- Email: [email protected]
Magnetic, structural and magnetoresistive properties of magnetron-sputtered thin NiFe films
Thm Sohd Fdms, 125 (1985) 251-256 ELECTRONICS AND OPTICS
251
MAGNETIC, STRUCTURAL AND MAGNETORESISTIVE PROPERTIES OF MAGNETRON-SPUTTERED THIN Ni-Fe FILMS* K. SOLT Research and Development, LGZ Lan&s und Gyr Zug A G, CH-6301 Zug (Swttzerland) (Received August 13, 1984, accepted October 12, 1984)
The resistivity, magnetoresistance ratio, magnetic hysteresis loop and structure of magnetron-sputtered permalloy (Nis 1Fel 9) films of thickness 200-1600 A were studied and compared with the corresponding properties of permalloy films prepared by electron beam evaporation. The r.f.-sputtered films were prepared in an argon plasma from a sputter gun with an aligning magnetic field of 400 Oe applied in the plane of the substrates. No external heating or substrate bias voltage was used during the deposition. The resistivities of sputtered permalloy films show a similar thickness dependence to that for evaporated films, although with higher values. The residual resistances of the sputtered films are also higher; this is attributed to the greater amount of incorporated gases and possibly less atomic ordering. From the thickness dependence of the resistivity as measured in the range 200-1600 A, the value of the boundary reflection coefficient was estimated and was found to be almost the same as that reported earlier for permalloy films deposited in an ultrahigh vacuum. X-ray diffraction techniques were employed to measure the intrinsic isotropic stress ¢r which was found to be about 6 × 109 dyn cm -2. The magnetoresistance ratio of the sputtered films is less than 2~, which might be explained by their higher resistivity values. The coercivity of the films is in the range 1.5-30e, which is higher than that of evaporated permalloy films. The same was true of the anisotropy field, which is between 5 and 6.50e.
l. INTRODUCTION The application of low coercivity magnetoresistive permalloy thin films as magnetic field detectors for magnetic bubble devices I or as miniaturized magnetic reading heads 2 has increased in importance during recent years. The magnetoresistive switching and the micromagnetic domain structure (which becomes important when the size of the magnetoresistive element is decreased to micrometre *Paper presentedat the Sixth InternationalConferenceon Thin Fdms, Stockholm,Sweden,August 13-17, 1984. 0040-6090/85/$3.30
© ElsevierSequola/Pnntedm The Netherlands
252
K. SOLT
dimensions) are both sensitive to structure and to impurities and are thus strongly dependent on the conditions of the film deposition. In this investigation Nis 1Fe 19 films were prepared by r.f. magnetron sputtering and were compared with evaporated films of the same thickness and composition. First, the dependence of the resistivity and magnetoresistance on the film thickness was studied. Then, the intrinsic stresses and the dependence of the magnetic properties on the film thickness were measured and compared with those of evaporated films. The effect of annealing on these properties was also investigated. 2.
EXPERIMENTAL PROCEDURE
The films used in our experiments were deposited onto thermally oxidized silicon wafers and were 200-1600 ~ thick. They were prepared in a diffusionpumped vacuum system with a base pressure in the low 10-7 Torr region, using a Sloan S-310 r.f. magnetron sputter gun in an argon plasma of pressure 7-10 mTorr. The composition of the target was 8 l~oNi-19~Fe. The films were deposited in an aligning magnetic field of 400 Oe applied in the plane of the substrates at a deposition rate of 70/~ min- 1. The magnetic holder was situated on a rotating substrate holder 60 mm from the target. No external cooling or heating was employed during the deposition. After deposition the permalloy films were annealed in a vacuum in the low 10- 7 Torr range at 250 °C for 3 h. During the annealing an aligning magnetic field of 400 Oe was applied in the plane of the films. The resistivity of the permalloy films was measured with the four-point method. The magnetoresistance measurements were made on strips 1 mm wide and 25 mm long using the method described by Collins and Sanders 3. The intrinsic isotropic stress tr was determined using X-ray diffraction techniques 4 from the equation 1
a = ~Ee =
2
E
flc°s0-D-li 4s~n0
where E is Young's modulus, e is the strain, 0 is the Bragg angle, 2 is the radiation wavelength, fl is the diffraction line halfwidth and D j_ is the grain size normal to the film surface. On the assumption of columnar growth ofgrains, Di is equal to the film thickness d. The magnetic properties of the film, i.e. the coercivity, the anisotropy field and the magnetization dispersion, were measured on samples of dimensions 20 mm x 20 mm using a B - H looper (B, induction; H, applied magnetic field). 3. RESULTS 3.1. Resistivity and magnetoresistance The resistivity of the sputtered permalloy films before and after annealing as a function of the film thickness is shown in Fig. 1. Resistivity values of our evaporated films of similar compositions, but deposited at an elevated substrate temperature, are also presented. The sputtered films show higher resistivity values (though reducible with post-deposition annealing) and greater variation with thickness.
PROPERTIES OF MAGNETRON-SPUTTERED
Ni-Fe
FILMS
253
50-
\'\.,,.,,.
A
I,-
~
30
S 20 0
400 800 FILM THICKNESS
1200 (,~)
1600
Fig 1. The reststivlty as a function of film thickness ---, before anneahng, evaporated at 200 °C (from ref 5).
• after anneahng; - - - ,
The residual resistance p(4.2 K)/p(293 K) of the sputtered films was also measured and found to be about 25% higher than those of evaporated films. From the dependence of the resistivity on the film thickness the boundary reflection coefficient was estimated using the approximate equations of Mola and Heras 6, which they derived from the Mayadas-Shatzkes model for the conductivity in thin polycrystalline films 7. Similar estimations were made earlier by Chapman et al. s for the boundary reflection coefficient of thin permalloy films evaporated under ultrahigh vacuum conditions. We found that for the non-annealed sputtered permalloy films the model is not applicable; however, the boundary reflection coefficient of annealed films proved to be 0.26, which is very near to the value ofR = 0.2 reported by Chapman et al. The magnetoresistance ratio is shown in Fig. 2 as a function of the film thickness. Since Ap does not depend on the thickness, the increase in Ap/p is due to the lower resistivity of the thicker films. Also the lower value of Ap/p for sputtered films might result from their higher resistivity. 3.2. Intrinsic stress o f the films
The instrinsic isotropic stress of sputtered permalloy films before and after annealing is presented in Fig. 3 together with the values reported earlier by Prutton 9 for films evaporated in an ordinary vacuum and by Chapman et al. s for films evaporated in an ultrahigh vacuum. The intrinsic stress values may not be very accurate since the columnar growth assumption used for the calculations is crude considering the deposition temperature used. However, they can give information about the changes in stresses due to the post-deposition heat treatments. The results show that the intrinsic stress in the films is reduced by annealing in vacuum. 3.3. Magnetic properties
The thickness dependences of the anisotropy field and coercivity of films before and after annealing are shown in Fig. 4. Both these quantities remain constant with
254
o
i<. SOLT
3
~10,
£
6 ~8
/,.s ,s
%. o
~,' S j j ~o 6 . W In
~4
n, o
U) 2.
Z
200
400
eoo
800
~6o
FILM THICKNESS ( , ~ )
SUBSTRATE TEMPERATURE
The magnetoreslstance ratio as a f u n c t ] o n o f film thickness' 200 °C (from ref 5) Fig 2
280 °C
, sputtered; - - - , evaporated at
F i g 3 Intrinsic lsotrop]c stress m sputtered permalloy films O, before anneahng, e , after anneahng, -- -, data from r e f 8, - - - , data from ref 9
15-
o
v
10
5
o
2bo 4'00 FILM
8'00
12'oo
16'o0
THICKNESS ( ~ )
F i g 4 The coerovlty and the amsotropy field as functions of the film thickness' O, before anneahng, e , after anneahng
increasing film thickness after annealing. A hysteresis loop of a sputtered permalloy film taken after annealing is shown in Fig. 5. 4. CONCLUSIONS The resistivity of sputtered permalloy films shows a similar thickness dependence to that of evaporated films. The higher resistivity and residual resistance values can be attributed to the higher amount of incorporated gases and possibly less atomic ordering. The magnetoresistance ratio of the sputtered films is 2% which might be explained as due to their higher resistivity values. The coercivity of the films is in the range 1.5-30e, which is higher than that of evaporated permalloy films. The same
PROPERTIES OF MAGNETRON-SPUTTEREDN i - F e FILMS
255
(a)
(b) Fig. 5. The hysteresis loop of a sputtered permalloy film, showing the amsotroplc behavlour of the magnetic films (/arc= 1 4 Oe; Hk = 5 1 Oe; ct = 2~): (a) rectangular along the easy axis; (b) linear and nearly closed along the hard axis was true of the anisotropy field, which is between 5 and 6 . 5 0 e . Both the coercivity and the anlsotropy field are independent of the film thickness in the range 2 0 0 1600 A. We found that the intrinsic stress in the films is reduced by heat treatment in vacuum. This observation is consistent with the contraction of the crystal lattice which is detectable by X-ray measurements. Both can possibly be explained by the diffusion of the incorporated gases into the grain boundaries, where compressive stresses can be set up 1°. The same mechanism can a c c o u n t for the decrease in resistivity by annealing since the resistivity of the films is found to be lower when the gaseous impurities are accumulated at the grain boundaries than when they are built into the grain lattice 11. ACKNOWLEDGMENTS The a u t h o r is indebted to G. Schneider for providing the magnetic measurements, to J.-M. Triscone for the low temperature measurements and to S. Veprek for carrying out the X-ray diffraction measurements and for valuable discussions. T h a n k s are due to H. Lienhard for continuously encouraging this work. REFERENCES 1 A.H. Eschenfelder, in H.-J. Quelsser (ed.), Magnetic Bubble Technology, Vol. 14, Solid State Sctences, Springer, Berlin, 1980, Chapter 4. 2 J -L Berchler, K Solt and T. Zajc, J. Appl. Phys, 55 (1984) 487.
256
K. SOLT
3 A J Colhnsandl J Sanders, ThmSohdFdms, 48 (1978) 247. 4 A Taylor, m X-Ray Metallography, Wiley, New York 5 K Solt, m J L de Segovla (ed.), Proc. 9th Int. Vacuum Congr. and5th Int. Conf. on Sohd Surfaces. Madrid, 1983, ass. Espanole del Vaclo, Madrid, 1983, p 122. 6 E E M o l a a n d J M. Heras, ThmSohdFilms, 18(1973) 137. 7 A F MayadasandM Shatzkes, Phys Rev. B, 1(1969) 1382. 8 U B Chapman, A J. ColhnsandR D. Garwood, Thm Solid Films, 89 (1982) 243. 9 M Prutton, Nature (London), 193 (1962) 565 10 R W Hoffman, Thin SohdFtlms, 34 (1976) 185 11 A Kubovy and M Janda, Thm SohdFdms, 42 (1977) 169
251
MAGNETIC, STRUCTURAL AND MAGNETORESISTIVE PROPERTIES OF MAGNETRON-SPUTTERED THIN Ni-Fe FILMS* K. SOLT Research and Development, LGZ Lan&s und Gyr Zug A G, CH-6301 Zug (Swttzerland) (Received August 13, 1984, accepted October 12, 1984)
The resistivity, magnetoresistance ratio, magnetic hysteresis loop and structure of magnetron-sputtered permalloy (Nis 1Fel 9) films of thickness 200-1600 A were studied and compared with the corresponding properties of permalloy films prepared by electron beam evaporation. The r.f.-sputtered films were prepared in an argon plasma from a sputter gun with an aligning magnetic field of 400 Oe applied in the plane of the substrates. No external heating or substrate bias voltage was used during the deposition. The resistivities of sputtered permalloy films show a similar thickness dependence to that for evaporated films, although with higher values. The residual resistances of the sputtered films are also higher; this is attributed to the greater amount of incorporated gases and possibly less atomic ordering. From the thickness dependence of the resistivity as measured in the range 200-1600 A, the value of the boundary reflection coefficient was estimated and was found to be almost the same as that reported earlier for permalloy films deposited in an ultrahigh vacuum. X-ray diffraction techniques were employed to measure the intrinsic isotropic stress ¢r which was found to be about 6 × 109 dyn cm -2. The magnetoresistance ratio of the sputtered films is less than 2~, which might be explained by their higher resistivity values. The coercivity of the films is in the range 1.5-30e, which is higher than that of evaporated permalloy films. The same was true of the anisotropy field, which is between 5 and 6.50e.
l. INTRODUCTION The application of low coercivity magnetoresistive permalloy thin films as magnetic field detectors for magnetic bubble devices I or as miniaturized magnetic reading heads 2 has increased in importance during recent years. The magnetoresistive switching and the micromagnetic domain structure (which becomes important when the size of the magnetoresistive element is decreased to micrometre *Paper presentedat the Sixth InternationalConferenceon Thin Fdms, Stockholm,Sweden,August 13-17, 1984. 0040-6090/85/$3.30
© ElsevierSequola/Pnntedm The Netherlands
252
K. SOLT
dimensions) are both sensitive to structure and to impurities and are thus strongly dependent on the conditions of the film deposition. In this investigation Nis 1Fe 19 films were prepared by r.f. magnetron sputtering and were compared with evaporated films of the same thickness and composition. First, the dependence of the resistivity and magnetoresistance on the film thickness was studied. Then, the intrinsic stresses and the dependence of the magnetic properties on the film thickness were measured and compared with those of evaporated films. The effect of annealing on these properties was also investigated. 2.
EXPERIMENTAL PROCEDURE
The films used in our experiments were deposited onto thermally oxidized silicon wafers and were 200-1600 ~ thick. They were prepared in a diffusionpumped vacuum system with a base pressure in the low 10-7 Torr region, using a Sloan S-310 r.f. magnetron sputter gun in an argon plasma of pressure 7-10 mTorr. The composition of the target was 8 l~oNi-19~Fe. The films were deposited in an aligning magnetic field of 400 Oe applied in the plane of the substrates at a deposition rate of 70/~ min- 1. The magnetic holder was situated on a rotating substrate holder 60 mm from the target. No external cooling or heating was employed during the deposition. After deposition the permalloy films were annealed in a vacuum in the low 10- 7 Torr range at 250 °C for 3 h. During the annealing an aligning magnetic field of 400 Oe was applied in the plane of the films. The resistivity of the permalloy films was measured with the four-point method. The magnetoresistance measurements were made on strips 1 mm wide and 25 mm long using the method described by Collins and Sanders 3. The intrinsic isotropic stress tr was determined using X-ray diffraction techniques 4 from the equation 1
a = ~Ee =
2
E
flc°s0-D-li 4s~n0
where E is Young's modulus, e is the strain, 0 is the Bragg angle, 2 is the radiation wavelength, fl is the diffraction line halfwidth and D j_ is the grain size normal to the film surface. On the assumption of columnar growth ofgrains, Di is equal to the film thickness d. The magnetic properties of the film, i.e. the coercivity, the anisotropy field and the magnetization dispersion, were measured on samples of dimensions 20 mm x 20 mm using a B - H looper (B, induction; H, applied magnetic field). 3. RESULTS 3.1. Resistivity and magnetoresistance The resistivity of the sputtered permalloy films before and after annealing as a function of the film thickness is shown in Fig. 1. Resistivity values of our evaporated films of similar compositions, but deposited at an elevated substrate temperature, are also presented. The sputtered films show higher resistivity values (though reducible with post-deposition annealing) and greater variation with thickness.
PROPERTIES OF MAGNETRON-SPUTTERED
Ni-Fe
FILMS
253
50-
\'\.,,.,,.
A
I,-
~
30
S 20 0
400 800 FILM THICKNESS
1200 (,~)
1600
Fig 1. The reststivlty as a function of film thickness ---, before anneahng, evaporated at 200 °C (from ref 5).
• after anneahng; - - - ,
The residual resistance p(4.2 K)/p(293 K) of the sputtered films was also measured and found to be about 25% higher than those of evaporated films. From the dependence of the resistivity on the film thickness the boundary reflection coefficient was estimated using the approximate equations of Mola and Heras 6, which they derived from the Mayadas-Shatzkes model for the conductivity in thin polycrystalline films 7. Similar estimations were made earlier by Chapman et al. s for the boundary reflection coefficient of thin permalloy films evaporated under ultrahigh vacuum conditions. We found that for the non-annealed sputtered permalloy films the model is not applicable; however, the boundary reflection coefficient of annealed films proved to be 0.26, which is very near to the value ofR = 0.2 reported by Chapman et al. The magnetoresistance ratio is shown in Fig. 2 as a function of the film thickness. Since Ap does not depend on the thickness, the increase in Ap/p is due to the lower resistivity of the thicker films. Also the lower value of Ap/p for sputtered films might result from their higher resistivity. 3.2. Intrinsic stress o f the films
The instrinsic isotropic stress of sputtered permalloy films before and after annealing is presented in Fig. 3 together with the values reported earlier by Prutton 9 for films evaporated in an ordinary vacuum and by Chapman et al. s for films evaporated in an ultrahigh vacuum. The intrinsic stress values may not be very accurate since the columnar growth assumption used for the calculations is crude considering the deposition temperature used. However, they can give information about the changes in stresses due to the post-deposition heat treatments. The results show that the intrinsic stress in the films is reduced by annealing in vacuum. 3.3. Magnetic properties
The thickness dependences of the anisotropy field and coercivity of films before and after annealing are shown in Fig. 4. Both these quantities remain constant with
254
o
i<. SOLT
3
~10,
£
6 ~8
/,.s ,s
%. o
~,' S j j ~o 6 . W In
~4
n, o
U) 2.
Z
200
400
eoo
800
~6o
FILM THICKNESS ( , ~ )
SUBSTRATE TEMPERATURE
The magnetoreslstance ratio as a f u n c t ] o n o f film thickness' 200 °C (from ref 5) Fig 2
280 °C
, sputtered; - - - , evaporated at
F i g 3 Intrinsic lsotrop]c stress m sputtered permalloy films O, before anneahng, e , after anneahng, -- -, data from r e f 8, - - - , data from ref 9
15-
o
v
10
5
o
2bo 4'00 FILM
8'00
12'oo
16'o0
THICKNESS ( ~ )
F i g 4 The coerovlty and the amsotropy field as functions of the film thickness' O, before anneahng, e , after anneahng
increasing film thickness after annealing. A hysteresis loop of a sputtered permalloy film taken after annealing is shown in Fig. 5. 4. CONCLUSIONS The resistivity of sputtered permalloy films shows a similar thickness dependence to that of evaporated films. The higher resistivity and residual resistance values can be attributed to the higher amount of incorporated gases and possibly less atomic ordering. The magnetoresistance ratio of the sputtered films is 2% which might be explained as due to their higher resistivity values. The coercivity of the films is in the range 1.5-30e, which is higher than that of evaporated permalloy films. The same
PROPERTIES OF MAGNETRON-SPUTTEREDN i - F e FILMS
255
(a)
(b) Fig. 5. The hysteresis loop of a sputtered permalloy film, showing the amsotroplc behavlour of the magnetic films (/arc= 1 4 Oe; Hk = 5 1 Oe; ct = 2~): (a) rectangular along the easy axis; (b) linear and nearly closed along the hard axis was true of the anisotropy field, which is between 5 and 6 . 5 0 e . Both the coercivity and the anlsotropy field are independent of the film thickness in the range 2 0 0 1600 A. We found that the intrinsic stress in the films is reduced by heat treatment in vacuum. This observation is consistent with the contraction of the crystal lattice which is detectable by X-ray measurements. Both can possibly be explained by the diffusion of the incorporated gases into the grain boundaries, where compressive stresses can be set up 1°. The same mechanism can a c c o u n t for the decrease in resistivity by annealing since the resistivity of the films is found to be lower when the gaseous impurities are accumulated at the grain boundaries than when they are built into the grain lattice 11. ACKNOWLEDGMENTS The a u t h o r is indebted to G. Schneider for providing the magnetic measurements, to J.-M. Triscone for the low temperature measurements and to S. Veprek for carrying out the X-ray diffraction measurements and for valuable discussions. T h a n k s are due to H. Lienhard for continuously encouraging this work. REFERENCES 1 A.H. Eschenfelder, in H.-J. Quelsser (ed.), Magnetic Bubble Technology, Vol. 14, Solid State Sctences, Springer, Berlin, 1980, Chapter 4. 2 J -L Berchler, K Solt and T. Zajc, J. Appl. Phys, 55 (1984) 487.
256
K. SOLT
3 A J Colhnsandl J Sanders, ThmSohdFdms, 48 (1978) 247. 4 A Taylor, m X-Ray Metallography, Wiley, New York 5 K Solt, m J L de Segovla (ed.), Proc. 9th Int. Vacuum Congr. and5th Int. Conf. on Sohd Surfaces. Madrid, 1983, ass. Espanole del Vaclo, Madrid, 1983, p 122. 6 E E M o l a a n d J M. Heras, ThmSohdFilms, 18(1973) 137. 7 A F MayadasandM Shatzkes, Phys Rev. B, 1(1969) 1382. 8 U B Chapman, A J. ColhnsandR D. Garwood, Thm Solid Films, 89 (1982) 243. 9 M Prutton, Nature (London), 193 (1962) 565 10 R W Hoffman, Thin SohdFtlms, 34 (1976) 185 11 A Kubovy and M Janda, Thm SohdFdms, 42 (1977) 169