Geri Dön

Amorf silisyumda staebler wronski etkisinin fotoiletkenlik parametresi üzerindeki rolü

The Role of staebler-wronski effect on the prameter of photocondictivity in amorphus silicon

  1. Tez No: 46397
  2. Yazar: SEVİLAY UĞUR
  3. Danışmanlar: DOÇ.DR. FATMA TEPEHAN
  4. Tez Türü: Doktora
  5. Konular: Fizik ve Fizik Mühendisliği, Physics and Physics Engineering
  6. Anahtar Kelimeler: Amorf silisyum, Fotoiletkenlik, Amorphous silicon, Photoconductivity
  7. Yıl: 1995
  8. Dil: Türkçe
  9. Üniversite: İstanbul Teknik Üniversitesi
  10. Enstitü: Fen Bilimleri Enstitüsü
  11. Ana Bilim Dalı: Belirtilmemiş.
  12. Bilim Dalı: Belirtilmemiş.
  13. Sayfa Sayısı: Belirtilmemiş.

Özet

ÖZET Amorf Silisyum (a-Si) ışığa çok duyarlı bir yarıiletken olarak, ışığın elektriğe çevrildiği, güneş pilleri, optik sensörler, fotokopi makinalarının tamburları gibi, pekçok yüksek teknoloji ürününde kullanılır. Ancak ışığa maruz kalan a-Si numunenin elektriksel özellikleri değişir. Staebler-Wronski etkisi adı verilen bu olay a-Si ile üretilen cihazların verimim etkileyen en önemli parametredir. Amorf silisyumda kristallerdeki düzen yoktur, ve bu malzemenin en küçük bir dış etkenle özellikleri değişir. Bu nedenle, başta Staebler-Wronski etkisi olmak üzere a-Si da görülen pekçok olayı açıklamak için, cihaz kalitesinde numuneler gerekir. TÜBİTAK, Marmara Araştırma Merkezi, Ulusal Metroloji Enstitüsü Amorf Silisyum Laboratuvannda yapılan çalışmalardan bir tanesi de a-Si numunelerde Staebler-Wronski etkisinin incelenmesidir. Amorf silisyum laboratuvanndaki gözlemlere göre, a-Si üzerine düşen ışık şiddetinin numuneden geçen akıma etkisini belirleyen 7 parametresi ile Staebler-Wronski etkisi arasmda bir ilişki bulunmaktadır. ^kF1 ilişkisinde oluşan akımı (7), uygulanan ışık şiddetine (F), bir sabitle (£), bağlayan 7 nın tam olarak çalışılabilmesi ve bu konuda yapılan teorik modellerle karşılaştırılabilmesi için yeni bir ölçüm tekniğinin geliştirilmesi gerekmektedir. Staebler-Wronski etkisiyle 7 arasındaki ilişkiyi inceleyen ve bu teze de konu olan araştırma kapsamında temel dört çalışma yapılmıştır: 1. a-Si numunelerin üretimi ve bu numunelerin cihaz kalitesinde olduklarının gösterilmesi, 2. Foton enerjisine ve ışık şiddetine bağlı olarak, 7 nın Staebler-Wronski etkisiyle nasıl değiştiğinin incelenmesi, 3. Yeni bir ölçüm tekniğinin geliştirilerek fotoakımın, ışık şiddetinin 1010 defadan daha geniş ışık şiddeti aralığında ölçülebilmesi, 4. Analiz edilen sonuçların yüksek kalitede amorf silisyum numunede 7 nın nasıl bir davranış göstermesi gerektiğini vurgulayan teori ile karşılaştırılması. Tamamlanan bu çalışmalarla yapılan numunelerin cihaz kalitesinde olduğu gösterilmiş, ayrıca dünyada ilk defa amorf silisyumda 14 basamak değişen ışık şiddetinin fonksiyonu olarak, fotoakım ve 7 nm ölçülmesine olanak sağlayan bir deney tekniği gehştirilmiştir. Alman verilerin analizi sonucunda Staebler- Wronski etkisi ile 7 nın ilişkisi belirlenmiş, ve sonuçların 1990 yılında Kocka et al. tarafından geliştirilen ancak deney tekniklerinin yetersizliğinden bugüne dek denenemeyen teoriye mükemmel bir şekilde uyduğu gösterilmiştir. viii

Özet (Çeviri)

SUMMARY THE ROLE OF STAEBLER-WRONSKI EFFECT ON THE 7- PARAMETER OF PHOTOCONDUCTIVITY IN AMORPHOUS SILICON There is a large interest in the optical and electrical properties of amorphous silicon (a-Si:H) produced by the discharge decomposition of silane [34,40,50]. Such films are deposited on the substrates at temperatures between 200 °C and 300 °C and have been extensively used for fabricating high quality junctions and efficient solar cell structures [6,7,35]. Traditional solid-state physics has emphasized the study of periodic systems, whose energy levels may be labeled by the wavevector. The concept of reciprocal space, or of allowed wavevector values, therefore provides a convenient and consistent framework for describing electronic, vibrational and magnetic properties of crystalline solids in terms of energy bands. The semiconductor properties of crystalline solids are succesfuly predicted by the band theory in terms of conduction and valance energy bands, Ec, and Ev. These bands correspond to energies where the densities of states that electrons and holes can occupy are high and the charge carriers in these extended or non-localized states can move freely through the material. The density of states between Ec and Ev on the other hand is zero so that no electrons or holes can exist between these energies. This results in a forbidden gap of energies which correspond to the classic semiconductor bandgap [1]. Furthermore, the periodic nature of the crystalline materials allows the charge carriers to have large mean free paths between collisions with an imperfection, which results in high mobilities of electrons and holes. In a nonperiodic system, the concept of a wavevector is not valid, therefore a fundamental concept that might be useful in the study of nonperiodic systems is needed. Like crystalline solids, amorphous materials also have extended states through which electrons and holes can move freely. However, the absence of long range order introduces many imperfections so that the carrier mean free paths become significiantly lower than in the crystalline material. When these energy dependent mean free paths become comparable to the interatomic distances, the charge carriers become localized. Consequently, amorphous semiconductors have boundaries between the non-localized and the localized states with the resulting energy bands which look similar to those in crystalline semiconductors. Ec and Ev correspond to these boundaries and are called the mobility edges [30]. The energy between Ec an Ev constitutes the mobility gap. In amorphous semiconductors mobility drops by more than two orders of ixmagnitude at the mobility edges [14]. In the crystalline case, there are no energy states available within the energy gap whereas in amorphous materials there is a continium of localized states. The states that are inherent to disorder have densities which decrease exponentially with energy from the mobility edges and rapidly decay to low values (<1016 cm'3eV“x) [20]. These tail states constitute traps into which the charge carriers fall as they move through the material. In the case of amorphous silicon, very large densities of defects exist primarily due to broken bonds between Si atoms. As a result, amorphous silicon makes a poor semiconductor. However it was found that amorphous silicon prepared by glow discharge or plasma-enhanced chemical vapor deposition, possessed a very low density of localized states when the deposited silicon was obtained from silane gas (SiH4) [4,5]. Subsequent work revealed that these films contained approximately 10 percent hydrogen, and therefore the films were actually alloys of amorphous silicon and hydrogen (hydrogenated amorphous silicon, a-Si:H). The low density of localized states was attributed in part to the saturation of dangling bonds by hydrogen. It is these low density of localized states near the middle of the gap of a-Si:H, as well as the similarity of its chemistry to that of crystalline silicon, that makes amorphous silicon such a good amorphous semiconductor. In an amorphous solid, the long range order is destroyed whereas the short range order, i.e. the interatomic distance and the valance angle is only slightly changed. The concept of density of states (DOS) is also applicable to non crystalline solids. Several models were proposed for the band structure of amorphous semiconductors. These are well known Cohen-Fritzche-Ovshinsky (CFO) [18], and Davis-Mott [14] models. These models have been extensively used to explain optical and electrical properties of these materials. CFO-model assumes that tail states extend across the gap in a structureless distribution. This gradual decrease of localized states destroys the sharpness of conduction and valance band edges. In Davis-Mott model, the tails of localized states which are extending into the forbidden gap should be rather narrow. They also proposed compansated levels near the middle of the gap, originating from defects in the random network. The conductivity for a-Si:H is described quite well by the relation [33] o= ojexptAEo/lcT) + oj exp[-(A£i + W1)lkT)]+ a2 exp[-(2afl+ W2)/kT\ (1) The first term corresponds to conduction by carriers excited into the extended states above Ec or below Ev respectively. The second term dominates at the lower temperatures approximetely above 150 °K and is due to phonon assisted hopping of conduction of carriers among band tail localized states. Each time an electron moves from one localized state to another, it will exchange energywith a phonon. If in this regime overlapping of wavefunctions are small, then charge carriers move through the material by multiple trapping [63,64], that is an electron spends some time in the shallow trap levels while moving in the conduction band. The third term is due to variable range hopping of carriers among localized gap states near EF. As the energy is lowered the number and the energy of the phonons decrease, so phonon assisted hopping becomes less favorable. Carriers will tend to hop larger distances in order to find sites which lie energetically closer than the nearest neighbors. The energy dependent optical absorption ct(hv) of amorphous semiconductors consists typically of three parts [41] i. A high absorption region (a ^ 104 cm”1) where the absorption coefficient is described by a=B2(hp-Eg)2/hp (2) This equation can be interpreted as resulting from optical transitions between valance and conduction band whose density of states depend on energy as N(E)<xE1/2. The optical gap Eg, may thus be obtained from straight line intercept of (ahv)m vs. hv curve. The optical gap Eg is not identical with the mobility gap Ec-Ev, but close to it. ii. An intermediate absorption region (l<a<104 cm“1 ) in which the absorption depends exponentially on photon energy a=a(/exp(hp/E1) (3) with a slope parameter Ej which is usually between 0.05 and 0.08 eV. iii. A weak absorption tail at low absorbtion constants (a^l cm”1) whose shape and magnitude depend on purity, thermal history and preparation conditions. When undoped or lightly doped a-Si:H are exposed to near infrared or visible light, its properties change. Both the conductivity and photoconductivity can be reduced by orders of magnitude. The original properties can be restored by annealing the films in the darkness for about two hours at 150 °C or shorter times [46]. This phenomenon is called Staebler-Wronski effect after the people first observed this effect. It is believed that photo-induced changes in Si-H bonds produce new defects whose states in the gap shift the Fermi level and increase the number of recombination centers. These photo-induced changes are metastable and relax upon annealing. The photocurrent, Ip, contributed by electrons at a given wavelength can be written as [84] ',= eW“ F(l-R)[l-ap{-ad)]Ew (4) xiwhere e: electronic charge, tj: quantum efficiency, /^electron mobility, t”: electron lifetime F: photon flux, jR: reflection coefficient, JE:electric field, w: length of electrodes. Assuming tj=1 for photon energies larger than optical gap Eg one readily obtains from Ip, the product n“Tn, because all other quantities can be measured independently. The photoconductivities and, consequently photocurrents of these films exhibit a dependence on F given by I,=kF> (5) where, 0.5<7<1 and k is a proportionality constant. Rose [52] has explained this 7 dependence using exponentially decreasing density of states model for the localized states which are extending into the mobility gap. The steady state photoconductivities and the dependence on the intensity of illumination could be used to characterize the recombination and the electron lifetimes. The free electron densities n, and the electron lifetimes t”, were obtained from the relations op=epji and n=grn (6) For 100% of electron-hole pair generation, volume generation rate, g, in cm“3 s”1 is given by g=F(l-R)[l-exp(-ad)]ld (7) where a is the optical absorbtion coefficient, and d is the film thickness. In TÜBİTAK- Amorphous Silicon Laboratory (ASL) thin film amorphous silicon samples were obtained in the ultra high vacuum plasma enhanced chemical vapor deposition system (UHV-PECVD). This system has two chambers, load- lock chamber and deposition chamber. A rotary pump is connected to the load- lock chamber and 300 It/sec. capacity turbomolecular pump, (TP) is connected to the deposition chamber. Pneumatically controlled transfer gate valve isolates two chambers from each other, and there is a pneumatically controlled high vacuum valve between TP and deposition chamber. Another feature of this system is the separate gas lines. Reaction gasses, silane (SiH4), methane (CH4), diborone ( B2H6), and phosphine (PH3) enter the deposition chamber separately or premixed. This way cross contamination which comes from gas lines are avoided. Flow rates of these gasses are controlled and xiimonitored by gas flow controllers. Glow discharge plasma in the reaction chamber is excited capacitively. In other words, during deposition gasses pass between two electrodes. One is kept at RF potential with a frequency of 13.56 MHz, and the other is grounded. Samples are loaded onto the grounded plate. Only 2% of the reaction gasses are used, because of high flow rates. The occurance of thickness gradients due to different positions of samples on the substrate holder is avoided this way. Remaining portion of very explosive and poisonous gasses are treated by homemade scrubber system [71] located at the end of exhaust lines. Carefully cleaned Corning-7059 glass substrates are first loaded into the load- lock chamber and this chamber is evacuated. Then substrates are transferred from the load-lock chamber into the deposition chamber which has been kept at deposition temperature and, pumped down to 10“8 torr pressure for at least 10 hours. TP continues to pump deposition chamber loaded with substrates further for about 12 hours. During this time, substrate temperature reaches the deposition temperature ( - 225 °C for intrinsic samples.) During deposition HV valve is closed between TP and the chamber, and deposition pressure (700 mTorr) was maintained by the throttle valve between chamber and the roots pump which is being used to evacuate the chamber during deposition. The plasma chemistry governing the growth of a-Si:H films is very complex. Plasma contains electrons, ions and silicon radicals. These silicon radicals are responsible for the growth of the film [73]. At the same time ions etch the surface of the growing film and this way eliminate the weak (less energetic) structures. Since inside the chamber plasma is far from equilibrium state, the film growth prevails over etching. For the electrical characterization, dark conductivity of produced films were measured as a function of temperature using a home made vacuum cryostat having a thermocouple in good thermal contact with the sample. Conductivity measurement as a function of temperature yields thermal activation energy, AE0 preexponenetial factor Oq, and room temperature conductivity aRT. All samples studied exhibited ohmic behavior and all the measurements were carried out with 100 V applied between coplanar electrodes 1.5 mm apart. Optical measurement produce thickness d, reflection coefficient n, optical absorbtion coefficient a, and the optical gap Eg, of the films. In order to obtain these quantities optical transmissions of films are measured between 400 nm and 1100 nm using a monochromator and a detector set-up which was build in ASL, xiiiand controlled by a computer. Using the computer program developed in ASL, optical transmission data is converted into optical density (OD). From the interference fringes appeared in OD curve sample thickness d, and reflection coefficient n, were calculated. Once the thickness and reflection coefficients are known, optical absorption constant a was obtained as a function of energy. The optical gap Eg is obtained graphically from the (ahp)112 vs. hv curve. For the experiments of photocurrent as a function of intensity again coplanar geometry is used for electrical contacts and experiments are performed in the vacuum cryostat build in ASL. In photocurrent measurements dependence of 7 parameter to light exposure and the change in the value of 7 while Staebler-Wronski effect occuring in the sample was investigated. The photocurrents were generated by monochromatic light having X=633 nm, 670 nm, 687nm, 835 nm and 1015 nm, and having light intensities ranging from 1*105 mW/cm2 to 1*10”9 mW/cm2. As a result, bulk photoconductivity ap was obtained directly from photocurrents and the known geometry, thickness and electric fields. In the experimental set-up, in order to achieve high dynamic range for intensity measurements, rotating neutral density filter (RNDF), and neutral density filter (NDF), combination was utilized. Intensity of the light beam coming out of a monochromatic light source was first changed by RNDF then separated into two beams by using a beam splitter. One of the beams was focused on a detector after chopped by a chopper with certain frequency, the intensity of light which has this frequency was amplified and measured by lock-in amplifier. Second beam was focused on the sample after passing through several NDF's. At the begining, all NDF's were in the system and RNDF was set to minimum intensity position. During the measurements, RNDF was rotated in order to increase the intensity. After a complete turn of RNDF, one or two of the NDF were taken out of the system and above measurement starting again from a minimum intensity position of RNDF repeated. This way 2-8 sets of data were obtained with increasing photocurrents but within the same intensity range. Using the fact that, in two different data sets, the same photocurrent generated by the same intensity, these sets were combined to give one photoconductivity vs. intensity curve which had large intensity interval. In these experiments samples are first annealed, then their photoconductivities were measured as a function of intensity using the same laser used for light soaking (annealed state), then samples are left under monochromatic light for xivlight soaking and the change in the slope parameter 7 was observed. This light soaking and then measuring cycle was continued until no further increase was observed in 7. For monochromatic light sources having photon energies larger than Eg and for the photocurrents larger than dark photocurrent we observed two values for the slope parameter 7, 7 seen at high light intensities yH increased as the exposure time increased. But the 7 seen at low intensities yL remained unchanged during long exposure times. For monochromatic light source having a photon energy 1.48 eV, slightly lower than Eg 7H increased as the exposure time increased and we also observed a larger yL value at lower intensities. For monochromatic light source having a photon energy 1.22 eV, considerably lower than Eg, yH was not observed, and higher yL value is obtained which is consistent with the photoluminesence results found in a-Si:H [93]. In order to test the model of Kocka et al. [70] as much as 14 orders of magnitude in intensity range is obtained through the new method used first in ASL, this way we could obtain photocurrents that are less than dark photocurrents and 7 exponent for the photocurrents which is less than dark current was 1, as Kocka et al. predicted. Three values of 7 exponent was obtained throughout 14 orders of magnitute light intensities, fits remarkably well to the Kocka's model. In conclusion with these thesis, it has been shown that, in TÜBİTAK-Gebze Amorphous Silicon Laboratory device quality thin film hydrogenated amorphous silicon samples are being produced. A new tecnique has been developed in order to measure the photoconductivity for over 14 orders of magnitude of light intensity which also enables the measurement of photocurrents which are less than the dark current. By using the new technique developed, it was possible to verify the theory developed by Kocka et. al. for the dependence of 7-exponent on light intensity. Almost perfect match of experimental data for the theory enables us to understand the behavior of 7 under various illumination conditions. Mutual recombination of electrons and holes determine the /*t product of steady state photoconductivity, /xt takes the saturation value of thermal equilibrium when photocurrent is less than the dark current so it is independent of generation rate and constant. As the generation rate increases, valance band tail states begin to be unoccupied and /it product of holes change a little but /it product for electrons decreases considerably. As the light intensity increased further (high generation rates) conduction band tail states begins to be occupied and assumes a dominant role in recombination traffic, then 7 exponent xvincreases. The relationship between photoconductivity y exponent and Staebler-Wronski effect was demonstrated, and we state that y exponent is a very sensitive parameter to the light induced degradation in hydrogenated amorphous silicon. XVI

Benzer Tezler

  1. A Study of the staeblerwronski effects in hydrogenated amorphous silicon films

    Başlık çevirisi yok

    AYDIN AKKAYA

    Yüksek Lisans

    İngilizce

    İngilizce

    1993

    Fizik ve Fizik MühendisliğiBoğaziçi Üniversitesi

    Fizik Ana Bilim Dalı

    DOÇ. DR. GÜLEN AKTAŞ

  2. Hidrojenlendirilmiş amorf silisyum ve silisyum-karbon alaşımlarında ışık ile yaratılan yarı-kararlı kusurların incelenmesi

    Light-induced metastable defect studies in hydrogenated amorphous silicon and silicon-carbon alloys

    ALP OSMAN KODOLBAŞ

    Doktora

    Türkçe

    Türkçe

    1999

    Fizik ve Fizik MühendisliğiHacettepe Üniversitesi

    Fizik Ana Bilim Dalı

    PROF. DR. ÖZCAN ÖKTÜ

  3. A-Si1-xCx: H filmlerin TSC (thermally stimulated conductivity) ve fotoiletkenlik ölçümleri ile incelenmesi

    Thermally stimulated conductivity (TSC) and photoconductivity measurements in a-Si1-xCx:H

    ALİ OSMAN KOPOLBAŞ

    Yüksek Lisans

    Türkçe

    Türkçe

    1993

    Fizik ve Fizik MühendisliğiHacettepe Üniversitesi

    Fizik Ana Bilim Dalı

    DOÇ. DR. HÜSEYİN TOLUNAY