Professional Documents
Culture Documents
L. Adnane, F. Dirisaglik, A. Cywar, K. Cil, Y. Zhu, C. Lam, A. F. M. Anwar, A. Gokirmak, and H. Silva
setup is used to generate a high frequency alternating mag- GST behaves like a p-type semiconductor,13 and the
netic field that results in heating of the chuck through contact conductivity ðrÞ can be expressed as
with steel surrounding plates. The temperature gradient can
also be controlled with the inductive heater by moving the rðT Þ ¼ elp ðTÞpðTÞ; (1)
coil from one side to another. The surrounding steel plates
form an oven-like enclosure for heat confinement and heat- where e is the elementary charge, p is the carrier density, and
ing of the probes and probe arms to reduce the probe cooling lp is the low field hole mobility. Both the carrier density and
effect. the mobility are temperature dependent.
When a temperature gradient is set across the sample at a The measured resistance of the sample is scaled to resis-
given stabilized temperature, the I-V characteristics obtained tivity using the resistivity value of the sample measured at
with the semiconductor parameter analyzer simultaneously room temperature using the van der Pauw (vdP) method14
provide the Seebeck voltage, which is the x-axis intercept of and assuming that the geometry factor does not change with
the line, and the resistance between the two contacts, which temperature. The resistivity of a square shaped GST thin film
is the slope of the line (Fig. 3). sample with known film thickness, from the same wafer, is
In the experiments, the maximum temperature differ- measured at room temperature using four probes to contact
ence recorded between the probe arm and the surface of the the film surface [Fig. 5(a)]. According to the vdP technique,
chuck does not exceed 150 C at a chuck temperature of the resistivity of the sample is given by
600 C. The probe needles (Cascade Microtech PTT-24/ R14; 23 þ R12;43
4–25 tungsten needles) are 20 mm long, and its tip diame- q ¼ 4:532 tf ; (2)
2
ter is 2.5 lm. An approximated geometry of a GST sample
with electrical contact (2D cylindrically symmetric tungsten where t is the thickness of the film, R14,23 ¼ V14/I23, and
probe tip on sample tungsten bottom contact) is simulated in R12,43 ¼ V12/I43, and the values of the coefficient f are tabu-
COMSOL to estimate the effect of the cooling by the probe lated for arbitrary values of the resistance ratio R14,23/
arms via a 10 lm diameter tungsten contact on the surface of R12,43.14
the GST film (Fig. 1). A worst-case scenario is simulated by The measured resistivity of the 100 nm thick fcc-GST
setting the temperature of the top surface of a 1 lm thick film in a square shape [configuration in Fig. 5(a)], at room
tungsten contact on the sample surface at 200 C, with the temperature, previously heated up to 200 C, is 504 6 3
bottom surface temperature of the sample (surface of the mXcm (assuming a 5% error on the film thickness).
chuck) at 400 C. The simulations show that outside the A second sample from the same wafer, with 4 inline
small area where the probe tips come into contact (10 lm metal contacts, as shown in Fig. 5(b), is also heated up to
diameter), the temperature of the sample surface is the same
as the chuck temperature (Fig. 4).
FIG. 4. Numerical modeling of the cooling of the sample by the probes. (a)
Temperature of the sample simulated after 60 s sample heating to 400 C.
FIG. 6. GST resistivity versus temperature measurements on the 200 nm
(b) Temperature of the simulated probe tip region while the tip surface is
GST film using the 4-point in-line measurement, scaled from R(T) using the
maintained at 200 C. (c) Temperature on the surface of the sample across
vdP measurement at room temperature after heating and cooling down the
the metal contact.
sample from 200 C and from 300 C at the same rates.
200 C, with the same heating rate (5 C/min), to measure configuration shown in Fig. 5(c) is used for simultaneous
the R(T) characteristic during heating and cooling back to R(T) and S(T) measurements of two different samples during
room temperature. The resistance between the two inner con- the same run. The temperature difference DT ¼ Thot Tcold
tacts on the surface of the sample is given by is the temperature difference between the two probe loca-
tions, which is varied between 10 C and þ10 C at given
R ¼ V23 =I14 : (3)
stable average temperatures TAVG¼(Thot þ Tcold)/2.
The results obtained from the 2 previous measurements are
III. RESULTS AND DISCUSSION
used to determine the resistivity value of the film at 200 C
A. Continuous R(T) measurements
qroomT
q200 ¼ R200 : (4) Continuously increasing temperature measurements of
RroomT
R(T) were performed on 50 nm, 100 nm, and 200 nm (target
This procedure is repeated for a second anneal temperature thicknesses) GST thin films up to the melting temperature
(300 C) and the R(T) characteristic obtained while heating and scaled to resistivity q(T) using the room-temperature
matches the R(T) obtained during the previous cooling from value after anneal at 200 C as described above (Fig. 7).
200 C (Fig. 6). The obtained resistivities at 200 C and During these measurements, the temperature was increased at
300 C are 75 mXcm and 11 mXcm, respectively. a constant rate of 5 C/min. The two transitions, from amor-
The measured R(T) characteristics for different samples phous to fcc (155 C) and from fcc to hcp (365 C), corre-
are then scaled to resistivity (Fig. 7) using the resistivity spond to the turning points in the curves. The as-deposited
value at 200 C, which assumes the same geometry factor
measured at room temperature.
The sign of the Seebeck coefficient indicates the
majority carrier type (S > 0 for p-type conduction). The
FIG. 12. XRD patterns of the 100 nm GST film annealed at different temper-
atures and then cooled down from 400 C to obtain the pattern for hcp-GST
at 30 C. The peaks around 47.5 and 55 , present in all patterns, likely orig-
inate from the sample holder and are not related to the GST sample.
0:9k
gs ¼ ; (7)
b cos h
ðRa RÞðR þ 2Rc Þ TABLE I. Amorphous and crystalline resistance pre-factors and activation
f ¼ ; (10)
3RðRa Rc Þ energies extracted from the R(T) measurements.
where R, Ra, and Rc are the resistances of the mixed phase, 50 nm thin film 200 nm thin film
the amorphous phase, and the fully crystalline fcc GST,
Ra0 (X) 77.05 56.99
respectively. Ea (meV) 419.8 376.7
The resistance of the mixed phase material at each crys- Rc0 (X) 3558 1104
tallinity fraction f follows an Arrhenius dependence:
125104-8 Adnane et al. J. Appl. Phys. 122, 125104 (2017)
FIG. 16. Crystallinity fraction f for the 200 nm GST thin film sample as a
function of the cubic grain size (gs)3 obtained from XRD patterns for the
mixed phase amorphous-fcc region, pre-annealed at increasing anneal tem-
peratures (180 C to 320 C). The intermediate linear region suggests growth
dominated crystallization for this anneal temperature range.
IV. SUMMARY
Simultaneous measurements of the temperature depen-
FIG. 15. (a) Calculated crystalline fraction f and experimental dS/dT as a
dent electrical resistance and Seebeck coefficient of GST
function of anneal temperature for the 50 nm and 200 nm thin GST films. (b)
Experimental dS/dT (/ m/p2/3) and room-temperature conductivity rRT (/ (Ge2Sb2Te5) thin films were performed up to 540 C.
ms/p) versus calculated crystalline fraction f for 50 nm and 200 nm thin film Repeated measurements to increasingly maximum tempera-
samples. tures allow characterization of the properties of each state (in
contrast to continuous measurements which show the convo-
versus calculated crystalline fraction f for the 50 nm and
luted effects of the temperature dependence of the transport
200 nm thin film samples. These relations can now be used to
parameters and material crystallization during heating, as
estimate S(T) and r(T) for a given amorphous-fcc mixed phase
also shown here up to melting temperature). The measured
material from a single, room-temperature value of the conduc-
resistance was scaled to resistivity using the 4 point room-
tivity. rRT can be used to determine f and the corresponding
temperature resistivity measurement performed on a sample
dS/dT and hence S(T), which can then be used to determine the
of the same film. The R(T) characteristics measured at differ-
conduction activation energy E (Fig. 11) to obtain the full r(T).
ent crystalline states follow an Arrhenius dependence with a
Since the model for conductivity we have used assumes
decreasing activation energy as the material crystallizes. The
spherical crystals in an amorphous matrix and f is proportional
S(T) results show p-type conduction until melting tempera-
to (gs)3 times the number of the grains in the sample, it is
ture and linear S(T) characteristics for each state above glass
interesting to look at the relationship between f and (gs)3
transition temperature, in agreement with the degenerate
obtained from XRD (Fig. 16), which appears to show three
semiconductor transport model. The measurements also
distinct regions for crystallization:
show that both the mixed amorphous-fcc and the mixed
(1) 180 C < T < 220 C: the crystallinity of the sample fcc–hcp phases exhibit a constant positive Thomson coeffi-
increases rapidly with slow grain growth (Df/Dgs3 cient (dS/dT). A linear relationship between the slope of
¼ 1.1 104 nm3), suggesting that nucleation in the S(T) characteristics and the activation energy of the
material is dominant. Arrhenius conduction for the GST material is also observed
(2) 220 C < T < 300 C: f changes linearly with (gs)3 at a for the mixed amorphous-fcc phase. This observation can be
lower rate compared to the previous region (Df/ useful to estimate the temperature-dependent conductivity or
Dgs3¼5.1 105 nm3), suggesting that the number of Seebeck coefficient of GST from a single measurement of
crystals is approximately constant and the increase in the Seebeck coefficient or a few conductivity measurements
crystallinity is mainly related to the growth of the at low temperature points (to obtain the conduction activa-
crystals. tion energy). In-situ XRD measurements on a 100 nm
(3) 300 C < T < 320 C: the crystallinity again increases thin GST film sample showed the plane-dependent thermal
rapidly with a small increase in grain size (Df/ expansion of the hcp phase, with a maximum thermal expan-
Dgs3¼1.1 104 nm3), implying that the critical size sion observed for the (004) plane.
125104-9 Adnane et al. J. Appl. Phys. 122, 125104 (2017)
11
A percolation model of conducting spheres in a lower L. W.-W. Fang, Z. Zhang, R. Zhao, J. Pan, M. Li, L. Shi, T.-C. Chong,
and Y.-C. Yeo, J. Appl. Phys. 108, 53708 (2010).
conductivity matrix (which predicts the sharp drop observed 12
L. Adnane, N. Williams, H. Silva, and A. Gokirmak, Rev. Sci. Instrum.
in the resistance with a relatively small change in the grain 86, 105119 (2015).
13
size, at a critical crystalline density threshold) was applied to A. Pirovano, A. A. L. Lacaita, A. Benvenuti, F. Pellizzer, and R. Bez,
the measured data to relate the electrical transport parame- IEEE Trans. Electron Devices 51, 452 (2004).
14
L. J. Van der Pauw, Philips Tech. Rev. 20, 220–224 (1958).
ters and the crystalline grain sizes obtained from XRD meas- 15
K. Cil, F. Dirisaglik, L. Adnane, M. Wennberg, A. King, A. Faraclas, M.
urements. The R(T) results of mixed phase amorphous-fcc B. Akbulut, Y. Zhu, C. Lam, A. Gokirmak, and H. Silva, IEEE Trans.
are then expressed in terms of the amorphous and fully crys- Electron Devices 60, 433 (2013).
16
talline fcc temperature-dependent resistance values using the I. Friedrich, V. Weidenhof, W. Njoroge, P. Franz, and M. Wuttig, J. Appl.
Phys. 87, 4130 (2000).
effective-medium theory to obtain the temperature depen- 17
D. Fugazza, D. Ielmini, G. Montemurro, and A. L. Lacaita, in Proceedings
dent crystallinity fraction of the material. of the International Electron Devices Meeting (IEEE, 2010), pp.
This study focused on the properties of GST up to the 29.3.1–29.3.4.
18
C. David Wright, M. Armand, and M. M. Aziz, IEEE Trans. Nanotechnol.
second phase transition (fcc–hcp), which is the more techno-
5, 50 (2006).
logically relevant range as phase-change memory devices do 19
J. A. Kalb, M. Wuttig, and F. Spaepen, J. Mater. Res. 22, 748 (2007).
20
not experience the slow fcc–hcp transition during switching. N. F. Mott and E. A. Davis, Electronic Processes in Non-Crystalline
A similar percolation and effective medium theory model Materials (Clarednon Press, Oxford, 1971), p. 591.
21
F. Rao, Z. Song, Y. Cheng, M. Xia, K. Ren, L. Wu, B. Liu, and S. Feng,
can however be applied to the mixed fcc–hcp phase after the Acta Mater. 60, 323 (2012).
second transition for a fuller understanding of the crystalliza- 22
A. Redaelli, A. Pirovano, I. Tortorelli, D. Ielmini, and A. L. Lacaita, IEEE
tion dynamics and transport properties in chalcogenide Electron Device Lett. 29, 41 (2008).
23
glasses. H.-K. K. Lyeo, D. G. Cahill, B.-S. S. Lee, J. R. Abelson, M.-H. H. Kwon,
K.-B. B. Kim, S. G. Bishop, and B. Cheong, Appl. Phys. Lett. 89, 151904
(2006).
ACKNOWLEDGMENTS 24
A. Voraud, M. Horprathum, P. Eiamchai, P. Muthitamongkol, B.
Chayasombat, C. Thanachayanont, A. Pankiew, A. Klamchuen, D.
The GST films were deposited at IBM T.J. Watson Naenkieng, T. Plirdpring, A. Harnwunggmoung, A. Charoenphakdee, W.
Research Center and characterized at UConn. This work was Somkhunthot, and T. Seetawan, J. Alloys Compd. 649, 380 (2015).
25
partially supported by the U.S. National Science Foundation T. Siegrist, P. Jost, H. Volker, M. Woda, P. Merkelbach, C.
Schlockermann, and M. Wuttig, Nat. Mater. 10, 202 (2011).
through Award Nos. ECCS 0925973 and ECCS 1150960. 26
M. Cutler and J. F. Leavy, Phys. Rev. 133, A1153 (1964).
The characterization and analysis efforts of L.A. were 27
G. J. Snyder and E. S. Toberer, Nat. Mater. 7, 105 (2008).
28
supported by the U.S. Department of Energy, Office of J. P. Heremans, V. Jovovic, E. S. Toberer, A. Saramat, K. Kurosaki, A.
Science, Basic Energy Sciences under Award No. DE- Charoenphakdee, S. Yamanaka, and G. J. Snyder, Science (80-.) 321,
1457 (2008).
SC005038. L.A. also acknowledges a graduate fellowship 29
F. Dirisaglik, G. Bakan, A. Faraclas, A. Gokirmak, and H. Silva, Int. J.
from the Graduate Assistance in Areas of National Need High Speed Electron. Syst. 23, 1450004 (2014).
30
(GAANN). F.D. and K.C. acknowledge graduate fellowships A. Faraclas, G. Bakan, L. Adnane, F. Dirisaglik, N. E. Williams, A.
Gokirmak, and H. Silva, IEEE Trans. Electron Devices 61, 372 (2014).
from the Republic of Turkey Ministry of National Education. 31
K. Cil, Dr. Diss. (2015).
A.C. was supported through an NSF GRFP fellowship. The 32
D. Dimitrov and H.-P. Shieh, Mater. Sci. Eng. B 107, 107 (2004).
33
authors would also like to thank Simone Raoux, Norma Sosa, N. Kato, I. Konomi, Y. Seno, and T. Motohiro, Appl. Surf. Sci. 244, 281
and Matthew BrightSky for their contributions to sample (2005).
34
Y. Yin, H. Sone, and S. Hosaka, Jpn. J. Appl. Phys., Part 1 44, 6208
fabrication at the IBM T.J. Watson Research Center. (2005).
35
A. Voraud, M. Horprathum, P. Eiamchai, P. Muthitamongkol, C.
1
Z. Sun, S. Kyrsta, D. Music, R. Ahuja, and J. M. Schneider, Solid State Thanachayanont, W. Somkhunthot, and T. Seetawan, Surf. Coat. Technol.
Commun. 143, 240 (2007). 291, 15 (2016).
2 36
D. Ielmini, in Proceedings of the International Conference on K. S. Bang, Y. J. Oh, and S.-Y. Lee, J. Electron. Mater. 44, 2712 (2015).
Simulations of Semiconductor Processes and Devices (SISPAD) (IEEE, 37
C.-A. Jong, W. Fang, C.-M. Lee, and T.-S. Chin, Jpn. J. Appl. Phys., Part 1
2009), pp. 1–8. 40, 3320 (2001).
3 38
D. T. Castro, L. Goux, G. A. M. Hurkx, K. Attenborough, R. Delhougne, A. L. Patterson, Phys. Rev. 56, 978 (1939).
J. Lisoni, F. J. Jedema, M. A. A. In’t Zandt, R. A. M. Wolters, D. J. 39
J. I. Langford and A. J. C. Wilson, IUCr J. Appl. Crystallogr. 11, 102
Gravesteijn, M. A. Verheijen, M. Kaiser, R. G. R. Weemaes, and D. J. (1978).
40
Wouters, IEEE Int. Electron Devices Meet. 315 (2007). B. Cullity, Elements of X-ray Diffraction (Addison–Wesley, Reading, MA,
4
H.-S. P. Wong, S. Raoux, S. Kim, J. Liang, J. P. Reifenberg, B. Rajendran, 1967), p. 262.
41
M. Asheghi, and K. E. Goodson, Proc. IEEE 98, 2201 (2010). Y. Hu, H. Zou, J. Zhang, J. Xue, Y. Sui, W. Wu, L. Yuan, X. Zhu, S.
5
J. Lee, T. Kodama, Y. Won, M. Asheghi, and K. E. Goodson, J. Appl. Song, and Z. Song, Appl. Phys. Lett. 107, 263105 (2015).
42
Phys. 112, 14902 (2012). Y.-H. Huang, Y.-J. Huang, and T.-E. Hsieh, J. Appl. Phys. 111, 123706
6
T. Kato and K. Tanaka, Jpn. J. Appl. Phys., Part 1 44, 7340 (2005). (2012).
7 43
N. Ciocchini, M. Laudato, and A. Leone, IEEE Trans. Electron Devices L. Rayleigh, London Edinburgh Dublin Philos. Mag. J. Sci. 34, 481
62, 3264 (2015). (1892).
8 44
J. Lee, M. Asheghi, and K. E. Goodson, Nanotechnology 23, 205201 O. Wiener, Phys. Z. 5, 332 (1904).
45
(2012). D. A. G. Bruggeman, Ann. Phys. 416, 636 (1935).
9 46
R. R. Heikes and R. W. Ure, Thermoelectricity: Science and Engineering R. Landauer, J. Appl. Phys. 23, 779 (1952).
47
(Interscience Publishers, 1961). D.-H. Kim, F. Merget, M. Laurenzis, P. H. Bolivar, and H. Kurz, J. Appl.
10
G. S. Nolas, J. Sharp, J. Goldsmid, and H. J. Goldsmid, Thermoelectrics: Phys. 97, 83538 (2005).
48
Basic Principles and New Materials Developments (Springer Science and R. C. McPhedran and D. R. McKenzie, Proc. R. Soc. London, Ser. A 359,
Business Media, 2013). 45 (1978).