aluminum.

Fig. S2. Rate performance of symmetric Al batteries using TAl and electrolyte of 1 m

Al(CF3SO3)3 in water. The current densities increase from 0.1 mA/cm2 to 0.5 mA/cm2.

Fig. S3. Symmetric Al battery performance using Al or TAl coupled with organic

electrolyte. The current density is 0.2 mA/cm2.

Fig. S4. The CV diagrams of Al-carbon fiber paper batteries. CV curves of cells using (A) Al

foil and (B) TAl foil as anode, carbon fiber paper (CP) as cathode, aqueous Al2(SO4)3 electrolyte

and Al(CF3SO3)3 electrolytes. The scanning rate is 1 mV/s and voltage region is from 0 V to 3 V.

After treating with ionic liquid electrolyte, the stability of Al2(SO4)3 is also slightly improved,

and the electrolytes with Al(CF3SO3)3 salts are remained.

Fig. S5. ATR-FTIR spectra of different electrolytes. Typical vibrations belong to the

functional group of CF3, SO3 and SO4 are detected in the electrolytes (32, 33), confirming the

salts of Al(CF3SO3)3 and Al2(SO4)3. When comparing the peak of water in different electrolyte,

the peaks show large blue shift for Al(CF3SO3)3 based electrolyte, indicating the interaction

between the large anion (CF3SO3-) and H2O. The interaction increases after increasing the

concentration of electrolyte.

Fig. S6. Cross-sectional SEM image of TAl anode and corresponding EDX mapping after

cycling in symmetric batteries. After cycling, a complicated SEI on Al is generated including

the element in salts of electrolytes. The organic layer is remained on the surface of Al because

the N element is still enriched. The AlCl3 is considered to be dissolved in the electrolyte to form

an acidic environment, which is not enriched on cycled TAl anode.

Fig. S7. SEM characterizations of MnO2 nanorod. (A), (B) SEM images and (C)

corresponding EDX mapping spectra of α-MnO2 nanorods.

Fig. S8. Galvanostatic discharge/charge curves of aqueous Al batteries using common Al

anode. (A) Using electrolyte of 1 m Al(CF3SO3)3 and (B) Using electrolyte of 2 m Al(CF3SO3)3

and Al2(SO4)3. The current density is 100 mA/g (MnO2). When using common Al as anode, the

batteries with 1 m Al(CF3SO3)3 can operate under large polarization. As comparisons, when

increasing the concentration of Al(CF3SO3)3 to 2m, the polarization is over 1 V. In that case, the

Al-MnO2 batteries can display a capacity near 300 mAh/g with the potential of about 0.3 V. For

Al2(SO4)3 electrolyte, the Al-MnO2 batteries can hardly work, and the charging process seem

endless during to the decomposition of electrolyte.

Fig. S9. Electrochemical properties of Al-MnO2 batteries using TAl anode–, Al(CF3SO3)3-,

and Al2(SO4)3-based aqueous electrolyte. Galvanostatic discharge/charge curves using (A) 1 m

Al(CF3SO3)3, (B) 3 m Al(CF3SO3)3, (C) 5 m Al(CF3SO3)3 (D) 1 m Al2(SO4)3 electrolyte at

current density of 100 mA/g; (E) corresponding cycling performance of 3 m Al(CF3SO3)3, 5 m

Al(CF3SO3)3, and 1 m Al2(SO4)3. All the electrolytes can operate coupled with TAl anode and

display narrow polarization. However, the cycling performance using higher concentration of

Al(CF3SO3)3 (3 m and 5 m) is not as good as low concentration, which may be caused by the low

ionic conductivity and thick SEI formed by high concentrated electrolyte.

Fig. S10. GITT profiles of Al-MnO2 batteries. (A) Using TAl, (B) using common Al. The

GITT profiles are obtained through the following steps: discharge/charge at 100 mA/g for 5 mins

and then rest for 120 mins. The profiles confirm that Al-stripping from Al anode is the sluggish

process for common Al anode. The artificial SEI formed by IL treating can largely facilitate this

process.

Fig. S11. Cycling performance comparisons with or without Mn(CF3SO3)2 addition at

current density of 200 mA/g. One reason of the capacity fading of Al-MnO2 batteries is due to

the dissolution of low valence manganese oxide. In this work, we find the cycling performance is

further improved by adding Mn(CF3SO3), which can deliver a capacity over 100 mAh/g after

100 cycles.

Fig. S12. Galvanostatic discharge/charge curves at different current densities.

Fig. S13. XRD patterns of MnO2 electrodes under different situations. XRD patterns of (A)

discharge/charge products on Ti foil and (B) discharge product on carbon fiber paper with or

without washing. In order to eliminate the effect of peaks belong to current collector, we also use

Ti foil current collector to detect the product. The results are similar with CP as current collector.

The new discharge peaks is water and ethanol soluble, and not belong to any peaks of salt in

electrolyte.

Fig. S14. SEM images of MnO2 electrode. Discharged electrode (A) without washing, (B) after

washing by water. Charged MnO2 electrode (C) without washing, (D) after washing by water.

After discharge, the morphologies of MnO2 electrode is covered with soft product, which can be

easily washed using water or ethanol. As comparisons, the charged product shows clear nanorod

structures even without washing.

Fig. S15. SEM images and selected positions for EDX studies. The figures A, B, C, D

corresponds to the discharged MnO2 electrode without washing, discharged MnO2 electrode after

washing, charged MnO2 electrode without washing, charged MnO2 electrode after washing. The

atomic ratio of Al and Mn is calculated as displayed in Fig. 4B and table S1. The ratio of Al and

Mn subtracting Al in electrolyte is calculated through the relationship of Al: S=1: 3 in

Al(CF3SO3)3. The voltage used for EDX analysis is 20 kV.

Fig. S16. TEM images of MnO2 electrodes. TEM images of (A) pristine MnO2 nanorod and (B)

discharged MnO2 nanorod.

Fig. S17. XPS Mn2p3/2 spectra of pristine MnO2, fully discharged MnO2 cathode, and fully

charged MnO2 cathode. The fitting of Mn2p3/2 has been widely used to analyze the oxidation of

Mn (34, 35). According to fitting curves of the XPS spectra, for pristine MnO2, most of oxidation

state of Mn is Mn4+, with small a bit of Mn2+, Mn3+. For fully discharge cathode, most of

oxidation state of Mn is Mn3+, indicating the reduction of MnO2. This process is reversible in

charged cathode.

Table S1. EDX analysis of points in fig. S15.

Element

(atom%)

C O F K Al S Mn Al/Mn Al/Mn

(subtract Al

from

electrolyte)

A1 39.11 24.33 19.24 0.88 6.65 3.31 6.48 1.03 0.86

A2 36.74 26.08 19.93 0.99 6.30 2.72 7.25 0.87 0.74

A3 35.77 22.28 20.99 1.00 8.64 3.85 7.46 1.16 0.99

B1 53.68 23.51 4.66 1.67 2.79 0.26 13.44 0.21 0,20

B2 62.82 16.42 7.70 1.27 1.89 0.14 9.75 0.19 0.19

B3 68.03 14.53 5.21 1.10 1.99 0.14 9.00 0.22 0.22

C1 9.95 48.01 15.82 1.97 4.70 2.96 16.59 0.28 0.22

C2 10.03 46.69 13.83 1.85 6.50 4.15 16.95 0.38 0.30

C3 13.09 44.34 15.67 1.89 5.45 3.53 16.03 0.34 0.27

D1 61.67 15.18 7.56 0.66 2.90 0.28 11.75 0.25 0.24

D2 42.05 30.81 8.72 1.34 1.98 0.20 14.90 0.13 0.13

D3 56.19 23.32 7.87 0.81 1.82 0.17 9.81 0.19 0.18

A1-D3 stands for the points in fig. S15. There are many sources of C and O element (such as

MnO2, super P, binder, substrate of SEM stage), which are always taking up the most part of

composition. The F exists both in Al(CF3SO3)3 and PVdF binder. K is introduced from the

preparation process (36). The S is originally exists in electrolyte, which share the atomic ratio

with Al at 3:1. After discharge, the ratio is much lower than 3:1, indicating Al also exist in

product, which is also proved by the high Al/Mn ratio. After washing by water, most of S and Al

are disappeared, indication the Al-based discharge product are water soluble. For charging

process, the ratio of Al/Mn is much smaller, indicating the Al striping from the discharge product.

Although the EDX spectra are semi-quantitative analysis methods, it can clearly illustrate the

trends of element change in MnO2 cathodes.