Electrochemically induced switching in SmHx thin films

Pushpendra Kumar, L.K. Malhotra*

Thin Film Laboratory, Department of Physics, Indian Institute of Technology, Hauz Khas, New Delhi 110016, India

Received in revised form 19 May 2004; accepted 21 May 2004

Available online 4 July 2004

Abstract

Discovery of optical and electrical switching in yttrium and lanthanum hydride thin films in 1996, has triggered large scale

investigations of rare earth hydride thin films. Both light and heavy rare earths have been shown to exhibit this so called

‘switchable mirror’ effect. Besides gas phase loading, solid as well as liquid electrolytes have been used as means of hydrogen

loading. In this paper, we report on the ability of samarium films to switch between the reflecting and transparent states

electrochemically by altering the hydrogen content. Samarium films of typical thickness 55 nm deposited by vacuum

evaporation and covered with Pd overlayer of thicknesses 5, 8 and 11 nm were loaded with hydrogen and deloaded in a

1 M KOH solution galvanostatically at room temperature. Our studies have shown that thickness of palladium overlayer plays

the most crucial role in observing fast transition between the as deposited metallic and the semiconducting nearly samarium

trihydride states as well as to obtain a very high optical contrast. The large reversible change in transparency between di and

trihydride states suggests possibilities of use of samarium films as an optical shutter as well as a hydrogen sensor material.

# 2004 Elsevier B.V. All rights reserved.

Keywords: Samarium hydrogenation; Palladium; Electrochemical behavior; Metal to semiconductor transition; Optical switching

1. Introduction

In the continuing research effort on rare earth

hydrides (REHx), the discovery of switchable mirror

effect by Huiberts et al. [1] occupies a special place.

By subjecting a 500 nm thin film of yttrium capped

with 20 nm Pd overlayer to 105 Pa of hydrogen gas, a

metal to semiconductor transition accompanied by

drastic changes in optical properties was observed.

This switchable mirror effect has since received a lot

of attention both from fundamental physics [2–4]

point of view as well as its technological applications

in smart windows, hydrogen sensors, solid state dis-

plays and electrochromic devices [5,6]. It is now

known that almost all rare earth metals and rare earth

based alloys exhibit this effect [7,8]. Hydrogenation of

these switchable mirrors can be achieved either by

exposing the sample to a hydrogen gas atmosphere or

electrochemically by polarization in a suitable elec-

trolyte solution [9–11]. Because of the hazards

involved in handling hydrogen gas, and the difficulty

in continuously monitoring the amount of hydrogen

getting incorporated into the films, gas phase loading

is not very practical from the application point of view.

These issues can be taken care of by hydrogenation via

electrochemical means. However, there are some

questions about cycling durability in the electroche-

mical process due to high reactivity of the metals and

their hydrides towards electrolyte solution and the

resulting enhanced corrosion. The cyclic durability

Applied Surface Science 236 (2004) 461–466

* Corresponding author. Tel.: þ91 1126 591325;

fax: þ91 1126 581114.

E-mail address: lalit@physics.iitd.ac.in (L.K. Malhotra).

0169-4332/$ – see front matter # 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.apsusc.2004.05.140

of switchable mirrors in KOH solution has been

investigated in detail by Janner et al. [12] by using

different sample stacks and growth conditions. The

electrolyte (water, ethanol and ethylene glycol) com-

patibility with these switchable mirrors for their dur-

ability and proton conductivity has been discussed in

detail by Matveeva et al. [13]. The role of Pd overlayer

has also been crucial in controlling the hydrogen

kinetics and stability in all the metal hydride thin film

devices. In this paper, we report on the ability of

samarium films to switch between the reflecting (gray)

and transparent (golden-greenish) states electroche-

mically and the effect of Pd overlayer on switching

behavior of SmHx thin films and their cyclic durability.

2. Experimental

Sm films of thickness 55 nm were deposited by

vacuum evaporation on 40 mm � 40 mm � 2 mm

ultrasonically cleaned glass substrates. The base pres-

sure in the vacuum system was 7 � 10�5 Pa prior to

deposition. Pd overlayers of different thicknesses 5, 8

and 11 nm were subsequently deposited on top of the

Sm films without breaking the vacuum. The geometry

shown in Fig. 1(a) was obtained by using a mask

during the evaporation process. The experimental

details for in situ measurement of film thickness

and rate of deposition have been described in one

of our earlier paper [14]. Two contacts A and B

(Fig. 1(a)) were then made for electrical measure-

ments using copper wires and silver paste. The elec-

trochemical measurements were performed in an

aqueous 1 M KOH solution using a Pt strip as a

counter electrode and an Hg/HgO electrode as a

reference electrode. Fig. 1(b) shows the experimental

setup used for electrochemical loading. The circular

portion of palladium capped samarium films was

immersed in the electrolyte, whereas the contacts were

kept in the atmosphere. The effective area of palla-

dium capped samarium films exposed to 1 M KOH

electrolyte solution was 4.9 cm2. All potentials were

measured with respect to the reference electrode using

a electrometer (Keithley, Model-6517A) and a con-

stant current source (Keithley, Model-224) was used to

apply a current. For in situ optical transmission, the

Pd-capped Sm film-working electrode (WE) was illu-

minated with a diode laser (Newport Corporation,

USA, LA 12-10—650 nm) and the transmission inten-

sity measured with a Photodyne radiometer/photo-

meter (Photodyne Inc., USA, Model 88XLA),

placed on the opposite end. A digital multimeter

was used to monitor the resistance. Before each

measurement, high purity argon gas was bubbled

through the solution at least for 15 min to remove

the oxygen and a constant argon flow was maintained

over the electrolyte during the measurements.

3. Results and discussion

Charging (loading) and discharging (unloading) of

hydrogen into/from Pd-capped Sm films (used as

working electrode in the electrolytic cell) was

achieved by polarizing/depolarizing (cathodically/

anodically) the working electrode, respectively.

Fig. 2(a)–(c) shows the development of potential on

charging and discharging of the working electrode and

Fig. 1. (a) Sample geometry and (b) electrochemical setup.

462 P. Kumar, L.K. Malhotra / Applied Surface Science 236 (2004) 461–466

the corresponding, in situ measured variation of trans-

mittance for 55 nm Sm films covered with Pd over-

layers of thicknesses 5, 8 and 11 nm, respectively. It is

evident that the potential falls sharply in each case

during cathodic polarization of the working electrode.

This fall in potential is indicated by the portion AB in

Fig. 2(a)–(c). The corresponding changes in the trans-

mittance are also shown in the same figures. The

metallic gray color of the film (corresponding to point

A in the figure) changes to the golden-greenish color

(point B). This change in color represents the trans-

formation of the samarium films from the metallic

state to the nearly trihydride state as also confirmed by

measurement of resistance. The switching time

(defined as the time required to reach 90% of the

maximum transmittance) varies with the Pd overlayer

thickness. The switching times obtained with three Pd

overlayers of thicknesses 5, 8 and 11 nm were found to

be 38, 30 and 22 s, respectively. We observed bubble

formation during cathodic polarization of working

electrodes while working with Sm films covered with

Pd overlayers of thicknesses 5 and 8 nm but no bubble

formation was observed for 11 nm Pd overlayer. The

presence and absence of bubble formation on working

electrode can be understood as follows.

During electrochemical loading, the electrolytic

reduction of a proton donating species, water in our

case, results in the following reaction:

H2O þ e� ! Had þ OH�

The mechanism of hydrogen entry into palladium

involves proton discharge H þ e� ! H�. The H

atoms subsequently diffuse into the underlying Sm

film and the reaction of hydrogen with Sm proceeds as

follows:

Sm þ 32

H2 ! SmH2 þ 12

H2 , SmH3

The second step is a reversible transition, which can

easily be induced by changing the polarity of the cell

or in open circuit condition, whereas the first step is

unidirectional. This is because of the relative small

heat of formation for the second step (�39.6 kJ/mol

H) compared to the heat of formation for the first step

(�202.6 kJ/mol H) [15]. The experimental evidence

for this reaction comes from the X-ray diffraction data

and change in transmittance as reported in one of our

earlier papers [16].

Since the surface energy of Pd cap layer is higher

than that of any rare earth metal [17], the nucleation

and growth theories suggest an activation barrier to the

nucleation of the condensed phase [18]. Pd will there-

fore deposit in the form of electrically disconnected

small clusters. At low thicknesses, these clusters do

not cover the entire surface area of the underlying Sm

films. Therefore, between two clusters oxide forma-

tion takes place. Hydrogen can only diffuse through

Fig. 2. Development of electrode potential E and in situ measured

optical transmission as a function of time for 55 nm Sm film

(working electrode) covered with different thicknesses of Pd

overlayer: (a) 5 nm; (b) 8 nm; (c) 11 nm during galvanostatic

hydrogen loading and unloading at current density 3.057 mA/cm2

in 1 M KOH.

P. Kumar, L.K. Malhotra / Applied Surface Science 236 (2004) 461–466 463

the Pd clusters but not through oxide covered areas.

Thus, all the H atoms liberated due to proton discharge

cannot diffuse into the Sm films. Since they can also

not exist in free state, two of them pair together

immediately to form H2 molecules, which are liber-

ated as hydrogen gas at the cathode. The bubble

formation in 5 and 8 nm Pd overlayer may be attrib-

uted to the excess hydrogen evolution at the working

electrode, the amount of hydrogen evolution exceed-

ing that getting incorporated in samarium films. Our

results suggest that for Pd overlayer of thickness

11 nm, the rate of evolution and absorption of hydro-

gen may be equal and this may be the reason for the

absence of bubble formation on working electrode.

The portion B to C in the transmittance and poten-

tial curves indicates no change in transmittance and

potential after the formation of nearly trihydride state

even though current density was maintained at the

initial value of 3.057 mA/cm2.

At point C, the films were anodically discharged by

changing the polarity. Comparison of transmittance

curves in Fig. 2(a)–(c) shows that the discharging

(unloading) is very slow in films covered with Pd

overlayer of thickness 5 nm. The rate of discharging

increases with the increase in the Pd overlayer thick-

ness. It has been theoretically predicted and experi-

mentally verified that the geometry and the electronic

band structure of the Pd clusters vary with cluster size

and they show corresponding change in their proper-

ties, for example, these clusters cease to show rever-

sible hydrogen deloading capacity [19–21] but they

continue to show catalytic hydrogen dissociation

behavior. This explains slow discharge of hydrogen

in Sm films capped with low thickness of Pd overlayer.

The reversibility between the charged and discharged

states also becomes better with increase in Pd layer

thickness. However, even at higher thickness of Pd, the

hydrogen from the charged electrode cannot be

removed completely by anodic discharging [12]. This

is consistent with the reported observations that rever-

sibility can only occur between dihydride and trihy-

dride states and that one cannot remove hydrogen from

the dihydride state. No degradation of the films was

observed even after 40 cycles of charging (loading)/

discharging (unloading) in 11 nm Pd covered film and

there was no trace of any bubble formation. The

reversible switching between dihydride and trihydride

states occurred within 6 s and was reproducibly

observed for 40 cycles. Fig. 3 shows the change in

transmittance after 40 cycles of loading and unloading

of hydrogen between dihydride and trihydride state for

the film covered with Pd (11 nm) overlayer.

In all the three cases discussed above, the fall in

potential is very high. This is due to the high applied

current density between working and counter elec-

trode. If current density is high, then the rate of

evolution of hydrogen ions at the electrode surface

will be high, resulting in a high chemical potential

generated in the solution. The overpotential

Z ¼ E � Er related to current density j is governed

by the Butler–Volmer equation [22]:

j

j0

¼ exp ð1 � aÞ nF

RTZ

� �� exp �a

nF

RTZ

� �

where j0 is the exchange current density, a the transfer

coefficient, and n the number of electrons involved in

the charge-transfer density reaction. E and Er repre-

sent the potential measured when a current density j is

passed across the WE–electrolyte interface and the

reversible potential corresponding to the system at

equilibrium with no current passing, respectively.

At high positive values of the over-potential, when

ðð1 � aÞnFZ=RTÞ 1, the second term on the right-

hand side of the above equation becomes negligible

with respect to the first one and the equation becomes:

j ¼ j0 exp ð1 � aÞ nF

RTZ

� �

This explains the direct correlation between Z and j.

0 50 100 150 200 25010

15

20

25

30

35

T (

%)

Time (sec)

Fig. 3. Change in transmittance with time after 40 cycles between

dihydride and trihydride states on loading and unloading of

hydrogen for the film covered with higher thickness of (11 nm) Pd

overlayer.

464 P. Kumar, L.K. Malhotra / Applied Surface Science 236 (2004) 461–466

Fig. 4 shows the variation of switching time with

applied current density. The switching time decreases

with increase in current density. The variation of

switching time with applied current density is related

to the equation [22]:

t1=2 ¼ p1=2nFD1=2C0

2j

where j is the applied current density, C0 the concen-

tration of hydrogen in the electrolyte and D the

diffusion coefficient. Our experimental results agree

with the variation suggested by this equation. The

decreased switching time can also be explained due to

the increased rate of reaction on the surface of the

working electrode, rate of reaction being high because

of a higher driving force at high current density.

Our studies show that in addition to thickness of Pd

overlayer, the interface between Sm and Pd films plays

an important role in controlling the switching time as

well as their chemical stability in the electrolyte and

cycle durability. We have found that the time duration

between deposition of Sm and Pd is an important

factor. In some cases, where the latter was not con-

trolled, no loading of hydrogen could be achieved.

Similar observations have also been reported earlier

[8].

4. Conclusions

The charging and discharging of 55 nm thick samar-

ium films covered with different thicknesses (5, 8 and

11 nm) of Pd overlayer, carried out at current density

of 3.057 mA/cm2 in 1 M KOH solution, reveal that the

thickness of Pd overlayer plays a crucial role in

controlling the switching time, chemical stability

and cycling durability. The minimum switching time

between dihydride and trihydride states was found to

be 6 s in Sm film covered with 11 nm thick Pd over-

layer. The current density has also been found to play a

crucial role in controlling the switching time. The

ability of samarium films to quickly switch between

the reflecting and transparent state electrochemically

by altering the hydrogen content, suggests potential

applications in optical windows as well as hydrogen

sensors.

References

[1] J.N. Huiberts, R. Griessen, J.H. Rector, R.J. Wijngaarden, J.P.

Dekker, D.G. de Groot, N.J. Koeman, Nature 380 (1996) 231.

[2] R. Eder, H. Pen, F.A. Sawatzky, Phys. Rev. B 56 (1997)

10115.

[3] P.J. Kelly, J.P. Dekker, R. Stumpf, Phys. Rev. Lett. 78 (1997)

1315.

[4] K.K. Ng, F.C. Zhang, V.I. Anisimov, T.M. Rice, Phys. Rev. B

59 (1999) 5398.

[5] J.W.J. Kerssemakers, S.J. van der Molen, R. Gunther, R.

Griessen, Nature 406 (2000) 489.

[6] R. Armitage, M. Rubin, T. Richardson, N. O’Brien, Y. Chen,

Appl. Phys. Lett. 75 (1999) 1863.

[7] K. von Rottkay, M. Rubin, P.A. Duine, J. Appl. Phys. 85

(1999) 408.

[8] K. von Rottkay, M. Rubin, F. Michalak, R. Armitage, T.

Richardson, J. Slack, P.A. Duine, Electrochim. Acta 44

(1999) 3093.

[9] F.J.A. den Broeder, S.J. van der Molen, M. Kremers, J.N.

Huiberts, D.G. Nagengast, A.T.M. van Gogh, W.H. Huisman,

N.J. Koeman, B. Dam, J.H. Rector, S. Plota, M. Haaksma,

R.M.N. Hanzen, R.M. Jungblut, P.A. Duine, R. Griessen,

Nature 394 (1998) 656.

[10] P.H.L. Notten, M. Kremers, R. Griessen, J. Electrochem. Soc.

143 (1996) 3348.

[11] M. Kremers, N.J. Koeman, R. Griessen, P.H.L. Notton, R.

Tolboom, P.J. Kelly, P.A. Duine, Phys. Rev. B 57 (1998) 4943.

[12] A.-M. Janner, P. van der Sluis, V. Mercier, Electrochim. Acta

46 (2001) 2173.

[13] E.S. Matveeva, R.J. Ortega Ramiro, A. Sanchez Bolinchez, C.

Ferrer Jimenez, V.P. Parkhutik, Sens. Actuators B 84 (2002)

83.

[14] P. Kumar, R. Philip, G.K. Mor, L.K. Malhotra, Jpn. J. Appl.

Phys. 41 (2001) 6023.

[15] M. Ouwerkerk, Solid State Ionics 431 (1998) 113.

[16] P. Kumar, L.K. Malhotra, Electrochim. Acta 49 (2004) 3355.

[17] A. Borgschulte, M. Rode, A. Jacob, J. Schoenes, J. Appl.

Phys. 90 (2001) 1147.

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

0

20

40

60

80

100

Sw

itch

ing

tim

e(s

ec)

Current density mA/cm2

Fig. 4. Variation of switching time with current density for a Pd

(11 nm) capped Sm (55 nm) film.

P. Kumar, L.K. Malhotra / Applied Surface Science 236 (2004) 461–466 465

[18] K.L. Chopra, in: Thin Film Phenomena, R.E. Krieger

Publishing Co., New York, 1979, Chapter 4.

[19] A. Zuttel, Ch. Nutzenadel, G. Schmid, Ch. Emmenegger, P.

Sudan, L. Schlapbach, Appl. Surf. Sci. 162 (2000) 571.

[20] I. Efremenko, J. Mol. Catal. A 173 (2001) 19.

[21] G.K. Mor, L.K. Malhotra, D. Bhattacharyya, J. Appl. Phys. 90

(2001) 1.

[22] E. Gileadi, E. Kirowa Eisner, J. Penciner, Interfacial

Electrochemistry: An Experimental Approach, Addison-

Wesley, 1975, pp. 47, 64.

466 P. Kumar, L.K. Malhotra / Applied Surface Science 236 (2004) 461–466