Thin Solid Films, 184 (1990) 15-19 15 P-TYPE D E L T A D O P I N G I N S I L I C O N MBE N. L. MATTEY, M. HOPKINSON*, R. F. HOUGHTON, M. G. DOWSETT, D. S. McPHAIL, T. E. WHALL AND E. H. C. PARKER Department of Physics, University of Warwick, Coventry CV4 7AL (U.K.) G. R. BOOKER AND J. WHITEHURST Department of Metallurgy and Science of Materials, Parks Road, Oxford OX13PH ( U.K.) (Received May 31, 1989) P-type delta-doped layers have been prepared for the first time in silicon molecular beam epitaxy by evaporation of elemental boron. The dopant and carrier distributions have been investigated using secondary ion mass spectrometry, transmission electron microscopy and capacitance-voltage measurements and it is deduced that the full width at half m a x i m u m of the delta layers is about 2 nm. Hall measurements indicate complete activation at 300 K and a mobility of 30 + 5 cm 2 V - 1 S - 1 for a sheet carrier concentration of 9 + 2 x 1012 cm -2. 1. INTRODUCTION The two-dimensional electron/hole gas formed in the quantum well produced by a delta-doped layer is of considerable scientific interest and finds application in a number of novel devices including buried channel field effect transistors ~ and infrared detectors 2. N-type delta doping in molecular beam epitaxy (MBE) silicon has been achieved using antimony in combination with solid phase epitaxy (SPE) 3 or low energy implantation 4. More recently the first growth of p-type delta layers in silicon, using both gallium 5 and boron 6 has been reported. In this paper we give a more detailed discussion of the growth of boron-doped delta layers, boron having advantages over gallium in that it is completely ionized at 3 0 0 K and shows no tendency to segregate at the surface. Antimony delta layers have also been prepared in this investigation 6 using SPE 3. 2. MBE GROWTH PROCEDURE Layers were grown on (100)-oriented p-substrates in a V G Semicon V80 MBE system fitted with a graphite-lined elemental boron source and pyrolytic boronnitride-lined antimony source. Standard back-damaged wafers were used as received. A 2 nm silicon cap was deposited at 400 °C and heated to 800 °C for 7 min * Present address: Department of Electrical Engineering, Universityof Sheffield, Sheffield,U.K. 0040-6090/90/$3.50 © ElsevierSequoia/Printed in The Netherlands 16 N.L. MATTEYet al. to desorb the native oxide. The layers were grown at a substrate temperature ~ of 710 °C and a silicon growth rate of 0.5 nm s - 1. The following procedure was adopted to minimize boron diffusion and broadening of the delta layers during growth. Initially, a 0.3 gm p-buffer layer was deposited, followed by an 0.4~tm n-layer of approximate concentration 1 X 1016 c m - 3. Growth was interrupted, T~ lowered to 400 °C and the layer exposed to a boron flux at 1011 atoms s 1 for 10-240s to produce surface coverages of between 1012 and 2 x 1013 c m - 2 . Finally growth of the n-layer was resumed at 700 °C for 1-4 min, leaving the delta layers 30-120 nm below the surface. 3. COMPOSITIONAL AND STRUCTURAL CHARACTERIZATION Secondary ion mass spectrometric (SIMS) analysis was carried out in a quadrupole instrument (EVA 2000) using 0 2 + primary ions of energies between 3.4 and 0.9 keV at normal incidence, the results are summarized in Table I. These data are indicative of ion-beam-induced atomic mixing which lowers the peak concentration, broadens the profile and gives rise to a leading edge which is sharper than the trailing edge. As the impact energy decreases, the profile narrows and the peak concentration increases. The apparent feature depth also increases owing to the differential shift caused by the shortening pre-equilibrium period. The areal density of the dopant is obtained from the area under the profile which remains constant with changing impact energy. TABLE 1 SIMS ANALYSIS OF BORON ~ LAYER Primary ion energy (keV) Apparent depth (gm) Peak concentration (1019 c m 3) F W H M (nm) Upslope per n m per decade Downslope per n m per decade 3.4 0.110_+0.005 1.1 +__0.1 6.8--0.5 4.2 + 0.2 8.7 4- 0.4 2.4 0.112_+0.005 1.4 + 0.1 5.8_+0.4 3.8 + 0.2 6.6 _+0.3 1.4 0.113+_0.005 1.7 + 0.2 4.5+0.4 3.6 + 0.2 6.2 + 0.3 0.9 0.125+0.005 1.9 ___0.2 3.5+0.4 3.3 + 0.2 5.2 +__0.3 Figure 1 shows the profile obtained at 0.9 keV impact energy, the full width at half m a x i m u m ( F W H M ) is 3.5 _+0.3 nm and the areal density is 1 × 1013 cm -2. This width must be regarded as an upper limit. Extrapolating the measured values to zero impact energy suggests a F W H M of about 2.5 nm and a peak concentration of about 2 x 1019 c m - 3 . This, however, assumes that any SIMS-induced broadening scales linearly with incident ion energy. A more precise measure of the layer width is possible from transmission electron microscopy (TEM) measurements: preliminary analysis was performed on a boron delta layer of areal density 1013 c m - 2 using a J E O L 200 CX microscope. Figure 2 shows a cross-sectional bright field image in which the boron layer is seen as a dark line of width approximately 2 nm at about 80 nm below the surface. The structure in the boron layer image is possibly due to precipitation, but further work is required to determine whether this is responsible for the contrast seen in the bright field image or whether it is due to strain or some other effect. |7 p - T Y P E DELTA D O P I N G IN SILICON MBE #o lO19. E 1018 I--- 1017 (_J lO16 1015 o.i . . . . . . . . DEPTH ~2 I 03 urn Fig. I. S I M S profile of a b o r o n delta layer, obtained using 0.9 keV 0 2 + ions at normal incidence. Surface ~ - ~ Layer ~ 50nrn Fig. 2. Cross-section T E M m i c r o g r a p h (bright field, two-beam, g = 400). The b o r o n layer can be seen as a dark line of width 2 nm at a depth 80 n m below the surface. 4. ELECTRICAL CHARACTERIZATION Electrical measurements were carried out on Hall bars of dimensions 280 [am x 30 [am defined by reactive ion etching. Contacts were made to the delta layer by ion implantation. To maintain the thermal budget below that of deposition, the implants were partially activated by annealing in a 1000 °C ambient for less than about 10 s and a room temperature plasma-enhanced oxidation process 7 was used to passivate the devices. Resistivity and Hall effect measurements were made on several samples at 300 K. A sheet carrier concentration of 9_+ 2 x 1012 c m - 2 and a 18 N.L. MATTEYet al. carrier mobility of 30__+5 c m 2 V - 1 S- 1 were obtained assuming one type of carrier and a Hall scattering factor of unity. A mobility of 30 cm 2 V - a s - 1 is lower than might be expected from the three-dimensional dopant concentration of about 2 x 1019cm -3 obtained from the SIMS measurements and is typical of uniform doping of concentrations 6 x 101° c m - 3 8. However, such a comparison is perhaps p r e m a t u r e - - f u r t h e r work is needed to clarify the situation with regard to the spatial distribution of the carriers and the role of quantum confinement in the delta layer. Capacitance-voltage (CV) measurements were made on delta layers located 75 nm below the silicon surface. Such measurements are not possible for sheet concentrations of greater than or equal to 3.5 x 1012cm 2 owing to avalanche breakdown effects 9 and have therefore been carried out on a sample of dopant concentration 2 x 1012 cm - 2, as determined by SIMS. Aluminium was used to make a Schottky barrier to the n-top surface and ohmic contact to the delta layer, the contact area to the latter being increased by a shallow-angle mechanical bevel. The carrier profile was obtained using the standard equation based on the depletion approximation and is shown in Fig. 3. As has been discussed by, for example, Blood 1° series resistance Rs and parallel resistance Rp will lead to errors of interpretation unless Rs/R p "~ 1 and Rs ~ 1/coo. Checks were made to ensure that these conditions were fulfilled in the present measurements, the second criterion being met by using a low operating frequency (300 Hz); the effects of gap states being assumed to be negligible for these high dopant concentrations. The sheet carrier concentration has been obtained from the area under the curve in Fig. 3 and is found to be about 3 x 1012 c m - 2 in reasonable agreement with the dopant concentration measured by SIMS (2 × 1012 cm-2). The observed F W H M of approximately 3 nm compares favourably with recent classical 9 and quantum 11 calculations which predict a F W H M of 7.8 nm and 3-5 nm (depending on carrier effective mass) respectively. 96//I] 10 20 . . . i . . . . . I . . . i . ' ' ' " I . . . . . . . . 10,9 + 'r I0~ , , , , I i , i i i , , , , I , , , , I IZX~id'h/~ Fig. 3. Carrier profile obtained from CV measurements on a boron delta layer of sheet concentration 2 x 1012cm-2 located 75 nm below the surface. p-TYPE DELTA DOPING IN SILICON MBE 19 5. CONCLUSIONS The first b o r o n delta-doped layers have been g r o w n in silicon. S I M S a n d T E M m e a s u r e m e n t s suggest that the layer width is a b o u t 2 nm. CV m e a s u r e m e n t s confirm a very n a r r o w carrier distribution. Hall m e a s u r e m e n t s show complete activation at 3 0 0 K with a m o b i l i t y of 3 0 c m 2 V - I s -1 for a sheet carrier c o n c e n t r a t i o n of 9 × 10~2cm -2. ACKNOWLEDGMENTS The a u t h o r s wish to t h a n k R. G. Biswas (Warwick) for assistance with the Hall m e a s u r e m e n t s a n d A. G u n d l a c h (Edinburgh), S. T a y l o r a n d W. Eccleston (Liverpool) for the fabrication of the Hall bars. REFERENCES 1 H.P. Zeindl, B. Bullemer,I. Eiseleand G. Tempel, J. Electrochem. Soc., 136 (1989) 1129. 2 G. Tempel, N. Schwarz, F. Muller, F. Koch, H. P. Zeindl and 1. Eisele, Thin Solid Films, 184 (1990) 171. 3 D.C. Streit, R. A. Metzger and F. G. Allen, Appl. Phys. Lett., 44 (1984) 234. 4 H.M. Li, M. Willander, W. Ni, K. F. Berggren, B. Serneliusand G. V. Hansson, Thin Solid Films, 183 (1989) 331. 5 H.P. Zeindl, T. Weghaupt and I. Eisele,5th European Workshop on MBE, Grainau, March 1989. 6 N.L. Mattey, M. Hopkinson, R. F. Houghton, T. E. Whall and E. H. C. Parker, 5th European Workshop on MBE, Grainau, March 1989. 7 8 9 10 I1 S. Taylor, W. Eccleston and P. Watkinson, Electron. Lett., 23 (14) (1987) 732. G. Masetti, M. Severiand S. Solmi, IEEE Trans. Electron. Devices, ED30(7) (1983) 764. A.A. Van Gorkum and K. Yamaguchi, IEEE Trans. Electron. Devices, ED6(2) (1989) 410. P. Blood, Semicond. Sci. Technol., 1 (1986) 7. E.F. Schubert and K. Ploog, Jpn. J. Appl. Phys., 25(7) (1986) 966.