• Current Photovoltaic Research
• /
• v.9 no.3
• /
• pp.75-83
• /
• 2021

The tunnel oxide passivated contact (TOPCon) structure got more consideration for development of high performance solar cells by the introduction of a tunnel oxide layer between the substrate and poly-Si is best for attaining interface passivation. The quality of passivation of the tunnel oxide layer clearly depends on the bond of SiO in the tunnel oxide layer, which is affected by the subsequent annealing and the tunnel oxide layer was formed in the suboxide region (SiO, Si₂O, Si₂O₃) at the interface with the substrate. In the suboxide region, an oxygen-rich bond is formed as a result of subsequent annealing that also improves the quality of passivation. To control the surface morphology, annealing profile, and acceleration rate, an oxide tunnel junction structure with a passivation characteristic of 700 mV or more (V_oc) on a p-type wafer could achieved. The quality of passivation of samples subjected to RTP annealing at temperatures above 900℃ declined rapidly. To improve the quality of passivation of the tunnel oxide layer, the physical properties and thermal stability of the thin layer must be considered. TOPCon silicon solar cell has a boron diffused front emitter, a tunnel-SiO_x/n⁺-poly-Si/SiN_x:H structure at the rear side, and screen-printed electrodes on both sides. The saturation currents J_o of this structure on polished surface is 1.3 fA/cm² and for textured silicon surfaces is 3.7 fA/cm² before printing the silver contacts. After printing the Ag contacts, the J_o of this structure increases to 50.7 fA/cm² on textured silicon surfaces, which is still manageably less for metal contacts. This structure was applied to TOPCon solar cells, resulting in a median efficiency of 23.91%, and a highest efficiency of 24.58%, independently. The conversion efficiency of interdigitated back-contact solar cells has reached up to 26% by enhancing the optoelectrical properties for both-sides-contacted of the cells.

• Journal of the Microelectronics and Packaging Society
• /
• v.10 no.4
• /
• pp.39-46
• /
• 2003

Sn-Cu eutectic solder bump was fabricated by electroplating for flip chip and its characteristics were studied. A Si-wafer was used as a substrate and the UBM(Under Bump Metallization) of Al(400 nm)/Cu(300 nm)/Ni(400 nm)/Au(20 nm) was coated sequentially from the substrate to the top by an electron beam evaporator. The experimental results showed that the plating ratio of the Sn-Cu increased from 0.25 to 2.7 $\mu\textrm{m}$/min with the current density of 1 to 8 A/d$\m^2$. In this range of current density the plated Sn-Cu maintains its composition nearly constant level as Sn-0.9∼1.4 wt%/Cu. The solder bump of typical mushroom shape with its stem diameter of 120 $\mu\textrm{m}$ was formed through plating at 5 A/d$\m^2$ for 2 hrs. The mushroom bump changed its shape to the spherical type of 140 $\mu\textrm{m}$ diameter by air reflow at $260^{\circ}C$. The homogeneity of chemical composition for the solder bump was examined, and Sn content in the mushroom bump appears to be uneven. However, the Sn distributed more uniformly through an air reflow.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.28 no.5
• /
• pp.179-184
• /
• 2018

To lessen oxygen concentrations in a wafer through modifying the length of graphite heaters, we investigated the influence of relative distance from heater to quartz crucible on temperature profile of hot-zone in Czochralski silicon-crystal growth by simulation. In particular, ATC temperature and power profiles as a function of different ingot body positions were investigated for five different heater designs; (a) typical side heater (SH), (b) short side heater-up (SSH-up), (c) short side heater-low (SSH-low), (d) bottom heater without side heater (Only-BH), and (e) side heater with bottom heater (SH + BH). It was confirmed that lower short side heater exhibited the highest ATC temperature, which was attributed to the longest distance from triple point to heater center. In addition, for the viewpoint of energy efficiency, it was observed that the typical side heater showed the lowest power because it heated more area of quartz crucible than that of others. This result provides the possibility to predict the feed-forward delta temperature profile as a function of various heater designs.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2009.06a
• /
• pp.456-456
• /
• 2009

Boron doping on an n-type Si wafer is requisite process for IBC (Interdigitated Back Contact) solar cells. Fiber laser annealing is one of boron doping methods. For the boron doping, uniformly coated or deposited film is highly required. Plasma enhanced chemical vapor deposition (PECVD) method provides a uniform dopant film or layer which can facilitate doping. Because amorphous silicon layer absorption range for the wavelength of fiber laser does not match well for the direct annealing. In this study, to enhance thermal affection on the existing p-a-Si:H layer, a ${\mu}c$-Si:H intrinsic layer was deposited on the p-a-Si:H layer additionally by PECVD. To improve heat transfer rate to the amorphous silicon layer, and as heating both sides and protecting boron eliminating from the amorphous silicon layer. For p-a-Si:H layer with the ratio of $SiH_4$ : $B_2H_6$ : $H_2$ = 30 : 30 : 120, at $200^{\circ}C$, 50 W, 0.2 Torr for 30 minutes, and for ${\mu}c$-Si:H intrinsic layer, $SiH_4$ : $H_2$ = 10 : 300, at $200^{\circ}C$, 30 W, 0.5 Torr for 60 minutes, 2 cm $\times$ 2 cm size wafers were used. In consequence of comparing the results of lifetime measurement and sheet resistance relation, the laser condition set of 20 ~ 27 % of power, 150 ~ 160 kHz, 20 ~ 50 mm/s of marking speed, and $10\;{\sim}\;50 {\mu}m$ spacing with continuous wave mode of scanner lens showed the correlation between lifetime and sheet resistance as $100\;{\Omega}/sq$ and $11.8\;{\mu}s$ vs. $17\;{\Omega}/sq$ and $8.2\;{\mu}s$. Comparing to the singly deposited p-a-Si:H layer case, the additional ${\mu}c$-Si:H layer for doping resulted in no trade-offs, but showed slight improvement of both lifetime and sheet resistance, however sheet resistance might be confined by the additional intrinsic layer. This might come from the ineffective crystallization of amorphous silicon layer. For the additional layer case, lifetime and sheet resistance were measured as $84.8\;{\Omega}/sq$ and $11.09\;{\mu}s$ vs. $79.8\;{\Omega}/sq$ and $11.93\;{\mu}s$. The co-existence of $n^+$layeronthesamesurfaceandeliminating the laser damage should be taken into account for an IBC solar cell structure. Heavily doped uniform boron layer by fiber laser brings not only basic and essential conditions for the beginning step of IBC solar cell fabrication processes, but also the controllable doping concentration and depth that can be established according to the deposition conditions of layers.

• Journal of Sensor Science and Technology
• /
• v.11 no.6
• /
• pp.342-349
• /
• 2002

New formation technique of metal oxide sensing film was proposed m this paper. Silicon wafer with Pt electrodes was used as a substrate for depositing metal Sn film. Metal Sn was deposited in the state of not continuous film but only island state. The samples were prepared to obtain the optimal condition of metal Sn deposition. The resistances of deposited Sn onto Pt electrodes amounted to $1\;k{\Omega}$, $5\;k{\Omega}$, $10\;k{\Omega}$ and $50\;k{\Omega}$, respectively. Also The sample with $1,500\;{\AA}$ thickness of Sn was prepared m order to compare sensing properties between conventional type and proposing type. After deposition of metal Sn, $SnO_2$ was formed by thermal oxidation method for 3 hrs. in $O_2$ ambient at $700^{\circ}C$. Surface morphology, crystal structure and surface roughness of oxidized-sensing film were examined by SEM, XRD, and AFM, respectively. From the results of these analyses, the optimal deposition condition of Sn was that the Pt electrode resistance became $10\;k{\Omega}(300\;{\AA})$. Also, the sensing characteristics of fabricated sensing film for various concentrations of butane, propane and carbon monoxide gases were measured at he operating temperatures of $250^{\circ}C$, $300^{\circ}C$ and $350^{\circ}C$, respectively. Although catalyst as not added to the sensing film, it has exhibited the high sensitivity to all the test gases.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.1
• /
• pp.603-608
• /
• 2018

Although the voltage-applied discharge method is most widely used in the semiconductor and display industries, periodic management costs are incurred because the method causes defects due to the absorption of ambient fine dust and causes emitter tip contamination due to the discharge. The emitter tip contamination problem is caused by the accumulation of fine particles in ambient air due to the corona discharge of the ionizer. Fuzzy ball generation accelerates the wear of the emitter tip and deteriorates the performance of the ionizer. Although a mechanical cleaning method using a manual brush or an automatic brush is effective for contaminant removal, it requires management of additional mechanical parts by the user. In some cases, contaminants accumulated in the emitter may be transferred to the wafer or product. In order to solve this problem, we developed an ionizer for a clean environment that can remove the pencil-type emitter tip and directly ionize the surrounding gas molecules using the tungsten wire located inside the ion tank. As a result of testing and certification by the Korea Institute of Machinery and Materials, the average concentration was $0.7572particles/ft^3$, the decay time was less than two seconds, and the ion valance was 7.6 V, which is satisfactory.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2010.06a
• /
• pp.264-264
• /
• 2010

Single-walled carbon nanotubes (SWCNTs) were grown on a Si wafer by using thermal chemical vapor deposition (t-CVD). We investigated the effect of the catalyst deposition rate on the types of CNTs grown on the substrate. In general, smaller islands of catalyst occur by agglomeration of a catalyst layer upon annealing as the catalyst layer becomes thinner, which results in the growth of CNTs with smaller diameters. For the same thickness of catalyst, a slower deposition rate will cause a more uniformly thin catalyst layer, which will be agglomerated during annealing, producing smaller catalyst islands. Thus, we can expect that the smaller-diameter CNTs will grow on the catalyst deposited with a lower rate even for the same thickness of catalyst. The 0.5-nm-thick Fe served as a catalyst, underneath which Al was coated as a catalyst support as well as a diffusion barrier on the Si substrate. The catalyst layers were. coated by using thermal evaporation. The deposition rates of the Al and Fe layers varied to be 90, 180 sec/nm and 70, 140 sec/nm, respectively. We prepared the four different combinations of the deposition rates of the AI and Fe layers. CNTs were synthesized for 10 min by flowing 60 sccm of Ar and 60 sccm of $H_2$ as a carrier gas and 20 sccm of $C_2H_2$ as a feedstock at 95 torr and $810^{\circ}C$. The substrates were subject to annealing for 20 sec for every case to form small catalyst islands prior to CNT growth. As-grown CNTs were characterized by using field emission scanning electron microscopy, high resolution transmission electron microscopy, Raman spectroscopy, UV-Vis NIR spectroscopy, and atomic force microscopy. The fast deposition of both the Al and Fe layers gave rise to the growth of thin multiwalled CNTs with the height of ${\sim}680\;{\mu}m$ for 10 min while the slow deposition caused the growth of ${\sim}800\;{\mu}m$ high SWCNTs. Several radial breathing mode (RBM) peaks in the Raman spectra were observed at the Raman shifts of $113.3{\sim}281.3\;cm^{-1}$, implying the presence of SWCNTs (or double-walled CNTs) with the tube diameters 2.07~0.83 nm. The Raman spectra of the as-grown SWCNTs showed very low G/D peak intensity ratios, indicating their low defect concentrations.

• Safety and Health at Work
• /
• v.2 no.3
• /
• pp.210-217
• /
• 2011

Objectives: The purpose of this study was to measure the concentration of volatile organic compound (VOC)s originated from the chemicals used and/or derived from the original parental chemicals in the photolithography processes of semiconductor manufacturing factories. Methods: A total of four photolithography processes in 4 Fabs at three different semiconductor manufacturing factories in Korea were selected for this study. This study investigated the types of chemicals used and generated during the photolithography process of each Fab, and the concentration levels of VOCs for each Fab. Results: A variety of organic compounds such as ketone, alcohol, and acetate compounds as well as aromatic compounds were used as solvents and developing agents in the processes. Also, the generation of by-products, such as toluene and phenol, was identified through a thermal decomposition experiment performed on a photoresist. The VOC concentration levels in the processes were lower than 5% of the threshold limit value (TLV)s. However, the air contaminated with chemical substances generated during the processes was re-circulated through the ventilation system, thereby affecting the airborne VOC concentrations in the photolithography processes. Conclusion: Tens of organic compounds were being used in the photolithography processes, though the types of chemical used varied with the factory. Also, by-products, such as aromatic compounds, could be generated during photoresist patterning by exposure to light. Although the airborne VOC concentrations resulting from the processes were lower than 5% of the TLVs, employees still could be exposed directly or indirectly to various types of VOCs.

• Journal of Sensor Science and Technology
• /
• v.1 no.1
• /
• pp.23-34
• /
• 1992

In order to fabricate magnetoresistive sensor, Fe-Ni and Co-Ni alleys were evaporated on the slide glass and the silicon wafers. Saturation magnetic induction($B_{s}$), coercive field strength($H_{c}$) and magnetoresistance were measured for fabricated samples. The evaporated Fe-Ni thin films show that the saturation magnetic induction was 0.65 T, and coercive field strength was 0.379 A/cm, and this value was changed to 0.370 A/cm(//), 0.390 A/cm(${\bot}$), respectively after magnetic annealing. For the measurement of coercive field strength, magnetizing frequency of 1 kHz was used. For the fabricated sensor element, the change of magnetoresistance (${\Delta}R/R$) was excessively unstable due to oxidation in the process of fabrication. The evaporated Co-Ni alloy thin films show that saturation magnetic induction was 0.66 T, and coercive field strengthes were 5.895 A/cm(//), 5.898 A/cm(${\bot}$), respectively, after magnetic annelaing. The change of magnetoresistance(${\Delta}R/R$) was $3.6{\sim}3.7%$ of which value was excessively stable to room temperature. Fe-Ni thin film could have many problems due to large affinity in the process of fabrication of magnetoresistance sensor, but Co-Ni thin film could be a suitable material for fabrication of magnetoresistance sensor, because of its small affinity and definite magnetoresistance effects.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.26 no.3
• /
• pp.270-280
• /
• 2016

This paper investigates the evolution of EM admittance of piezoelectric transducers mounted on a notched beam from wave propagation perspective. A finite element analysis is adopted to obtain numerical solutions for Lamb waves reverberating on the notched beam. The mode-converted Lamb wave signals due to a notch are extracted by using the polarization characteristics of piezoelectric transducers collocated on the beam. Then, a series of temporal spectrums are computed to demonstrate the evolution of EM admittance through fast Fourier transform of the mode-converted Lamb wave signals which are consecutively truncated in the time domain. When truncation time is relatively small, the corresponding temporal spectrum is governed by the characteristics of the input driving frequency. As truncation time becomes large, however, the modal characteristics of the notched beam play a crucial role in the temporal spectrum within the input driving frequency band. This implies that mode-converted Lamb waves reverberating on the beam contributes to the resonance of the beam. The root mean square values are computed for the temporal spectrums in the vicinity of each resonance frequency. The root mean square values increase monotonically with respect to truncation time for any resonance frequencies. Finally the implications of the numerical observation are discussed in the context of damage detection of a beam.