NCPV and Solar Program Review Meeting 2003

NREL/CD-520-33586 Page 866

DLTS Characterization of CIGS Cells

L.L. Kerr, S. S. Li*, T.J. Anderson and O.D. Crisalle Dept. of Chemical Engineering, University of Florida, Gainesville, FL, 32611 *Dept. of Electrical and Computer Engineering, University of Florida, Gainesville, FL, 32611 S. Johnston, Jehad Abushama and R. Noufi, National Renewable Energy Lab, Golden, CO, 80401 ABSTRACT EXPERIMENTAL PROCEDURE Deep Level Transient Spectroscopy (DLTS) and Capacitance-Voltage (C-V) measurements are employed to study deep- level electron and hole traps in CIGS solar cells fabricated at two different locations (EPV and NREL). The activation energy and trap density as well as suggested defect origins are given. DLTS is used to measure the transient capacitance change of a junction device after deep- level traps in the space charge region have been filled with either the majority or minority carrier charges, and the trap density and energy levels of these deep level traps can be extracted from DLTS measurements. By scanning the transient capacitance over a range of temperature, the emission rate as a function of INTRODUCTION temperature can be obtained. The activation energy for an It is well known that CuInGaSe2 (CIGS) has a large electron trap can be estimated from an Arrhenius plot: homogeneity range, and such deviation from stoichiometry in Ea ) (1) compound semiconductors is often attributed to antisite en = n v th N c exp( kT defects, vacancies and defect clusters in the material. Zhang [1] has calculated the transition energy for a large number of where en is the electron emission rate, n is the capture defects in CIS, and the results are summarized in Figure 1. cross section, and th is the electron thermal velocity. The Knowledge of the defect energies as well as their densities are electron trap density, NT, can be calculated by important inputs to developing meaningful device models. DLTS is a standard diagnostic technique for determining the 2 C Na (2) trap properties such as trap energy level, capture cross N T = C0 section, and trap concentration. Equation (2) is valid for NT < Na. C is the height of capacitance change due to the emission of charge carriers from the trap level, which is proportional to the probed trap concentration. Carrier concentration, Na, is determined by the C-V meaurements, which give the average carrier VSe(2+/0) concentration in the CIGS absorber layer. C0 is the zero bias capacitance that is obtained from the C-T data at the corresponding DLTS peak temperature. RESULTS AND DISCUSSION 1. EPV CIGS sample

The DLTS spectra measured from the EPV sample with a reverse bias voltage, VR =-0.1V, a trap-filling pulse of 0.3 V, and a saturation pulse width of 10 ms is shown in Fig. 2. The EPV sample showed a deep majority carrier trap (negative signal) at a temperature around 270 K. The DLTS peak shifts to higher temperatures with shorter rate windows. Since the CIGS film is p-type material, the majority carrier trap is referred to as a hole trap, and the minority carrier trap as an electron trap.

Fig. 1. Defect Transition Energy Diagram of -CuInSe2 [1]

NCPV and Solar Program Review Meeting 2003

NREL/CD-520-33586 Page 867

1.00E+00 5.00E- 01 0.00E+00 0.2 ms

Capacitance (F)

450x10

-12

1.4x10

-3

400
350
300
250
200
150
0.0

0.1ms

Conductance (S)

EPV_77K_C EPV_150K_C EPV_200K_C EPV_250K_C EPV_300K_C

1.2

1.0

EPV_77K_G EPV_150K_G EPV_200K_G EPV_250K_G EPV_300K_G

0.8
0.6
0.4
0.2

0.05ms 0.00E+00 5.00E+01 1.00E+02 1.50E+02 2.00E+02 2.50E+02 3.00E+02 3.50E+02 - 5.00E- 01 0.02 ms -1.00E+00 5 ms 2 ms 1 ms 0.5 ms 100 ms 50 ms -3.00E+00 -3.50E+00 -4.00E+00 20 ms 10 ms

C (PF)

-1.50E+00 -2.00E+00 -2.50E+00

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Reverse Bias (V)

50x10
18
4
3

Reverse Bias (V)

EPV_77K_N EPV_150K_N EPV_200K_N EPV_250K_N EPV_300K_N

EP V CIGS Reverse Bias=0.1 V P ulse Height = 0.3 V P ulse Width = 10 ms

40
N (cm^-3)
16

1/C^2

30

10

9 8 7
6

20

10

EPV_77K_C2 EPV_150K_C2 EPV_200K_C2 EPV_250K_C2 EPV_300K_C2

0.0
0.2

5
4

Temperature (K)

0.4

0.6

0.8

1.0

0.2

0.3

0.4
Depletion Width (um)

0.5

0.6

Reverse Bias (V)

Fig. 2. DLTS scan for EPV cell at a reverse bias of 0.1 V, a pulse height of 0.3 V, and pulse width of 10 ms. The activation energy calculated from the Arrhenius plot is 0.94 eV, which may be attributed to the Se vacancy according to the diagram shown in Figure 1.

Fig. 4. C-V data of EPV sample at different temperatures. From the C-T scan (Fig. 5), C0 in Eq. (2) was found to be 15 -3 13 -3 3.0110 cm , and NT was determined to be 6.510 cm .
4.0E+02 3.5E+02 3.0E+02 2.5E+02 C (PF) 2.0E+02 1.5E+02 1.0E+02 5.0E+01 0.0E+00

0
-2

0.002

0.004

0.006

-4

ln(T2)

-6

Ea=0.94 eV EPV CIGS Reverse Bias=0.1 V Pulse Height = 0.3 V Pulse Width = 10 ms 1/T (K-1)
2

EPV CIGS Reverse B ias=0.1 V P ulse Height = 0.3 V P ulse Width = 10 ms

-8

0.E+00 1.E+02 2.E+02 3.E+02 4.E+02 Temperature (K)

-1 0

Fig. 5. The C-T scan for EPV cell at a reversed bias -0.1V. 2. NREL CIGS sample

-1 2

A pulse amplitude equal to 0.4V was applied to the NREL CIGS sample at a reverse set-off voltage of -0.5 V. A minority carrier trap peak was observed at a lower temperature of Figure 4 shows the C-V data measured on the same EPV cell around 100 K as shown in Figure 6. The activation energy is as that used in the DLTS measurement. The average carrier 0.067 eV below the conduction band as shown in Figure 7. A 15 concentration at 270 K from the C-V measurements is 310 possible physical origin of this electron trap is the Se vacancy -3 13 -3 cm . (Vse).[1] The trap density is estimated to be 4.210 cm . The DLTS scan at higher temperature showed another possible positive peak, however, the activation energy for this trap could not be resolved by heating sample above 300 K. Fig. 3. The Arrhenius plot of emission rate en/T vs. 1/T.

NCPV and Solar Program Review Meeting 2003

NREL/CD-520-33586 Page 868

3.0E+00

1.2E+00 1.0E+00 8.0E-01 6.0E-01 4.0E-01

E2
NREL CIGS Reverse B ias=0.5 V P ulse Height = 0.4 V P ulse Width = 10 ms
0.2 ms 0.1 ms 0.05ms 0.02 ms 5 ms

2.5E+00

E1

2.0E+00 1.5E+00

NREL CIGS Applying 0.1V forward bias Cool down to 77 K, t hen heat up t o 350 K Reverse Bias=0.5 V Pulse Height = 0.7 V Pulse Widt h = 10 ms

C (PF)

2.0E-01 0.0E+00 0.E+00 -2.0E-01 -4.0E-01 -6.0E-01

1 ms 0.05 ms

C (PF)

2 ms

1.0E+00
0.2 ms

1.E+02

2.E+02

3.E+02

100 ms 50 ms 20 ms 10 ms

4.E+02

5.0E-01

0.1 ms 0.05 ms 0.02 ms

E3

0.0E+00 0.E+00 -5.0E-01

1.E+02

2.E+02

3.E+02

4.E+02

Temperature (K)

Temperature (K)

Fig. 6. DLTS data for NREL sample measured at a reverse bias -0.5 V, pulse amplitude 0.4 V, and pulse width 10 ms.
0.5 0 -0.5 0 0.005 0.01 0.015

Fig. 8. DLTS scan for the NREL sample under forward bias (0.1 V) injection, and heat up with reverse bias of 0.5V, pulse height of 0.7V, and pulse width of 10 ms.

1
Ea=0.07 eV

ln(T2)

-1 -1.5 -2 -2.5
NREL CIGS Reverse Bias=0.5 V Pulse Height = 0.4 V Pulse Width = 10 ms

0.5 0 C (PF) 0.0115 -0.5 -1 -1.5 -2 1/T (K-1) 0.012 E1=0.14 eV 0.0125 0.013 0.0135 E =0. 1 4 eV
1

1/T (K-1)

Fig. 7. Arrhenius plot of the emission rate en/T vs. 1/T. Therefore, it is noted that the minority traps are likely the dominant traps in the NREL sample. To detect other possible minority traps, a small forward bias (0.1 V) was applied and the sample cooled to 77 K to freeze the minority carriers. The sample was then heated with VR =0.5 V, trap-filling pulse amplitude of 0.7 V, and a saturation pulse width of 10ms to observe the capacitance change due to the minority carrier emission. Figure 8 shows three minority carrier traps with the addition of a small peak around 250 K, which was not apparent in the DLTS spectra of the 0.4 V pulse amplitude. The Arrhenius plot in Figure 9 showed that the activation energy for the trap at 80 K is 0.14 eV. The trap density is estimated to be 13 -3 2.5410 cm . The electron trap observed at 80 K is possibly related to a vacancy on a Se site, which might be the same trap level as the one observed with the 0.4 V pulse amplitude as shown in Figure 6. The activation energy at 335 K could not be estimated since the peak shift does not follow the trend; the shorter the rate window, the higher the temperature at which the peak appears.

Fig.9. Arrhenius plot of emission rate en/T vs. 1/T for the electron trap E1.

CONCLUSIONS The DLTS and C-V measurements have been performed on both EPV and NREL CIGS samples, and defect activation energy and density were determined from these measurements. The EPV sample showed a majority carrier (hole) trap with activation energy and trap density determined. The NREL CIGS sample showed three minority carrier traps. Possible defect origins were depicted from this DLTS study. A summary of the DLTS and C-V results is given in Table 1.

NCPV and Solar Program Review Meeting 2003

NREL/CD-520-33586 Page 869

Table 1. Summary of DLTS and C-V data for the EPV and NREL samples. EPV Approximate Peak Temperature Activation Energy (eV) DLTS peak Trap Carrier Type Na (cm ) NT (cm ) Possible defect I.D. Capture Cross Section 2 (cm )
-3 -3

NREL 100K Ec-0.067 Positive Minority 2.2510

13 15

270 K Ev+0.94 Negative Majority 310

15

250 K

335 K

Positive Minority

Positive Minority

6.510 ED2 VSe

4.210 ED2 VSe 610

13

-18

ACKNOWLEDGEMENTS This work Partnership subcontract study were program. is funded by the DOE/NREL Thin Film PV Program and administrated by NREL under No. XAF-8-17619-32. The samples used in this delivered by the NREL thin film partnership

REFERENCES [1] S.B. Zhang, S-H Wei, and A. Zunger, Phys. Rev. B, Vol 57, No. 16 pp 9642-9668 (1998).