Part 3: Nanostructured Cathode Materials for Lithium Ion Batteries 79
6.3 Results
6.3.2 Electrochemical Measurements
that while further reduction in particle size is not occurring as rapidly, the material continues to be better mixed with the carbon matrix, which influences the environment of the Fe atoms.
The XRD patterns and M¨ossbauer spectra of the the ball-milled carbon-FeF3 nanocomposite used in this study are shown in Fig. 6.3(b) and (d) in comparison with pristine FeF3 (a) and (c).
The material milled with carbon for 36 h was selected because the fraction of superparamagnetic iron in the material is dominant over the antiferromagnetic without indication of any additional phase formation in the XRD. The average crystallite size of 18 nm is small enough to benefit from reduced particle size effects commonly observed in nanocomposite cathode materials [99,98].
Bright-field and dark-field TEM was performed on the ball-milled carbon-FeF3 nanocomposite.
Images acquired from the same region of the sample are shown in Fig. 6.4. The bright-field image in Fig. 6.4(a) shows the aggregated FeF3 nanoparticles surrounded by amorphous carbon. The dark-field image in Fig. 6.4(b) was taken using the FeF3 (100) diffraction ring (d = 3.60 ˚A). An average particle size of 15±8 nm was determined by examining several regions of the sample, which is consistent with the value of 18 nm obtained from XRD. In the electron diffraction pattern of the carbon=FeF3 nanocomposite shown in Fig. 6.4(c), the innermost fine ring corresponds to the 10µm SAD aperture used to obtain the diffraction pattern. The remaining diffraction rings visible in Fig. 6.4(d) match well with indexes of FeF3 as a ReO3 cubic structure, consistent with the XRD pattern of Fig. 6.3(b). A number of electron diffraction patterns were taken and impurities from ball-milling such as Fe or Mn metals were not found, nor were any additional elements visible by energy-dispersive X-ray spectroscopy (EDS) analysis of the milled materials.
4
3
2
1
Voltage (V)
250 200 150 100 50 0
Capacity (mAh/g) 4.5V - 1.75V
4
3
2
1
Voltage (V)
500 400 300 200 100 0
Capacity (mAh/g)
4.5V - 1.5V
4
3
2
1
Voltage (V)
800 600
400 200
0
Capacity (mAh/g)
4.5V - 1V 4
3
2
1
Voltage (V)
250 200 150 100 50 0
Capacity (mAh/g)
4.5V - 2V
Figure 6.5: Cycling curves for coin cells cycled at ±142 mA/g between 4.5V and 1.0V, 1.5V, and 1.75V, 2.0V. The first ten cycles are shown, with the initial cycle in blue.
with cycling. A substantial initial capacity is observed during discharge to 1.0 V, reaching over 800 mAh/g on the first cycle, compared to only 220 mAh/g on the first cycle discharged to 2.0 V.
The specific capacity and specific energy density as a function of cycle number are shown in Fig. 6.6. The largest capacity and energy density come from cycling to a lower voltage cutoff of 1.0 V, but this benefit may be offset by the significant capacity fade from over 800 mAh/g to less than 500 mAh/g after just ten cycles. In contrast, cycling to a higher voltage cutoff of 2.0 V yields a much more stable capacity and energy density over 10 cycles, although the capacity is only∼200 mAh/g.
Thus, a second set of tests were undertaken to investigate discharge voltages of 1.0 and 2.0 V over a larger number of cycles. Figure6.7 shows voltage versus capacity curves for selected cycles over 100-cycle tests between 1.0 to 4.5 V, and 2.0 to 4.5 V. The cells undergoing the ‘deep discharge’ to 1.0 V showed a large voltage hysteresis between charge and discharge, indicating very low Coulomb efficiency. Their discharge profiles had large slopes, and their discharge capacity faded quickly during cycling. The cells undergoing the ‘shallow discharge’ to 2.0 V showed a voltage plateau between 3.0 and 2.7 V that undergoes little change up to 20 cycles. Their capacity decreased slowly upon cycling, with the slope of the discharge curve increasing gradually. The voltage hysteresis remained around 0.7V to the 100th cycle. The voltage hysteresis during cycling with a “shallow discharge” was much smaller than for the “deep discharge”, consistent with previous cyclic voltammetry results [109]
which are discussed in the next section.
The performance of the cathode material may be better demonstrated by comparing the relative capacities versus cycle number for the different depths of discharge. In Fig. 6.8(a), the charge and discharge capacities relative to the charge capacity in the third cycle are plotted versus cycle number. Figure6.8(b) shows that even over 10 cycles, there are obvious differences in the capacity fade as a function of minimum discharge voltage. When discharged to 2.0 V, the capacity loss after
1.0
0.8
0.6
0.4
Energy density ratio
10 8
6 4
2
Cycle number 2V
1.75V 1.5V 1V 900
800 700 600 500 400 300 200 100
Specific capacity (mAh/g)
10 8
6 4
2
Cycle number
2V 1.75V 1.5V 1V
Figure 6.6: Specific capacity and energy density shown as a function of cycle number for charge (solid symbols) and discharge (empty symbols) corresponding to the voltage profiles shown in Fig.6.5. Coin cells were cycled at±142 mA/g between 4.5V and 1.0V, 1.5V, and 1.75V, 2.0V.
4.0 3.5 3.0 2.5 2.0
Voltage (V)
100 80
60 40 20 0
Capacity (mAh/g)
3 100 50 20
10 5 4.0
3.0
2.0
1.0
Voltage (V)
240 200 160 120 80 40 0
Capacity (mAh/g) 100 50
20 10
5 3
a.
b.
Figure 6.7: Cycling curves for coin cells cycled at±142 mA/g between (a) 4.5V and 1.0V and (b) 4.5V and 2.0V. Each cell underwent 100 cycles. The 3rd, 5th, 10th, 20th, 50th, and 100th cycles are shown.
100 90 80 70
Percent Relative Capacity
10 9 8 7 6 5 4 3
Cycle Number 4.5-2V
4.5-1.75V 4.5-1.5V
4.5-1.25 4.5-1V
100 80 60 40 Percent Relative Capacity 20
3 4 5 6 7 8
10 Cycle Number2 3 4 5 6 7 8100 4.5-2V
4.5-1V
a.
b.
Figure 6.8: Coulombic efficiency relative to the third cycle versus cycle number for capacities during charge (solid circles) and discharge (empty circles). (a) Capacities in the extended tests shown in Fig.6.7. (b) Capacities from shorter tests.
10 cycles was 4%, whereas discharging to 1.75 V and 1.5 V caused capacity losses of 5% and 6%, respectively. For deeper discharges below 1.5 V, the capacity losses in 10 cycles jumped to greater than 30%. These differences were even more prominent after further cycles, as seen from Fig. 6.8(b).
For shallow cycling between 4.5 and 2.0 V, both charge and discharge capacities were stable for the first ten cycles, decreased approximately 10% between cycles 10 and 20, and reached 100 cycles with over 60% capacity remaining. Deep cycling between 4.5 and 1.0 V gave a fade in capacity to 50%
after 10 cycles, an additional fade to 40% between cycles 10 and 20, and less than 20% capacity remained after 100 cycles.