Growth of InAsPSb Quaternary Alloy by Gas Source Molecular Beam Epitaxy

在三五族半導體材料中，銻磷化銦材料(InAsSb)具有最小的直接能隙，從砷化銦(InAs)的0.417 eV到銻化銦(InSb) 0.235 eV，是作為中紅外線(Mid-Infrared)發光元件或偵測器主動層的極佳材料。在本論文中，我們使用分子束磊晶技術成長銻磷砷化銦合金與砷化銦基板上，用以研究其塊材表面形態、組成與光學特性。在成長塊材實驗中，我們發現昇高長晶基板溫度會增加砷的含量，相對的也減少了的銻的成份。顯微鏡觀察顯示，銻成份的增加使塊材產生鬆弛(relaxation)而導至X光繞射訊號半寬變寬，表面形態變差。使用光激發螢光放光分析(photoluminescence, PL)塊材之光學特性可得知，我們所成長之銻砷化銦材料其躍遷能量範圍為0.2 ~ 0.4 eV (3 ~ 5 micron)；以此材料與砷化銦一同成長量子井結構時，在基板成長溫度450oC時有最強的PL放光。由4K時之光激發螢光放光結果得知，銻砷化銦/砷化銦(InAsSb/InAs)量子井結構之能帶排列屬第二型量子井(type-II)排列；利用一系列不同銻成份(Sb = 0.06 ~ 0.13)之量子井結構PL放光分析與能量躍遷計算，銻砷化銦材料能隙之彎曲係數並非全落在導電帶(conduction band)；而是與價電帶呈40%與60%之比例(conduction band : valence band)。此外，銻成份0.12之量子井結構樣品因為放光波長為4.2 micron，正可使用於偵測二氧化碳之偵器上。一方面，本研究也首次以氣態源分子束磊晶技術成長銻磷化銦材料(InPSb)於砷化銦基板上，其中晶格最匹配的樣器其X光繞射結果之磊晶層訊號半寬只有65秒；由於混溶隙(miscibility gap)的存在，成長單一晶相(single phase)合金的成長條件非常嚴格，稍許差異都會造成相分離(phase separation)。在砷化鎵與磷化銦基板上的成長也顯示了不同晶格常數基板對銻磷化銦材料成長所造成的影響。在電性方面，霍爾效應量測(Hall effect measurement)指出，鈹(Be)與矽(Si)可用於分子束磊晶成長銻磷化銦材料之p型與n型摻雜雜質。對四元材料銻磷砷化銦(InAsPSb)的成長，文中利用理論計算其混溶隙區域的範圍，以晶格匹配於砷化銦基板之四元組成分析，在成長溫度470oC時產生相分離之最小砷成份為0.39，此外，根據文獻資料，我們也計算了InAsPSb成份組成與能隙大小與溫度變動時之能隙改變關係，這些理論值亦用於其後之四元材料特性分析。一系列四元材料樣品中，由顯微鏡表面分析，砷含量較少(深入混溶隙區域)之樣品表面型態較差，搭配X光繞射與電子束微探儀(electron-probe microanalyzer, EPMA)定量分析，這些樣品均呈現了程度不一之相分離；低溫時，樣品InAs0.04P0.67Sb0.29光激發螢光放光能量較能隙小約223 meV。在此光激發螢光放光光譜中，解析出兩種放光模式，其中我們將類似高斯曲線的放光模式歸因於深階能階(deep level)所造成，它的特性可以使用組態座標模型(configuration coordinate model)來解釋，這些深階能階主要可能是晶格中空洞與雜質複合產生，例：碳原子雜質與銦原子空洞(VIn-CAs or VIn-CIn)；另一個解析出較低能量區域之放光模式則歸因於帶尾能態的載子複合機制造成。高砷成份之四元樣品變溫光激發螢光放光則顯示不同的放光機制，S型峰值溫度變化可由低溫時帶尾能態的載子複合機制解釋，較高溫時隨能隙變化而下降之放光峰值則為導電帶至電洞能階放光(conduction band - acceptor level recombination)。後，我們試著以前所述之四元材料製作一p-i-n結構光偵測器，此一光偵測器為表面收光，採用傳統的溼蝕刻製程製作。所製作光偵測器偵測範圍為1 ~ 3 micron，峰值位於2.6 micron，光反應度(responsivity)為0.3 A/W，探測能力為1.7×109 cmHz1/2/W，足可比擬市售硒化鉛(PbSe)偵測器於77K時之表現。Among III-V compound semiconductors, InAsSb material system has the lowestandgap energy ranging from 0.417 eV (InAs) to 0.235 eV (InSb) and is the best activeayer candidate for IR gas detectors. In this study, series of InAsSb alloy samples wererown using solid source molecular beam epitaxy on InAs substrates to investigate theirurface morphology, structural and optical characteristics. Increasing growthemperature enhances the As incorporation and leads to the decreasing of Sb moleraction. Under microscope investigation, XRD FWHM broadens and relaxationnduced crosshatch begins to appear on epilayer surface with increasing Sboncentration. Photoluminescence (PL) measurement was also taken to acquire allamples’ bandgap energy and qualitatively compare each sample’s opticalharacteristics. Transition energies of above mentioned samples are within the range of.2~0.4 eV (3~5 um). For InAsSb/InAs multiple quantum wells grown at differentubstrate temperature, sample exhibits better optical quality with growth temperaturelose to 450oC and 4 K photoluminescence result shows that the band alignment wasetermined to be staggered type-II. From samples with the Sb mole fraction rangingrom 0.06 to 0.13, it was found that both the conduction and valence bands of InAsSblloy exhibit some bowing by comparing the emission peak energies with a transitionnergy calculation. The bowing parameters were determined to be in the ratio of 4:6.or a sample with Sb composition ~0.12 in the quantum well the photoluminescencemission band covers the CO2 absorption peak making it suitable for use in sources forO2 detection.nPSb ternary sample were also successfully grown on InAs substrate which weelieve that it is the first demonstration of MBE grown single phase InPSb bulk layer onnAs. The XRD FWHM of InPSb epilayer signal is only 65arcsec. Substrate effect andhase separation were observed for InPSb grown on InP and GaAs substrates. Halleasurements shows that Be and Si can be p-type and n-type dopants for InPSbaterial respectively.or InAsPSb quaternary, bandgap energies of quaternary as well as miscibilityap using strictly regular solution approximation are calculated. For InAs latticeatched InAsPSb samples grown at 470oC, the As composition limit is 0.39. Rougherurface morphology can be seen as the arsenic composition in quaternary bulkncreasing. Near band edge photoluminescence emission is observed for high As moleraction sample while 223 meV of deviation between PL energy and calculated bandgapnergy for InAs0.04P0.67Sb0.29. Two photoluminescence bands are resolved. Theaussian-like line shape is attributed to be the result of deep-level defects which can bellustrated by configuration coordination model. The deep levels are possibly due toacancy-impurity complexes that are composed of a substitutional carbon and an Inacancy, i.e., a VIn-CAs or VIn-CIn complex. The other band located lat low energy sidehose peak redshifts and intensity decreases as temperature increasing is ascribed to thearrier recombination in the tail states.nAsPSb p-i-n photodetectors were made to be operated in room temperaturesing gas source molecular beam epitaxy. The surface illuminated photodetector with aesa structure were fabricated by a conventional device process. The detectableavelength ranges from 1um to near 3 um with peak responsivity equal to 0.3 A/Wocated at 2.6 um. The Johnson noise limited detectivity of 1.7×109 cmHz1/2/W isomparable with PbSe detector operates at 77 K and outperforms that at 300 K.中文摘要 ibstract iiiontents vigure Index viiable Index xiiihapter 1 Introduction 1.1 Historical reviews 1.2 Experiment 4.2.1 Epitaxy 4.2.2 Composition determination 6.2.3 Photoluminescence measurement 6.3 Thesis structure 8hapter 2 InAsSb and InPSb 9.1 InAsSb bulk 9.2 InAsSb/InAs multiple quantum wells 14.3 InPSb 21hapter 3 InAsPSb 29.1 Miscibility gap calculation 29.2 Bandgap Estimation 34.3 Growth of InAsPSb 37hapter 4 Photoluminescence study of InAsPSb 45.1 PL study of sample outside MG 45.1.1 InAs0.68P0.22Sb0.10 48.1.2 InAs0.56P0.28Sb0.16 56.1.3 InAs0.36P0.41Sb0.23 58.2 Deep level recombination of samples inside MG 61.2.1 InAs0.22P0.52Sb0.26 61.2.2 InAs0.09P0.63Sb0.28 63.2.3 InAs0.04P0.66Sb0.30 65.2.4 Deep-level defect analysis 68hapter 5 InAsPSb photodetector 79.1 Introduction 79.2 Fabrication 80.3 Optical response measurements 82.4 Result and conclusion 85hapter 6 Conclusion 91ibliography 9