6-3.1 Leakage current suppressed by oxygen plasma treatment
Figure 6-1 gives the leakage current-voltage relationship of the as-deposited BST films. The leakage current under a reverse bias is markedly higher than that under a positive one. In general, Schottky emission (SE) and Poole-Frenke (PF) conduction can be considered to interpret the leakage current of BST films. Many reports indicated the asymmetrical leakage current profiles of the BST capacitors biased at positive/negative voltage are caused by SE contact at Pt/BST interface of single side [154]. Chapter 4 also shows the similar results. The barrier height of SE (interface limited) is dependent on the work function of the electrode metal, electron affinity of the dielectric, and the surface states. The PF transport mechanism (bulk limited) is a result of the lowering of the barrier height of traps in the dielectrics. Hence, the PF transport mechanism does not exhibit polarity dependence [142 – 144, 150, 152]. Figure 6-1 shows the asymmetrical current-voltage profile of the capacitor
biased at positive/negative voltage.
Since both of the top and bottom electrodes are made from Pt materials, this asymmetrical current-voltage profile shows the Schottky barrier is formed at the interface of the BST/bottom-electrode. In addition, the interface states in the upper electrode
Figure 6-1 The typical curve of the leakage current versus the electric field for the as-deposited BST films.
-500 -400 -300 -200 -100 0 100 200 300 400 500
result in high leakage current at a negative bias [149, 151, 155]. The interface defects are caused by high concentration of oxygen vacancies at top surface, which will be discussed later. Obviously, the BST film must be fabricated at a substrate surface temperature of 300oC above (410oC at wafer backside), and exposed to a plasma environment with high concentration Ar+ during sputter process. Hence, the oxygen atoms escape from top surface, and then oxygen vacancies are subsequently generated according to the following equation:
O0 ↔ V0++
+ 2e- + 1/2O2, where O0, V0++
and e− denote the oxygen ion at its normal site, an oxygen vacancy and an electron, respectively. A high concentration of oxygen vacancies led to n-type conductivity of the BST materials due to the electrons generated, causing large leakage currents. The oxygen vacancies can
be greatly improved by an oxygen plasma treatment, which gives rise to active oxygen atoms, decreasing the oxygen vacancies. Figure 6-2 shows the leakage current, measured at various negative electric fields, is a function of the duration time of O2
plasma treatment. The leakage current is significantly reduced for the samples post O2 plasma treated with 3, 5 and 10 min, and,
Figure 6-2 Dependence of the the leakage current measured at various electric fields with a delay time of 30 s on duration of O2 plasma treatment after BST film deposition.
10-7
particularly, the 5 min treatment can reach a minimum value of leakage current. Besides, if the plasma treatment time is longer than 10 min, the leakage current will gradually increase. The substrate in oxygen plasma equipment is DC-biased. A DC-biased substrate will induce serious plasma bombardment damage. The oxygen plasma can effectively passivate the oxygen vacancies on the top surface of the BST film, but the treatment duration must be carefully controlled. If the oxygen plasma treatment with long duration is applied, the plasma bombardment damage will induce even more defects in the BST films.
6-3.2 Material and electrical characterizations of BST films post treated by oxygen plasma
The improvement of the leakage current may be attributed to several mechanisms, such as the improvement of surface roughness [155], variation of grain size [156, 157] or compensation of oxygen vacancies near the surface [142 – 144, 149, 152]. Detailed measurements were taken to clarify the reason why the leakage characteristics were enhanced. The surface morphologies of the as-deposited samples and samples O2-plasma treated for 5 min samples were measured using the AFM, as shown in Fig. 6-3. The root-mean square (rms) roughness of the as-deposited and the 5-min O2-plasma-treated samples were 7.581 nm and 7.584 nm, respectively. According to this figure, the surface roughness of the as-deposited BST films is not obviously different from that of samples treated by O2 plasma for 5 min. Figure 6-4 shows the XRD patterns of the BST films post treated by oxygen plasma at various conditions. All of the samples, as-deposited and oxygen plasma treated, do not show obvious differences in XRD results, so the oxygen
plasma treatment does not change the BST crystalline structure. The preferential crystal orientation of each BST film, as confirmed by XRD, was (110). According to the Scherrer formula, the average grain size (<100nm) can be estimated by
D=0.9λ/(B*cosθ),
where D is the grain size, λ is the X-ray wavelength (~0.15428nm), B is the full-width at half-maximum (FWHM) of the XRD peak, and θ is the diffraction angle [153]. The FWHMs of all samples analyzed by XRD were identical, about 0.01725. Hence, the estimated grain size of each sample are almost the same, about 8.3 nm, as shown in Fig.
6-5. Those results of XRD analysis and grain size estimation are consistent with dielectric measurements. The dielectric constants of BST films do not shows greatly changes after post treatment using oxygen plasma, as indicated in Fig. 6-5.
On the other hand, the Auger electron spectroscopy (AES) profiles reveal the oxygen concentration distribution versus sputtering time, as depicted in Fig. 6-6. The
Figure 6-3 The surface roughness of (a) as-deposited BST thin films and (b) 5-min O2-plasma-treated BST film.
(a) (b)
sputtering time is proportional to the BST film depth, the direction of which is perpendicular to the top surface. The AES results show the samples post treated by oxygen plasma exhibit great oxygen passivation on BST surface.
Obviously, the oxygen concentration at the top surface can be gradually improved by oxygen plasma treatment as the duration time increasing. Therefore, the improvement of leakage current at negative bias is primarily attributed to the compensation of surface oxygen vacancies.
Moreover, the leakage currents in the high-electric-field region (0.3, -0.3 and 0.5 MV/cm) were also improved, and the PF transport mechanism dominated the leakage mechanism in such a region [142 – 144, 150, 151]. Therefore, the oxygen plasma treatment can compensate the oxygen vacancies not only on the surface but also in the bulk. The oxygen radicals in oxygen plasma are very active that they are able to oxidize BST films and compensate oxygen vacancies. However, the leakage current becomes degrading for the samples treated with long duration (> 5 min), and this degradation mechanism has been discussed above. Therefore, careful control of the duration time of the oxygen plasma treatment can obtain an optimal condition of the leakage current suppression.
20 30 40 50
Figure 6-4 X-ray diffraction patterns of BST thin films with O2 plasma treatment for different durations after BST film deposition.
0
Figure 6-7 shows the TDDB characteristic of BST films without and with 5 min of oxygen plasma treatment. The lifetime extrapolation using constant-voltage stress gives a lifetime of ten years. For the 5-min-plasma treated samples, lifetime longer than 10 years in a stress field of 2 MV/cm can be obtained. In contrasting, the as-deposited samples exhibit very poor TDDB characteristics, and the 10 years breakdown field is as low as 0.6 MV/cm. In addition, the leakage current of the as-deposited Pt/BST/Pt capacitor is higher than that of 5 min oxygen-plasma-treated samples, reflecting that the increased leakage current accelerates the electrical degradation of the capacitors. Hence, the properly controlled treatment using oxygen-plasma can effectively improve the lifetime characteristics of the BST capacitors.
Figure 6-5 The dielectric constant and the estimated grain size versus the duration of O2 plasma treatment on the BST films.
Figure 6-6 AES profiles of the BST films post treated by oxygen plasma under various treatment durations.
0 400 800 1200 160
6-4 Summary
BST thin films fabricated by the RF sputtering technique or post surface annealing exhibits serious oxygen deficiency on top surface, and therefore the leakage current is severely degraded. Applying oxygen plasma treatment to the BST films can effectively passivate the oxygen vacancies at low substrate temperatures. The plasma technique significantly improves the leakage currents by reducing the concentration of oxygen vacancies in the as-deposited BST films. Excellent electrical characteristics, including a low leakage current (1.5x10-8 A/cm2) under 0.1 MV/cm, high dielectric constant (288), and lifetime longer than 10 years under 2 MV/cm can be achieved. The suitable duration of the oxygen plasma treatment can effectively improve the leakage current, but a very extended duration will damage the BST film surface. Consequently, the proper and carefully controlled oxygen plasma treatment greatly suppresses the leakage current at
Figure 6-7 TDDB characteristics of the as-deposited/oxygen-plasma-treated BST films.
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.
100 102 104 106 108 1010
As-deposited
O2 plasma 5 min 10 years
Time to Breakdown (s)
Stress Electric Field (MV/cm)
low temperature, and therefore this post treatment could be one of the most promising technologies for the integration of IC process.