Hydrogen in a-Si:H Deposited by an Expanding Thermal Plasma: A Temperature, Growth Rate and Isotope Study

  • PDF / 413,956 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 8 Downloads / 152 Views

DOWNLOAD

REPORT


electrons can be neglected as no power is coupled in downstream, leading to a low electron temperature ( z 0.2 eV). The films are deposited at substrates positioned 32 cm from the arc outlet at a temperature controlled substrate holder. A more detailed description is given in [2]. The setup conditions used are listed in Table I, as well as the growth rate and silane consumption determined respectively by an in situ ellipsometer at 632.8 nm and by a residual gas analyzer. Gas flows and arc current have been chosen in such a way that the relative silane consumption is about equal for the three conditions and that the silane dissociation is dominated by atomic hydrogen. Replacement of hydrogen by deuterium leads to a higher growth rate and silane consumption as its larger mass leads to a more efficient operation of the arc and a larger interaction with silane. Table I: Setup processing values for three different conditions. The values in brackets refer to the replacement of hydrogen by deuterium. condition 1 2 3

Ar-flow (slm) 3.3 3.3 2.1

H 2(D 2)-flow

(slm) 0.6 0.3 0.12

SiH 4-flow (slm) 0.6 0.2 0.04

arc current (A) 45 25 12.5

growth rate (nm/s) 11(21) 3(7) 0.3(0.7) 1

silane consumption 12%(22%) 10%(21 %) 8%(17%)

Determination of hydrogen content The hydrogen content has been examined by two different techniques. From elastic recoil detection (ERD) combined with Rutherford Backscattering (RBS), using a beam of cyclotron generated 4 MeV 4He+ ions, the hydrogen and silicon density have been obtained. The results have been verified by a second ERD setup using a beam of 54 MeV 65Cu ions generated by a Van der Graaff accelerator. More information on both ERD techniques can be found in [4,5]. Although ERD leads to unambiguous results a major disadvantage is that no information on the nature of the hydrogen bonds is obtained. This information has been obtained by the second technique, i.e., infrared (IR) absorption spectroscopy. From this, the total hydrogen content can be obtained from the hydrogen wagging mode at 640 cmI and from the sum of the stretching modes at about 2000 and 2100 cm . These correspond with monohydride (SiH) and dihydride (SiH 2) bonding modes respectively [6]. Trihydride bonded hydrogen (SiH 3) is normally absent. The concentration Ni of a bond in a specific mode is linked with the integrated absorbance Ii =cfi/oidoiat the frequency (oi under consideration by a proportionality constant Ai which is theoretically defined as [7]

A

2 *2 Ai-cno i

(1)

27c ei where c is the velocity of light, n the refractive index, jt the reduced mass, ei the effective charge of the dipole and ci the absorption coefficient. Note that the reduced mass also appears in the eigenfrequency. These proportionality constants are given in literature [7,8] but because of inconsistancies they have been recalibrated from combining our ERD/RBS and IR spectroscopy results on a large number of films deposited under several conditions. To obtain

530

proportionality constants for silicon-deuterium bondings SiH 4 has been replac