Effective energy storage from a triboelectric nanogenerator

To sustainably power electronics by harvesting mechanical energy using nanogenerators, energy storage is essential to supply a regulated and stable electric output, which is traditionally realized by a direct connection between the two components through a rectifier. However, this may lead to low energy-storage efficiency. Here, we rationally design a charging cycle to maximize energy-storage efficiency by modulating the charge flow in the system, which is demonstrated on a triboelectric nanogenerator by adding a motion-triggered switch. Both theoretical and experimental comparisons show that the designed charging cycle can enhance the charging rate, improve the maximum energy-storage efficiency by up to 50% and promote the saturation voltage by at least a factor of two. This represents a progress to effectively store the energy harvested by nanogenerators with the aim to utilize ambient mechanical energy to drive portable/wearable/implantable electronics.

The maximum E C for the direct charging cycle E C,designed,max The maximum E C for the designed charging cycle Q C The amount of charge flowing to the energy storage unit (proportional to the average charging rate) Q C,direct Q C for the direct charging cycle Q C,designed Q C for the designed charging cycle η The energy storage efficiency, as the stored energy per cycle over E m η direct The maximum η of the direct charging cycle η designed The maximum η of the designed charging cycle V Sat The saturation voltage (the largest possible charging voltage) V Sat,direct V Sat of the direct charging cycle V Sat,designed V Sat of the designed charging cycle

Supplementary Note 1: The equations for boundaries of CMEO with infinite load R:
As demonstrated previously 1 , Q = 0 and Q = Q SC,max are the two boundaries which are parallel to V -axis. The other two boundaries are: The equations (1) and (2) can be used to calculate coordinates of important points in the charging cycles.
Supplementary Note 2: The change of V C during one charging cycle: For the batteries, we can use the region close to the plateau voltage, then V C will be nearly constant during charging process. For capacitors, most of load capacitors used for charging have a large capacitance C L . In fact as reported, to achieve efficient charging process, the optimum value of C L should be much larger than the largest value of C(x) 2 . The capacitance between two electrodes of TENG can be written as a variable C(x) related to the displacement x, and it is usually very small. Therefore: Here, ΔQ and ΔV C represent the amount of charges flowing to the capacitor and the variation of V C in the capacitor during a half charging cycle. Then we can deduct that min{V OC,max , V' max } >> ΔV C , which means we can assume that charging voltage V C does not change significantly during one charging cycle.

Supplementary Note 3: The process of TENG charging the capacitor/battery:
During the charging process of a capacitor, the electrons from one electrode of the TENG are driven to the negative electrode of the capacitor. Consequently, the additional positive charge is induced in the positive electrode by the electrostatic induction, and the additional electrons are released from the positive electrode in the capacitor, and driven to the other electrode of the TENG.
During the charging process of one fabricated lithium-ion battery or several batteries in series, the electrons from one electrode of the TENG are driven to the cathode of the first battery, to facilitate the following reaction in the cathode: With the consumption of the lithium ions, the ones in the anode immigrate to the cathode to balance the concentration. Therefore, the following reaction in the anode is promoted to release more lithium ions: . From status V to VI, all of the rectifiers are off, then Q is kept and status VI is Similarly, for the designed charging cycle, statuses I to III are calculated as (0, 0), (0, V C ) and (Q SC,max (1 -V C /V OC,max ), V C ), respectively. In status IV which still satisfies Supplementary Equation (1), the switch is on to make V = 0. Then status IV is calculated as (Q SC,max , 0). The status V is given by Q = Q SC,max since all of the rectifiers are off. So status V is (Q SC,max , -V C ). The status VI which is with x = 0 is as same as status V in the direct charging cycle.

Supplementary Note 5: The decrease of Q C as increase of V C in the charging cycles:
The calculation method of coordinates has been stated in Supplementary Note 4. Then we can extract Q C in both cases as: As we can see, both Q C decrease with increase of V C . The reason is discussed below: When x is approaching x max from 0, a certain charge of SC,max C OC,max / Q V V should stay in the 1 st electrode (the electrode overlapped by the dielectric when x = 0 in SFT mode) to the maintained voltage of V C ; similarly, when x is approaching 0 from x max , a certain charge of SC,max C max / ' Q V V should stay in the 2 nd electrode (the electrode overlapped by the dielectric when x = x max in SFT mode) due to the maintained voltage of -V C . And these amounts of charges are both proportional to V C . Even through in the designed charging cycle, the switch is turned on to make charges fully transferred, these additionally transferred charges are through the switch other than the energy storage unit.
Thus, at increased V C , as limited by these parts of charges, Q C for both charging cycles decrease.
The difference between Q C for both charging cycles is: So in the designed charging cycle, Q C will decrease slower than the direct charging cycle.
The fundamental reason of that is stated as below: During switch-on operation from statuses III to IV, the charges were fully transferred to Q SC,max , so that in the next half-cycle, the charges available for flowing to the energy storage unit change from SC,max statues IV and V in the direct charging cycle) to SC,max between statues V and VI in the designed charging cycle); Similarly, during switch-on operation from statuses VI to I, the charges were fully transferred to 0, so that in the next half-cycle, the charges available for flowing to the energy storage unit change from SC,max between statues III and VI in the direct charging cycle) to SC,max (differences in Q between statues III and II in the designed charging cycle);

Supplementary Note 6: Q C for both the direct and designed charging cycles:
Theoretically, only when the voltage V between the electrodes of TENG achieves the charging voltage V C , the charging process of the batteries/capacitors can proceed. In both the shapes of the direct and the designed charging cycles, the sides parallel to the Q axis means V is kept at a constant voltage as ± V C , therefore, Q C is the total length of these sides. To double confirm that, we measured Q C directly by connecting an Ammeter in series with the batteries/capacitor in the circuit for the direct charging cycles, as shown in If V C ≥ V OC,max (we can assume V OC,max ≥ V' max as we stated in Supplementary Note 7), the rectifiers cannot be opened, then no energy harvesting in both the direct and designed cycles; (Please check Equation (4) (2)): From status III to IV (except right at status III or IV), |V| < V C , so the rectifiers cannot be opened again. And in following cycles, TENG oscillates between status III and status IV.
Consequently, |V| < V C is always valid for 0 < x < x max , rectifiers cannot be opened, and there is no energy stored. Only when x = 0 or x max , |V| might equal to V C , but since the displacement x cannot change further (as fixed in the range of 0 to x max ), no charge can be pushed into the energy storage unit. In summary, in this case, the energy can only be stored in the 1 st half cycle, and there is no energy storage during steady-state of the direct charging cycle. So in the continuous direct charging cycle for a capacitor, when V C is approaching V Sat,direct , less and less energy is stored in the capacitor; and V C > V Sat,direct is not accessible since there is no energy stored once V C = V Sat,direct .
For the designed charging cycle, as shown in Supplementary Figure 7b, the process is still cycled from status I to VI, as same as that in the manuscript. The reason is that after statuses III and VI, the switch is turned on to make V = 0 (statuses IV and I), so that |V| can achieve |V C | much easier in the following steps when all of the rectifiers are all off (status IV to V and status I to II). However, in this case the encircled area only contains area 2 which can only be stored by the designed charging cycle, and area 3 which is released through the switch. The area 1 that can be stored by both the direct and designed charging cycles as described in Figure 3 is missing due to the reason as stated above. So in this designed charging cycle, the charging process can continue after V C surpass V Sat,direct , until V C is approaching max{V OC,max , V' max }.