Oxidized starch as a superior binder for silicon anodes in lithium-ion batteries

Abstract

Commercial oxidized starch (OS) containing oxidized amylose and oxidized amylopectin is proposed as a binder for silicon anodes. The deformable OS with flexible helix-type structure and elastic α-linkages in the backbone, and the multi-dimensional interaction between OS and silicon can effectively improve the cycle performance of strong-volume-change silicon anodes.

---

As an important anode material in lithium ion batteries with a high theoretical capacity (4200 mA h g⁻¹, Li₂₂Si₅) and moderate lithiation potential (0.4–0.5 V vs. Li/Li⁺),¹ silicon has been intensively investigated in the past decade. However, the poor conductivity of Si and dramatic volume effect upon lithiation and delithiation lead to an unstable solid electrolyte interface (SEI), pulverization of Si particles and breakdown of the conductive network. Consequently, its electrochemical performance is greatly limited.^2–4 Enormous efforts have been dedicated to the design of Si based anode materials,³ such as porous structure,^5–7 nano size Si⁸ and Si/C composites.⁹ In addition to the active material, the binder plays a very critical role in a silicon electrode and directly affects the electrochemical performance.¹⁰ It should be mentioned that in recent years there have been many publications concerning binder-free electrode structure and fabrication.^11–13 Compared with traditional binder-containing electrodes, the binder-free strategy should pay more attention to the feasibility of large-scale production and high loading of active materials.

Compared to conventional poly(vinylidenefluoride) (PVDF), binders containing carboxylic acid and hydroxyl functional groups such as sodium carboxymethyl cellulose (CMC),¹⁴ sodium alginate (SA),¹⁰ polyacrylic acid (PAA)¹⁵ can obtain superior cycling performance utilizing hydrogen bonding interaction with nature oxidized silicon surface. In addition to strong adhesion between binder and silicon,¹⁶ the mechanical properties of binders are also important.^17–20 Multi-branched structure²¹ and cross-linking between polymers (PAA/CMC²² and PAA/polyvinyl alcohol (PVA)²³) have been reported to be effective in increasing the rigidity and improving cycle retention. Nevertheless, too rigid or stiff polymers are unable to completely relieve the stress caused by volume expansion of Si during lithiation, leading to macroscopic cracks on electrode and poor cycle performance.^23,24 Pullulan (based on α-glycosidic linkages) cross-linked with PAA shows more elastic properties than CMC (based on β-glycosidic linkages) cross-linked with PAA, resulting in the better cycle retention.²⁵ And pectin (based on α-glycosidic linkages) is reported to be a better binder for Si anode than CMC.²⁴ The difference of elasticity between pullulan, pectin and CMC comes from the different types of glycosidic linkages between constituent units or residues.^24,25 The β linkages are regarded to provide stiffer mechanical properties than their counterparts.²⁶

In an effort to find superior polysaccharide binders for practical application of silicon anode, we resort to combine elastic α linkages, multi-branched structure and excellent adhesion together in one binder. In particular, we pay attention to commercial oxidized starch (OS) from corn starch (Fig. 1), which contains 25–30% amylose and 70–75% amylopectin. When OS is applied as binder, Si anode shows much better electrochemical properties than anodes with frequently used SA and CMC, owing to its flexible structure and the strong multi-dimensional adhesion resulting from abundant polar groups in linear and branched polymers.

Fig. 1 Molecular structures of CMC, SA and OS.

As shown in Fig. 1, CMC is composed of D-glucose units or their derivatives linked via β-(1 → 4) linkages and SA is a copolymer of β-(1 → 4) linked D-mannuronic acid and α-(1 → 4) linked L-guluronic acid.¹⁰ The structure of OS is also shown in Fig. 1. The oxidation reaction of starch occurs mainly on C₁ and C₆ in the pyranose, breaking down the linkages and transforming –CH₂OH into –CHO or –COOH.^27,28 So OS is a mixture of oxidized amylopectin containing α-(1 → 4) and α-(1 → 6) linked D-glucose units or their derivatives and oxidized amylose composed of α-(1 → 4) linked D-glucose units or their derivatives. Since the β glycosidic linkages provide stiffer mechanical properties than their α counterparts,²⁶ both CMC and SA show less elastic properties than OS. The major content of OS, oxidized amylopectin, is multi-branched, which is proved to be effective for binding Si.²¹ The abundant polar groups in OS can develop a robust interaction between OS and Si.

Fig. 2 shows the deformation of α-(1 → 4) and β-(1 → 4) linkages,²⁵ and the corresponding third level structure of their linear polymers. For α-(1 → 4) linkages, the oxygen atoms linked on C₁ and C₄ are in the same side of the pyranose rings, thus will not blocking the changes from chair type to boat type. In the opposite, the oxygen atoms linked on C₁ and C₄ in β-(1 → 4) linkages are in different sides of the pyranose rings, which fixes the rings in chair type. Moreover, the third level structure of linear polymer based on α-(1 → 4) linkages is flexible helix-type while stiff stretched ribbon-type corresponds to β-(1 → 4) linkages. As a result, oxidized amylose composed of α-(1 → 4) linked D-glucose units or their derivatives is more flexible than CMC and SA containing β-(1 → 4) linkages, tending to accommodate the volume change of silicon during cycling. Beyond that, oxidized amylopectin is disordered cluster structure, containing numerous side chains. This multi-branched structure provides omnibearing strong adhesion sites like octopus, resulting in multi-dimensional connection between oxidized amylopectin and silicon.

Fig. 2 Molecular-level conformational changes of pyranose units of different glycosidic linkages during elongation and the corresponding third level structures.

To investigate the polar groups of OS and the reaction between OS and Si, FTIR spectra have been recorded (Fig. 3(a)). For OS, a broad absorption band at 3200–3600 cm⁻² is related to the stretching of H–O in C–OH and a peak at 2937 cm⁻² corresponds to the stretching of C–H in –CH₂.¹⁷ Peaks at 1651 cm⁻¹ and 1420 cm⁻¹ are characteristic of –COOH.²⁹ And the peaks at 1162 cm⁻¹, 1086 cm⁻¹ and 1010 cm⁻¹ are related to C–O–C in pyranose rings. –OH and –COOH in OS can form hydrogen bonds with –OH on the surface of silicon and build a firm connection between binder and silicon. After OS was mixed with Si, the bands assigned to O–H and –COOH obviously shift to lower wave numbers, demonstrating that an interaction occurs between the OS binder and Si.³⁰ To evaluate the binding strength of binder, conventional 180° peeling tests are conducted on electrodes with OS, CMC and SA.^25,31 The film thickness of each electrode was 18 μm on average with a mass ratio of Si/SP/binder = 3 : 1 : 1. The tests were repeated for 3 times to guarantee the accuracy and all electrodes are under the same test conditions, including composition and loading. Fig. 3(b) shows the results of max detached forces for electrodes with OS, CMC and SA. The value for OS electrode is 3.022 N, higher than 2.879 N for CMC and 1.288 N for SA. These results indicate the excellent mechanical strength of the coating layer and binding effect to current collector in OS electrode. It should be mentioned that the mechanical properties of the polymer itself, which are measured by traditional tension tests and nano-scratch tests, may be different when the electrolyte solution exist.

Fig. 3 (a) FTIR spectra of OS and OS + Si. (b) Results of peeling tests for OS, CMC, SA electrodes (speed: 100.0 mm s⁻¹).

Fig. 4(a) shows the cycle performance and the corresponding coulombic efficiencies of OS, CMC and SA electrodes with silicon loading of ca. 0.8 mg cm⁻². Initial capacities and efficiencies of CMC and SA electrodes are 3127.2 mA h g⁻¹ with 85.6% and 2252.0 mA h g⁻¹ with 81.5%, respectively. The low initial capacity and efficiency of SA electrode may due to the poor dispersion ability of Super P in SA solution. In contrast, OS electrode delivers high initial capacity of 3295.4 mA h g⁻¹ with 85.1% coulombic efficiency, corresponding to ca. 2.5 mA h cm⁻². After 120 cycles, a specific capacity of 1904.2 mA h g⁻¹ is obtained for OS electrode with capacity retention of 74% at the same current. However, conventional CMC and SA electrodes present only 918.3 mA h g⁻¹ and 657.7 mA h g⁻¹ after the same cycles, corresponding to capacity retention of 36% and 31%, respectively. The superior cycle performance of OS electrode should benefit from its strong binding strength and proper elasticity. Moreover, the coulombic efficiencies of CMC and SA electrodes fluctuate a lot during cycling while OS electrode exhibits stable coulombic efficiencies around 99%, indicating more stable interface and electrode structure.

Fig. 4 (a) The cycle performance and coulombic efficiencies of OS, CMC and SA electrode. 0.1 A g⁻¹ is for first two cycles and 0.3 A g⁻¹ for next two cycles. And for the rest cycles, discharge current is 0.5 A g⁻¹ and charge current is 2.0 A g⁻¹. (b) The rate performance of OS, CMC and SA electrodes.

Fig. 4(b) shows the rate performance of OS, CMC and SA electrodes with silicon loading of ca. 0.5 mg cm⁻². At the current density of 0.1 A g⁻¹, CMC and OS electrodes show similar capacities around 3500 mA h g⁻¹ while SA electrode exhibits capacity about 2500 mA h g⁻¹. These results are in accord with Fig. 4(a). When the current density increased to 0.5 A g⁻¹, the difference between CMC and OS electrodes becomes obvious. At 4.0 A g⁻¹, OS electrode exhibits 2183.0 mA h g⁻¹, corresponding to 63% of the capacity at 0.1 A g⁻¹, while CMC electrode shows only 765.7 mA h g⁻¹, corresponding to 22% of the capacity at 0.1 A g⁻¹. And a reversible capacity of 1454.5 mA h g⁻¹ is obtained for OS electrode at high current density of 8.0 A g⁻¹ while CMC and SA electrodes deliver capacities of 111.1 mA h g⁻¹ and 526.7 mA h g⁻¹, respectively. The excellent rate performance of OS electrode should be attributed to its effective conductive network formed by uniformly dispersed Super P and good electrolyte uptake ability of OS, which ensures fast lithium ion diffusion in the electrode. The proper mechanical properties and strong adhesion strength of OS also contribute to the stable electrode structure and high Si utilization at large current densities.

The charge and discharge curves of different electrodes corresponding to Fig. 4(a) are shown in Fig. 5. The differences between 1st cycle and 5th cycle are due to different current densities. After 120 cycles, all electrodes show capacity decay and voltage polarization. Especially for CMC and SA electrodes, the charge and discharge platforms shift largely, indicating aggravated voltage polarization accompanied with serious structure degradation. In comparison, OS electrode exhibits fairly stable voltage trends, suggesting relatively stable electrode structure owing to strong binding strength and excellent mechanical properties of OS.

Fig. 5 The charge and discharge curves of the electrodes with OS, CMC and SA binders. 0.1 A g⁻¹ is for the first two cycles. The charge current is 2.0 A g⁻¹ and discharge current is 0.5 A g⁻¹ from 5th cycle.

It has been reported that the distinct binder effects are reflected in the surface morphology,^24,32 so we imaged the electrode surfaces in different cycling stages using scanning electron microscopy (SEM) (Fig. 6). The electrodes composition and discharge–charge conditions are the same as in Fig. 4(a). Distinct cracks of 2–4 μm on conventional CMC and SA electrodes after cycling indicate that such linear stiff binders are unable to endure the repeated silicon expansion and extraction during cycling, which results in poor cycle performance in Fig. 4(a) and serious polarization in Fig. 5. On the contrary, OS can stabilize the electrode morphology and only slight crevices can be observed after cycling.

Fig. 6 SEM images of the Si electrode surfaces containing (a) OS, (b) CMC, and (c) SA before and after battery cycles.

Fig. S1† shows the Nyquist plots of OS electrode after 5, 20 and 35 cycles. The depressed semicircle in the high–middle frequency region mainly corresponds to the charge transfer resistance (R_ct), and an oblique straight line in the low frequency region relates to the ion diffusion kinetics.³³ Compared to the 5th cycle, R_ct of the cell in 20th cycle declines slightly, which may result from the electrode activation in the initial cycling stage. After 20 cycles, no obvious change of R_ct is observed, further indicating the stability of OS electrode structure during cycling.

Table S1† presents a detailed literature comparison of competitive binders for silicon. Due to the large number of variables (regarding both compositional and testing conditions), it is difficult to make a strict comparison between different systems. Since areal capacity has a direct relationship with electrode loading and binder effect, we specially list it as a part of the electrode performance. Obviously, OS electrode in our work performs better than all other samples in Table S1,† further proving that OS is a promising binder for the silicon electrode.

Conclusions

In summary, OS has been proposed as a superior polysaccharide binder for silicon anodes due to its molecular-level conformation, multi-branched structure and strong adhesion force. Abundant polar groups in OS and the multi-branched structure of oxidized amylopectin provide omnibearing adhesion sites like octopus, thus leading to a strong adhesion between binder and Si. Beyond that, deformable OS with flexible helix-type structure and the elastic α-linkages in backbone can effectively accommodate the large volume change upon cycling, resulting in excellent cycling stability. The Si anode with areal capacity of 2.5 mA h cm⁻² can cycle for 120 times with capacity retention of 74%. Moreover, the OS electrode also exhibits excellent rate performance with a reversible capacity of 1454.5 mA h g⁻¹ at high current density of 8.0 A g⁻¹. The outstanding binding capability, low cost, and nontoxicity make OS an excellent binder for Si anodes.

Notes and references

1. H. K. Liu, Z. P. Guo, J. Z. Wang and K. Konstantinov, J. Mater. Chem., 2010, 20, 10055–10057.
2. S. F. Lux, I. T. Lucas, E. Pollak, S. Passerini, M. Winter and R. Kostecki, Electrochem. Commun., 2012, 14, 47–50.
3. M. A. Rahman, G. Song, A. I. Bhatt, Y. C. Wong and C. Wen, Adv. Funct. Mater., 2016, 26, 647–678.
4. X. H. Liu, H. Zheng, L. Zhong, S. Huan, K. Karki, L. Q. Zhang, Y. Liu, A. Kushima, W. T. Liang, J. W. Wang, J. H. Cho, E. Epstein, S. A. Dayeh, S. T. Picraux, T. Zhu, J. Li, J. P. Sullivan, J. Cumings, C. S. Wang, S. X. Mao, Z. Z. Ye, S. L. Zhang and J. Y. Huang, Nano Lett., 2011, 11, 3312–3318.
5. K. X. Wang, X. H. Li and J. S. Chen, Adv. Mater., 2015, 27, 527–545.
6. Z. D. Lu, N. Liu, H. W. Lee, J. Zhao, W. Y. Li, Y. Z. Li and Y. Cui, ACS Nano, 2015, 9, 2540–2547.
7. J. Ryu, D. Hong, S. Choi and S. Park, ACS Nano, 2016, 10, 2843–2851.
8. L. Zong, B. Zhu, Z. Lu, Y. Tan, Y. Jin, N. Liu, Y. Hu, S. Gu, J. Zhu and Y. Cui, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 13473–13477.
9. L. Zhang, R. Rajagopalan, H. Guo, X. Hu, S. Dou and H. Liu, Adv. Funct. Mater., 2016, 26, 440–446.
10. I. Kovalenko, B. Zdyrko, A. Magasinski, B. Hertzberg, Z. Milicev, R. Burtovyy, I. Luzinov and G. Yushin, Science, 2011, 334, 75–79.
11. G. Zhu, Z. He, J. Chen, J. Zhao, X. Feng, Y. Ma, Q. Fan, L. Wang and W. Huang, Nanoscale, 2014, 6, 1079–1085.
12. Y. Liu, W. Wang, Y. Wang, Y. Ying, L. Sun and X. Peng, RSC Adv., 2014, 4, 16374.
13. M. Zhang, F. Yan, X. Tang, Q. Li, T. Wang and G. Cao, J. Mater. Chem. A, 2014, 2, 5890.
14. H. Buqa, M. Holzapfel, F. Krumeich, C. Veit and P. Novák, J. Power Sources, 2006, 161, 617–622.
15. A. Magasinski, B. Zdyrko, I. Kovalenko, B. Hertzberg, R. Burtovyy, C. F. Huebner, T. F. Fuller, I. Luzinov and G. Yushin, ACS Appl. Mater. Interfaces, 2010, 2, 3004–3010.
16. M.-H. Ryou, J. Kim, I. Lee, S. Kim, Y. K. Jeong, S. Hong, J. H. Ryu, T.-S. Kim, J.-K. Park, H. Lee and J. W. Choi, Adv. Mater., 2013, 25, 1571–1576.
17. M. Ling, Y. N. Xu, H. Zhao, X. X. Gu, J. X. Qiu, S. Li, M. Y. Wu, X. Y. Song, C. Yan, G. Liu and S. Q. Zhang, Nano Energy, 2015, 12, 178–185.
18. C. Wang, H. Wu, Z. Chen, M. T. McDowell, Y. Cui and Z. A. Bao, Nat. Chem., 2013, 5, 1042–1048.
19. Z. Chen, C. Wang, J. Lopez, Z. Lu, Y. Cui and Z. Bao, Adv. Energy Mater, 2015, 5, 1401826.
20. Y. Bie, J. Yang, X. Liu, J. Wang, Y. Nuli and W. Lu, ACS Appl. Mater. Interfaces, 2016, 8, 2899–2904.
21. M. S. Park, E. Park, J. Lee, G. Jeong, K. J. Kim, J. H. Kim, Y. J. Kim and H. Kim, ACS Appl. Mater. Interfaces, 2014, 6, 9608–9613.
22. B. Koo, H. Kim, Y. Cho, K. T. Lee, N.-S. Choi and J. Cho, Angew. Chem., Int. Ed., 2012, 51, 8762–8767.
23. J. X. Song, M. J. Zhou, R. Yi, T. Xu, M. L. Gordin, D. H. Tang, Z. X. Yu, M. Regula and D. H. Wang, Adv. Funct. Mater., 2014, 24, 5904–5910.
24. D. E. Yoon, C. Hwang, N. R. Kang, U. Lee, D. Ahn, J. Y. Kim and H. K. Song, ACS Appl. Mater. Interfaces, 2016, 8, 4042–4047.
25. C. Hwang, S. Joo, N. R. Kang, U. Lee, T. H. Kim, Y. Jeon, J. Kim, Y. J. Kim, J. Y. Kim, S. K. Kwak and H. K. Song, Sci. Rep., 2015, 5, 1–10.
26. P. Marszalek, A. Oberhauser, Y. Pang and J. Fernandez, Nature, 1998, 396, 661–664.
27. G. Felton, F. Farleyf and R. M. Hixon, Cereal Chem., 1938, 15, 678.
28. F. F. Farley and R. M. Hixon, Ind. Eng. Chem., 1942, 34, 677–681.
29. J. Song, M. Zhou, R. Yi, T. Xu, M. L. Gordin, D. Tang, Z. Yu, M. Regula and D. Wang, Adv. Funct. Mater., 2014, 24, 5904–5910.
30. J. Liu, Q. Zhang, T. Zhang, J.-T. Li, L. Huang and S.-G. Sun, Adv. Funct. Mater., 2015, 25, 3599–3605.
31. Y. K. Jeong, T.-w. Kwon, I. Lee, T.-S. Kim, A. Coskun and J. W. Choi, Nano Lett., 2014, 14, 864–870.
32. Y. K. Jeong, T. W. Kwon, I. Lee, T. S. Kim, A. Coskun and J. W. Choi, Energy Environ. Sci., 2015, 8, 1224–1230.
33. J. Guo, A. Sun, X. Chen, C. Wang and A. Manivannan, Electrochim. Acta, 2011, 56, 3981–3987.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20560g

---