HAL Id: jpa-00220781
https://hal.archives-ouvertes.fr/jpa-00220781
Submitted on 1 Jan 1981
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
HIGH TEMPERATURE 1H NMR IN a-Si : H
W. Carlos, P. Taylor
To cite this version:
W. Carlos, P. Taylor. HIGH TEMPERATURE 1H NMR IN a-Si : H. Journal de Physique Colloques,
1981, 42 (C4), pp.C4-725-C4-727. �10.1051/jphyscol:19814158�. �jpa-00220781�
JOURKAL DE PHYSIQUE
CoZZoque C 4 , supp26ment au nO1O, Tome 4 2 , octobre 1981 page C4-725
H I G H TEMPERATURE NFlR I N a - S i : H
W.E. Carlos and P.C. Taylor
NavaZ Research Laboratcry, Washington, D. C . 20375, U. S. A .
Abstract.- Pulsed NMR measurements of H have been employed to study the 1 chemical bonding and diffusion of hydrogen atoms in hydrogenated amorphous silicon (a-Si:H) films at elevated temperatures (20-530°C). The spin-lattice relaxation time T 1 is observed to go through a sharp drop above -400°C. The previously observed low temperature T minimum near 30K is essentially unaf- fected by annealing at temperatures up to -500°C where over 50% of the hydrogen 1 has evolved.
Introduction.- Proton nuclear magnetic resonance (NMR) has proved to be a valuable probe of both the bonding and the interactions with the lattice of the hydrogen atoms in a-Si:H (1-3). The NMR signal in virtually all samples consists of two overlapping lines--a narrow (3-4kHz) Lorentzian line which is probably due to ran- domly distributed hydrogen (-3-6 at.%) in monohydride bonding sites, and a broad
(-20-25kHz) Gaussian line which is generally attributed to hydrogen at defects such as internal surfaces or dihydride bonding sites. The broad line is seen even in films which exhibit only monohydride vibrational features in infrared spectroscopy.
In all glow-discharge-prepared films which we have studied the spin-lattice relaxa- tion time, T1, has a minimum at a temperature of about 30K. In addition at lower temperatures (below 30K) T is roughly proportional to the NMR frequency, vo. There is considerable evidence tiat most of the protons are relaxed by hydrogen-containing relaxation centers (disorder modes) (3). The relaxation of these centers to the lattice yields the temperature and frequency dependences of TI. Kowever, there remain some significant difficulties with the details of the interpretation ( 4 1 , and no entirely satisfactory model for these relaxation centers has been developed.
In this work we report the results of experiments which probe the effects of annealing on the low temperature minimum and lineshape as well as the spin lattice interaction at elevated temperatures. The sample used in this work was obtained from RCA Laboratories. It was deposited onto a high temperature (330°C) substrate by the glow discharge of silane and originally contained -12 at.% hydrogen. Infra- red measurements showed only SiH vibrations, and ESR indicated -5x1015 electronic spins/cc. In a previous NMR study of the effects of annealing on a-Si:H, Reimer et al. (5) reported results for a film deposited at room temperature containing 21 at.%
hydrogen. They report-considerable changes in the NMR lines after relatively low temperature anneals (T-300°C) including redistribution of the hydrogen from the broad to the narrow line. In the high-temperature substrate film of lower hydrogen content which we employed for this study, we see no evidence for either significant low temperature (T<300°C) hydrogen evolution or redistribution of the hydrogen.
Results.- The sample of a-Si:H was heated for periods of about 6 hours during which time T was measured. This heating naturally annealed the sample, and measurable hydrogen was driven off at temperatures above -400°C. 1 The effects of the annealing on the NPlR lineshape and the hydrogen content are sunnnarized in Table I. As seen in the table, the annealing produces a narrowing of both the broad and narrow lines due to a decrease in the hydrogen content (and hence an increase in the average hydrogen- hydrogen separation) in both regions. However, the ratio of the intensities in the
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19814158
C4-726 JOURNAL DE PHYSIQUE
broad and narrow lines is also decreasing which indicates that the hydrogen in the more clustered regions is driven off at somewhat lower temperatures.
Table I
Annealing Linewidth* Linewidth* Intensity Ratio Total Electroni Temp. (OC) (Narrow line) (Broad line) (Broad/Narrow) H(at .%) Spins (cm-%)
(Before
annealing) 3.7 kHz 25.6 kHz 3.2 12 15
5x1017
450 2.5 kHz 21.0 kHz 2.5 7 4x1018
500 2.0 Idlz 15.0 kHz 2 4.7 2x1018
530 1.8 MIZ
- o
1.3 7x10*Full width at half maximum.
We estimate the total hydrogen content after annealing by assuming that the volume of the film containing the low density (narrow line) phase does not change with annealing. Since the narrow line is Lorentzian the measured linewidth is proportional to the concentration of hydrogen. Therefore since we know the initial concentration in that line from previous quantitative measurements and we measure the intensity ratio of the broad line to the narrow line, the hydrogen content remaining in each line after annealing can be calculated. As a check on this proce- dure we use as a calibration the broad Gaussian line where the width is proportional to the square root of the concentration and repeat the annealing calculations. The two procedures yield total concentrations within -1 at.% of each other. An average of the two methods is given in Table I. Unlike the results of Reimer et al. (5) on a film with greater hydrogen content prepared on a room temperature substrate, we observe no significant hydrogen loss near 300°C and no redistribution of hydrogen from the broad component into the narrow component above 300°C. However, the tem- peratures at which approximately half the hydrogen has evolved are similar in the two films (-450°C). This similarity may simply reflect the fact that the hydrogen bonding after redistribution (annealing) in the high hydrogen content film is simi- lar to that which exists in our films which were prepared on a high temperature substrate. As we anneal the sample to high enough temperature to drive off hydrogen, the silicon "dangling bond" ESR signal increases (see Table I) as expected since one of the roles of the hydrogen in these films is to passify the silicon dangling bonds.
After the higher temperature runs T was again measured at low temperatures, and the results are shown in Fig. 1.
hi
minimum at T-30K remains even after annealing the film to 500°C when half of the hydrogen has evolved. Furthermore, the depth of the minimum changes very little. After heating the sample to 530°C the hydrogen content drops to -1 at.% and T1 at low temperatures rises by 1-2 orders of magnitude. At this point the minimum has either disappeared or has increased by over two orders of magnitude. It is surprising that the relaxation at low tempera- tures is not significantly affected by annealing at 500°C even though the ESR signal has increased by a factor of -1000 and the hydrogen concentration has decreased by over a factor of two. Clearly any model of the hydrogen-associated relaxation centers must account for the stability of these centers at such high annealing temperatures. The hydrogen in these centers must be as strongly bonded to the silicon matrix as the additional remaining hydrogen.The spin lattice relaxation rate (1/T ) measured at high temperatures is shown in Fig. 2. The rate is essentially constant up to T-300°C, after which it increases 1 with increasing temperature. Although there is some scatter in the data, the general increase in the rate can be described by a process having an activation energy of -0.2 eV. It is interesting to note that we obtain the same activation energy from the changes in the width of the narrow line with annealing temperature. In addition, these two estimates agree reasonably well with results reported for hydrogen evolu- tion from similar films (-0.3 eV) (6).
Fig. 1. Spin lattice relaxation time of Fig.12. Spin lattice relaxation rate H in a-Si:H as a function of reciprocal of H in a-Si:H at elevated tempera- temperature after annealing at T
A' tures. The dashed line is an aid to Smooth curves have been drawn through the the eye.
data.
At temperatures above 400°C measurable amounts of hydrogen evolve from the film. The changes in the spin lattice relaxation rate, however, are not due to the macroscopic motion of protons within the lattice because no raotional narrowing is observed. Furthermore, the T measurements, which took approximately 5-6 hours, were started after the sample had been heated at a given temperature for -30 minutes 1 at which time no additional hydrogen was evolving from the sample.
Summary.- Data for the spin lattice relaxation time of H in a-Si:H at high tempera- tures have been presented. T is seen to decrease fof temperatures above -300°C.
In addition annealing the film to temperatures where -50% the hydrogen has evolved 1 has little effect on the T minimum which indicates that the relaxation centers must be associated with rather lightly bound hydrogen.
Acknowledgements.- D. E. Carlson and J. Dresner are thanked for the sample used in this work. This work was supported in part by the Solar Energy Research Institute (D.O.E. contract No. DE-AI02-80CS83116). One of us (WEC) was an NRL-NRC Research Associate when the work was performed.
References
1. REIMER J.A., VAUGHAN R.W. and KNIGHTS J.C., Phys. Rev. Lett.
44
(1980), 193, Phys. Rev. B2
(1981) 2567, and Phys. Rev. 3 to be published.2. LOWRY M., JEFFREY F.R., BARNES R.G. and TORGESON D.R., Solid State Comm., to be published.
3. CARLOS W.E. and TAYLOR P.C., Phys. Rev. Lett.
45
(1980) 358, and J. Phys. Soc.Japan
49A
(1980) 1193.4. MOVAGHAR B. and SCHWEITZER L., Proceedings Tetr. Bonded Amorphous Semicond.
Conf., Carefree, Arizona, 12-14 Mar 1981, in press.
5. REIMER J.A., VAUGHAN R.W. and KNIGHTS J.C., Solid State Comm.
37
(1981) 161.6. TSAI C.C., FRITZCHE H., TANIELIAN M.H., GACZI P.J., PERSONS P.D. and VESAGHI M.A. in Amorphous and Liquid Semiconductors, ed. W. E. Spear (Univ.
of Edinburgh, 1977), p. 339.
