High deposition rate aSi:H layers from pure SiH 4 and from a 10% dilution of SiH 4 in H 2

 Retail

 76 views
of 4

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Description
High deposition rate aSi:H layers from pure SiH 4 and from a 10% dilution of SiH 4 in H 2
Share
Tags
Transcript
  Ž . Thin Solid Films 373 2000 176  179 High deposition rate a-Si:H layers from pure SiH and from a 4 10% dilution of SiH in H 4 2 M. Estrada a,  , A. Cerdeira a , I. Pereyra b , S. Soto a a  Depto. de Ingenierıa Electrica    S.E.E.S., CINVESTAV, A    . Instituto Politecnico Nacional 2508, 07300 Mexico D.F., Mexico ´ ´ ´ ´ b  Laboratory of Microelectronics, Uni    ersity of Sao Paulo, Sao Paulo, Brazil  Abstract In this paper, we present the results on deposition rates and characterization of a-Si:H layers deposited from pure SiH in 4 Ž . 13.56-MHz plasma enhanced chemical vapor deposition PECVD equipment, where the use of parallel plates of equal area, longgas residence as well as the optimization of process parameters doubled the previously reported deposition rates for this r.f.frequency and gas. A deposition process using a 10% dilution of SiH in H was also optimized to increase the deposition rate to 4 2 1.5   m  h. I  V and I  T curves of PIN diodes up to 18   m thick fabricated on these high deposition rate a-Si:H layers werecharacterized. The density of ionized states at deep depletion were determined and compared with those obtained for diodesfabricated with other standard and high deposition rate methods.    2000 Elsevier Science S.A. All rights reserved.  Keywords:  Plasma processing and deposition; Amorphous materials; Silicon 1. Introduction In the past few years, much effort has been dedi-cated to fabricate thick amorphous PIN diodes that canbe used in applications such as radiation detectors formedical applications. PIN diodes several micrometersthick are already commercially available for the indi-rect detection of medical X-rays, used in conjunction with scintillating layers to transform the radiation into   photons 1 . Direct detection is still an objective to beachieved. Layers tens or hundreds of micrometers thick,stable and with good electrical parameters are re-quired, and so it is necessary to increase the depositionrate of the PECDV films, while maintaining or improv-ing their quality. To achieve this objective, severalapproaches have been used, such as increasing the r.f.frequency, increasing the plasma density or dilutingSiH in other gases. Using an r.f. frequency of 85 MHz, 4  Corresponding author. Ž .  E-mail address:  mestrada@mail.cinvestav.mx M. Estrada . good quality a-Si:H layers with deposition rates above 2  m  h have been obtained by PECVD using pure SiH 4   or mixtures of SiH in He, and in H 2 . At 13.56 4 2 MHz, a deposition rate of 2.88   m  h has been re-   ported for a dilution of SiH in He 3 . For pure SiH 4 4 the deposition rate is usually less than 1   m  h in     standard PECVD equipment 4 . In Will et al. 5 , adeposition rate of 11   m  h was reported using amodified reactor where the substrates were located inthe electrode with the smallest area. However, someauthors recognize that the as-deposited films had aninferior quality when compared to conventionally de-posited material. A deposition rate using a gas mixture of 10% SiH 4 in H that provides deposition rates in the order of 0.9 2  m  h was already reported by us when a specially   designed injector was used 6 . In Estrada and Pereyra   7 we reported a deposition rate of up to 2.5   m  h forpure SiH using a plasma frequency of 13.56 MHz. 4 In this work we characterize films and diodes de-posited in standard PECVD equipment, at 13.56 MHz Ž . from pure SiH process PA1 and 10% SiH in H 4 4 2 0040-6090  00  $ - see front matter    2000 Elsevier Science S.A. All rights reserved. Ž . PII: S 0 0 4 0 - 6 0 9 0 0 0 0 1 1 2 9 - 9  ( ) M. Estrada et al.  Thin Solid Films 373 2000 176  179  177 Ž . process PA2 , when an injector and parallel plates of equal area are used. 2. Experimental details The process parameters that can increase the deposi-tion rate in a PECVD equipment are the power den-sity, the deposition temperature, the pressure and theflow rate, as well as the geometry of the reactor and gasflow mixture. Increasing power density may negativelyinfluence the properties of the deposited layers, due tohigher ion bombardment and  or gas phase polymeriza-tion, which can also occur when increasing the flowrate.Increasing the deposition temperature may increasethe gas reactivity and the rate of the surface reaction,but the concentration of hydrogen atoms inside thea-Si:H layers is also related to the temperature of deposition and a total concentration smaller than 20%must be achieved to obtain good quality with lowdensity of dangling bonds and film stress. As described by Paschen’s Law, for each individualgeometry of the deposition equipment there are tworegimes: the low pressure regime, where the depositionis limited by surface reaction, and the high pressureregime, where deposition is limited by chemicalprocesses in the gaseous phase. In this later case, Ž . formation of Si  H bonding of higher levels SiH , etc. 2 are expected compared to the first regime, where thepredominant Si  H bonding is expected to be SiH. BothPA1 and PA2 processes were adjusted to work in thelow pressure regime.In our experiments the geometry of the reactors wasmodified to provide parallel electrodes of equal area.The use of electrodes of equal area results in a greatersheath voltage for the electrode where the substratesare located, so the deposition rate is increased.Process PA1 was first adjusted in the reactor   schematically shown in Fig. 1 6 , and later in thereactor shown in Fig. 2, which indicates that the re-ported process is repeatable. The layer thickness wasmeasured by profilometry and reflection spectrometry.The presence of Si  H bonds was determined by in- Ž . frared Fourier transform spectrometry FTIR .The flow rate can be increased when diluted SiH is 4 used, but it is limited by the extraction capacity of the vacuum equipment to maintain the condition of thelow pressure regime. The use of a gas injector thatprovides the gas to flow parallel to the surface of thesubstrate where the layer is being deposited was al-ready reported to have great importance for obtaininghigh deposition rates using 10% dilution of SiH in H 4 2   6 . Under these conditions, we could enhance theconcentration of the reactive species near the substratesurface, without increasing gas phase polymerization. Fig. 1. Schematic diagram of reactor No. 1 used for process PA1. Cr deposited on silicon and glass substrates wasused. Before deposition, the substrates were carefullycleaned in trichloroethylene, acetone, a mixture of sulfuric acid and H O , a mixture of H O with 2 2 2 2 NH OH, and a mixture of H O with ClH, followed by 4 2 2 a rinse in deionized water. Cleaning of the substrate was of much importance to increase film quality and toavoid peel off.Pin diodes were fabricated on silicon substrates pre- viously covered with a Cr layer, to define the backelectrode of the diodes. An n  type amorphous layer170 nm thick was deposited from a mixture of 10%SiH in hydrogen and 1% PH in H ; p  layers 300 4 3 2 nm thick were deposited from a mixture of 10% SiH 4 in H , and 1% of B H in H . For the PA1 process, 2 2 6 2 the i-layer was deposited using 20 sccm of pure SiH at 4 Fig. 2. Schematic diagram of reactor No. 3 used for processes PA1and PA2.  ( ) M. Estrada et al.  Thin Solid Films 373 2000 176  179 178 43 Pa, 50 mW  cm 2 and 200  C; for all samples, the p  layer was covered by another layer of Cr, to form thecontact region. Afterwards, a photolithography wasmade, on the Cr layer, the p  layer and part of thei-layer, in order to leave the metal and the p  layeronly in the areas defined for the independent diodes.The etching of part of the i-layer around the diode areais done to prevent leakage current through the borderof the p  i junction. Finally, annealing to form theohmic contacts was done.To characterize PIN radiation detectors, it is impor-tant to determine the voltage,  V   , required to reach fd the condition of full depletion of the intrinsic layer. A  Ž . space charge region SCR across this layer and anelectric field to help recollect the generated carriersare both required for the device to work efficiently.The voltage  V   can be determined by several meth- fd ods. In this work we used the I  V curves where thecurrent, for voltages greater than  V   , depends expo- fd nentially with voltage due to field enhanced emission.For this reason,  V   is determined as the voltage where fd a linear region in a semi-logarithmic graphic of the I  Vcurve appears.To compare with results previously reported for otherhigh deposition methods, we will also use the crys-talline-like approximation where:  q 2 Ž . V       N     d  . 1  fd o i 2   0  i  q  is the electron charge;    ,    are the absolute and 0  i relative dielectric constant, respectively;  d  is the in- i trinsic film thickness; and  N   is a density of ionized  o deep states when electron emission equals hole emis-sion.To characterize the quality of the films we de-termined the position of the Fermi level  E  using the F0 I  V  T curves of the PIN diodes. If a measurement of the steady state current for a given voltage at differenttemperatures is made, the activation energy obtainedfrom these curves should correspond to  E    E  , C Fq  where  E  is the quasi-Fermi level. If, however, the Fq current is measured for each temperature with a timedelay of 1 s after bias, the thermal generation current will exceed its steady state value and the activation   energy will correspond to  E    E  8 . C F0 3. Results and discussions  A deposition rate for a-Si:H layers of 2.5   m  h wasobtained in three different reactors using parallel elec-trodes of equal area, long gas residence and optimizingthe other deposition parameters to 20 sccm of SiH , 43 4 Pa, 50 mW  cm 2 and 200  C.   The deposition rate obtained in Soto et al. 6 was Table 1Values of deposition rate and charged density of states for severaldeposition processesDeposition process Deposition rate  N   o  3 Ž . Ž .  m  h cm 15   Standard SiH 4 1.0 1.28  10 415   SiH in He 3 2.88 1.7  10 415 PA2 2.5   10 15 PA1 1.5   10 increased to 1.5   m  h with the addition of parallelelectrodes of equal area.The gap  E  was determined by transmittance, ob- g taining values within the interval 1.7  1.75 eV, for allfilms using both processes.The presence of SiH bonds was determined throughthe absorption peaks near 640 cm  1 and 2000 cm  1 .The presence of SiH bonds, through the absorption 2 peaks around 850 and 2090 cm  1 . Since the peak at2090 cm  1 is usually difficult to separate from the peakat 2000 cm  1 , an indication of a smaller or greateramount of SiH bonds with respect to SiH bonds is 2 obtained by comparing the intensity   width ratio of thepeaks at 630 cm  1 and near 2000 cm  1 , normalized with respect to the film thickness. The selected deposi-tion regime revealed a small amount of SiH radicals. 2 Diodes 8 and 18   m fabricated with process PA1reached full depletion voltage at 60 V and 220 V,respectively, giving a constant density of charged statesof 10 15 and 8  10 14 . In Table 1, these results arecompared to those previously reported. As can be seen,the two new high deposition methods produce similar values of   N   as previously reported for deposition at 80  o MHz and dilutions of silane in He.Diodes 3   m thick fabricated with process PA2reached full constant density of charged states of   N     o 9  10 14 cm  3 . Ž . Ž . Ž . Fig. 3. Log  I   vs.  V   curves 1 , 2 and 3 for devices 3, 8 and 18 D nom  m thick under reverse bias.  V    V   V   . nom fd  ( ) M. Estrada et al.  Thin Solid Films 373 2000 176  179  179Fig. 4. I vs. 1000  T curves of an 18-  m thick device, for voltages   Ž .    Ž .  between   50 V curve 1 and   250 V curve 5 , in steps of    50V. The I  V curves for diodes 3, 8 and 18   m thick areshown in Fig. 3, where  V    V   V   for each device, is norm fd plotted along the  x -axis. As can be seen, at  V    1, norm the current in the three cases starts to depend expo-nentially with the voltage.In I  V  T experiments, devices were fixed at a giventemperature and I  V curves were measured with atime delay of 1 s. Fig. 4 presents the curves I vs.1000  T for reverse voltages starting from  V    50 V D to  V    250 V with steps of 50 V for a typical device D prepared with process PA2.The slope for these curves in the temperature rangebetween   45 and 45  C and for all the devices ex-amined, remains practically independent of the voltage,giving an average activation energy in the order of 0.65eV, which also points to the fact that the Fermi level inequilibrium for these devices is on the order of 0.2 Vabove the middle of the gap as expected. 4. Conclusions  A deposition rate of 0.9   m  h previously obtainedfor a gas mixture of 10% SiH in H at 300  C, 66.7 Pa, 4 2 r.f. power of 50 mW  cm 2 and flow rates of 150 sccm, when an injector was incorporated to the equipment, was increased to 1.5   m  h by making the parallel,non-symmetric electrodes into electrodes of equal ar-eas.The electrical characterization of thick PIN diodesfabricated with processes PA1 and PA2 demonstratesthat the new deposition processes developed to dupli-cate the deposition rate of the films, provided a con-stant density of charged states similar to the valuesreported for standard deposition methods and for highdeposition rate using SiH diluted in He. The Fermi 4 level was determined to be very close to the expected value at 0.65 eV below the conduction band.  Acknowledgements The authors acknowledge Olga Gallego, M.I. Alayoand C.A. Villacorta for sample preparation. This work was financially supported by CONACYT projects28092A, 3192P-A9607 and FAPESP project   96  7183-6. References   1 R.L. Weisfield, in: 1998 International Electron Device MeetingTechnical Digest, San Francisco, USA, 6  9 December 1998, p.21.   2 W. Hong, A. Mireshghi, J. Drewery, T. Ying, Y. Kitsumo, S. Ž . Kaplan, V. Perez Mendez, IEEE Trans. Nucl. Sci. 42 1995240.   3 T. Pochet, A. Ilie, A. Brambilla, B. Equere, IEEE Trans. Nucl. Ž . Ž . Sci. 41 4 1994 1014.    Ž . Ž . 4 R. Curtins, N. Wyrsch, A.V. Shan, Electron. Lett. 23 5 1987228.   5 S. Will, H. Mell, M. Poschenrieder, W. Fuhsm, in: Proceedingsof the 17th Inter. Conf. on Amorphous and MicrocrystallineSemiconductors, Budapest, Hungary, 25  29 August 1997, p.131.   6 S. Soto, M. Estrada, A. Merkulov, R. Asomoza, Thin Solid Ž . Films 330 1998 83.    Ž . 7 M. Estrada, I. Pereyra, Thin Solid Films 346 1999 255.    Ž . Ž . 8 R.A. Street, Appl. Phys. Lett. 57 13 1990 1334.
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks