Indian Journal of ChemistryVol. 34A, February 1995, pp. 111-115

Effect of UV radiation on the growth and breakdown voltage of anodicoxide films on niobium

Sanjcev Makcn*. Mohindcr Singh. K C Kalra & K C Singh

Dcparrmcru of Chemistry. Maharshi Dayanand University, Rohtak 124001 (India)

Received 24 March 1994; revised 30 August 1994; accepted 13 September 1994

Formation rates of anodic Nb.O, films grown under galvanostatic conditions decrease in thepresence of UV radiation, unlike those grown in the absence of UV radiation. This may be due tothe development of a positive space charge near the solution/oxide interface which is responsible foran increase in electronic current in the film during its formation. Value of breakdown voltage alsoincreases in the presence of UV radiations. The effect of current density and resistivity of the solu-tion upon the breakdown voltage, both in the presente and absence of UV radiation, is discussed interms of Ikonopisov theory of breakdown voltage.

A considerable interest has been witnessed in therecent years in the studies of growth and electri-cal breakdown during the anodization of niob-ium!". The phenomenon is important both forelectrolytic capacitor industry and for anodizationin general. The formation rate of anodic oxidefilms on Ta5·7 and AlH is enhanced by UV radia-tion, which also causes permanent structuralchanges. These changes modify the properties ofphotogrown oxide films as compared to thosegrown in natural light. In order to see how oxida-tion of niobium, a typical valve metal, would beaffected in the presence of UV radiation, thepresent work has been carried out. In this paperthe results of ionic conduction and breakdownphenomena for anodic oxide films formed undergalvanostatic conditions in the presence and abs-ence of UV radiation are presented.

Materials and MethodsSpecimens of a~ea 2 cm ' were cut, using a die

from 0.25 mm thick sheet of 99.99% pure niob-ium (supplied by Aldrich). The method of thepreparation of the specimens, chemical polishingprocedure and the method of anodic oxidationwas the same as that described previously". Apolished specimen was then placed in a glass cell,having a platinum wire gauge counter electrodeand 0.1 N oxalic acid as the electrolyte. The cellwas then placed in a UV chemical reactor (Shrin-ivasan-Griffan Rayonet type, manufactured byGupta Scientific Industries, Ambala Cantt),

Anodic polarization of the specimen was carri-ed out at constant current, adjusted through an

electronically operated constant current generator(General Electronics, Ambala Cantt) designedspecially for this purpose. The method of anodicpolarization was same as described earlier", Thepolarization under galvanostatic conditions werecarried up to the breakdown voltage with ARgrade oxalic acid electrolyte at different currentdensities and concentrations in the presence andabsence of UV radiation. During the polarizationthe solution was stirred continuously with a mag-netic stirrer. The anodic and cathodic reactionsduring the polarization were

(a) atanode[Nb-Nb+:i+5e-]x22Nb+:i + lOOH- - Nb-O, + 5H~O

(b) at cathode lOH+ + lOe- -+ 5H, t

The thickness of the oxide film was establishedusing simple Faraday's law calculation. Accordingto this the thickness (D in em) of the film is calcu-lated by D = MQ!2zFAp (assuming z = 5, and100% current efficiency for the film) where p isdensity of the oxide film, A is the total area of thefilm in ern", M is the molar mass of film material,z is the valency of the film, Q is amount of chargepassed in coulombs and F is equal to 96488 cou-lombs. The density of Nb~O~ was taken to be4.36 gm cm -.' as suggested by Holtzberg et al.".The capacitance of the film in the dissolution ex-periments was measured using a capacitancebridge (Model CB lS4-D/3, BPL, India). Eachexperiment was repeated at least five times andthe breakdown voltage and dissolution data showa very good reproducibility.

112 INDIAN J CHEM. SEe. A FEBRUARY 1995

5.0mAcm2

100

III!::o>~

c:J'IV

~ 50>

0.1N Oxalic acidTemp '298.15K

• In presence of UVo In absence of UV

o 250 500 750 1000

Anodization time/seconds

Fig. 1 - Variation of volta~e of formation with time of anodization

Results and Discussion

Formation and dissolution rates of anodic oxidefilms in the presence and absence of UV radia-tion

Anodic polarization data are plotted in theform of formation voltage versus time of anodiza-tion in Fig. 1. The results presented in these line-ar plots are for films formed up to 125 V. How-ever a slight deviation from the linearity behav-iour was observed beyond 125 V (not shown inFig. 1). This indicates that the total current pass-ing through the film does not remain totally ionicbut part of it becomes electronic beyond 125 V.The rates of formation in the presence of UV ra-diation are found to be lower. This observation iscontrary to the observations on the rate of forma-tion of oxide films on other valve metals like Ta,A15-8etc.

Nb20s films grown to 125 V in the presenceand absence of UV radiation were dissolved in24% HF solution. The data are plotted as recip-rocal film capacitance versus time of dissolutionof the film in Fig. 2. It is observed from Fig. 2that the initial rate of dissolution is about threetimes faster for the film grown in the absence ofUV radiation but it becomes very slow afterward.It seems that almost whole of the film has dis-solved in few seconds, while film grown in UV ra-diation always has a slower rate of dissolutionand it decreases further with time. This might in-dicate that the film grown in the presence of UVradiation is less porous (or more compact) andhas a fewer flaw as compared to the film grown inthe absence of UV radiation.

Tltlllll lOt.15K

• InprHltllC1t of UVo In abSltnclt Of UV

NEU

TI&.~-•uc•:t: ItuI•••30

~u•a:

2

o 50 100

TlIII. of dinolution I Sltconds

Fig. 2 - Variation of reciprocal capacitance with time of dis-solution in 24°;', HF

The slower formation and dissolution rate ofNb20s films in the presence of UV radiation canbe explained by a model similar to that proposedby Vermilyea 7. During the formation of an anodicoxide film under high field, a positive spacecharge is created as a consequence of the excita-tion of valence electrons of the oxide into the

MAKEN et al.: EFFECT OF UV RADIATION ON ANODIC NbP~ FILMS 113

conduction band thus leaving positive holes in theoxide. As there would be a small barrier for thepassage of electrons from the oxide, it would bedifficult for electrons to jump from the solution tothe full band of the oxide in order to fill the posi-tive holes. As a result, a positive space chargenear the oxide/solution interface would developand the field at the interface would increase untilthe level of the holes in the full band decreases tothe level of the electrons or ions in the solution{Fig. 3).

The fermi energy of Nb205 film is not knownand hence an accurate estimate of the field re-quired cannot be obtained. Initially, the thicknessof film is very small, therefore, most of the radia-tion would be absorbed into the metal rather thaninto the oxide and little space charge would de-velop. Thus, the rate of formation of the filmwould be almost the same as that in the absenceof UV radiation.

After the film had grown to a thickness of afew hundred A, most of the light would be ab-sorbed into the film, a positive space chargewould develop and the field near the oxide solu-tion interface would increase until the level ofpositive holes was equal to the level of electronsin the solution. To maintain the photocurrent,some electrons from the solution may jump di-rectly to the oxide to neutralize the positive spacecharge at the interface. Thus the total current inthe film would not remain ionic and part of itwould become electronic. This would lead to theslower formation rate of the film. This has beenobserved experimentally for our system (Fig. 1).

As the electronic current in the film is in-creased the formation of the oxide would besmoother and its lattice would have fewer defects.Conversely, the film grown in the presence of UVradiation would be less porous than that grown inthe absence of UV radiation. Thus the rate of dis-solution of the less porous films would be sloweras compared to the more porous film. This is alsoobserved in our system (Fig. 2).

Fig. 3 - Schematic drawing of the electronic levels in the oxidefilm and on anions in the solution during the passage of

photocurrent.

Breakdown phenomenaA number of voltages were proposed II to de-

fine breakdown voltage. Breakdown voltage hasbeen reported II to be independent of currentdensity, temperature, topography and method ofpre-treatment of electrode surface. Further break-down voltage has been tound":" to be sensitiveto the nature of anodized metal or compositionand concentration of the electrolyte. In our earlierwork we found that breakdown voltage dependsupon concentration, current density, temperatureand nature of electrolyte. In the case of niobiummetal, breakdown is accompanied by the evolu-tion of the oxygen gas at the surface. Under gal-vanostatic conditions, rate of gas evolution in-creases with time and voltage becomes static forsometime (say 100 s). This particular voltage istaken as breakdown voltage (UB) for our investi-gations. The effect of UV radiation on the break-down voltage was determined at various currentdensities (j in A cm,2) and range of resistivities (pin Q em) in an oxalic acid solution. Figure 4 rep-resents the variation of the breakdown voltage(UB) with log j, while the plot of UB versus logpare presented in Fig. 5 in the presence and abs-ence of UV radiation. At the same current densityVB is greater in the presence of UV radiation.However, the difference increases with the in-crease in the current density (Fig. 4). It can be ob-served from Fig. 5 that VB values in the presenceof UV radiation are always higher than those in

Temp:298.15K.119 ;.3 OxaliC acid

• In p'HenCt of UVo In absence of UV

400 !!-0>-...co0-••::o>c~o'0oX

300:~

•

o

~~------------~----------~~~-3 -2 -1

109 CI / Amp cm 2)

Fig. 4 - Plot of breakdown voltage versus log j.

114 INDIAN 1 CHEM. SEe. A, FEBRUARY 1995

the absence of UV radiation for all resistivities ofthe solution.

According to lkonopisov theory II of breakdownvoltage during anodization, the electrolytic con-tact not only provides ions for oxidation, but alsoinjects electrons into the conduction band of theoxide. The high field strength can accelerate theseelectrons to an energy which is sufficient to freeother (secondary) electrons by impact ionization,causing an avalanche multiplication to occur.

The avalanches effect cause the density 0.) ofthe electronics current to increase with distance(x) within the anodic films according to

jJx) =L(O)Pexp[(eE/ Em)/( 1/x + 1/xr)] ... (1)

where j.(O) is the primary current density injectedat the electrolyte/oxide interface (at x = 0); e is theelectronic charge; x, is the recombination length;P is the probability for an electron to reach theunstable energy which is required for surmountingphonon scattering and Em is the difference be-tween the mean energy. of an electron necessaryfor ionization and the mean energy of an electronemerging from an ionization event.

The general equation for the breakdown vol-tage is given by

... (2)

where r = a/ (a, + 1) < 1and x = a.D, where D is the thickness of the filmand jH is the electronic current which is sufficientfor the oxide destruction. The dependence of thebreakdown voltage on the resistivity of the elec-trolyte can be expressed by the relation

... (3)

where p is the resistivity of electrolyte. Parame-ters a, and be are related to UO) at constant tem-perature and field; for a particular metal and elec-trolyte

When the film is grown in the presence of UVradiation, a positive space charge develops at theinterface and most of the electrons injected in theoxide are neutralized at the interface and thus themean value of the recombination constant, r,would decrease. At the same time, UV radiationalso activates electrons in the oxide, and Em in-crease in the presence of UV radiation. There-fore, from Eq. (3), UB in the presence of UV radi-

ation will be always higher at all resisnvmes ascompared to UB in the absence of UV radiation.

The primary electronic current density je(O) atthe interface also depends on the concentration ofanions in the solution. As the concentration ofelectrolyte decreases (or logp increases), the num-ber of electrons injected from the solution intothe oxide will also decrease. This means that therecombination constant (r) will increase with anincrease of logp, and the difference between UB

values in the presence and absence of UV radia-tion will continue to decrease as has been ob-served in our system (Fig. 5).

Variation of UH with current density (j) or fieldstrength E is given by

... (4)

where a. and Po are constants and relatedthrough the relation

when temperature and resistivity of electrolyte areconstant. According to Eq (4) UB should decreasewith increasing field (E). This is in contradictionto the experimental observation for Ta, AI andNb oxide films. UB values are found to be inde-pendent of E for Ta and AI oxide films!"!" whilethese values increase with an increase of currentdensity for Nb20S films", Ikonopisov" has pro-posed that E has little effect on UB• This is be-cause large variations in the current density onlycause small changes, in the field. Thus UB values

T""p 298 .15K Oxalic acid

1t00

~"0>..•..CJ's"0,.c~ 3000..,oK

'"•••..In

• In pretlnce of UVo In ab5lnce( of UV

200~--~ ~ ~2 3 It

109 (,/ACII)

Fig. 5 - Plot of breakdown voltage versus logp.

MAKEN et al.: EFFECT OF UV RADIATION ON ANODJC Nb20, FILMS

400III~a>t:;-O'••~ec!300'0

t••.CD

Oxalic acidTlmp = 29t.15K

• In presence of UVo In absenci of UV

200~ ~ ~~ __ ~~ __ ~2.0 2.1 2.2 2.3

~2 II!E I( V/cm)

Fig. 6-Plot of breakdown voltage versus E1/2.

are found to be independent of the field. ForNb205 films, however, a hundred-fold increase inthe current density causes an increase in VB of120 V. The hundred-fold change of current dens-ity changes the field strength (E1!2) only by 18%(ref. 9). Therefore, either VB is very sensitive toapplied field or some other constant (Em or r) alsochanges simultaneously with E. By definition, Emis the difference between the mean energy of anelectron necessary for ionization and the meanenergy of electron emerging from an ionizationevent. As the value of E increases, the mean ener-gy associated with the electron emerging tromionization event will also increase. Thus, Em willalso increase. The VB should increase with an in-creaseofE.

When the film is irradiated with UV radiation,the electrons in the oxide would be excited andthe difference between the mean energy of anelectron jumping from solution to oxide and themean energy of the electron emerging from theionization even would increase, or Em would in-crease. Thus VB values at particular values of Ewould be increased in the presence of UV radia-tion as compared to those in the absence of radia-tion. When the field is increased in the presenceof UV radiation the value of Em also increasesfurther, and differences between VB values in thepresence and the absence of UV radiation alsoincrease with an increase of the field. In fact, thishas been observed experimentally (Fig. 3 andFig. 6). Since the electronic current during forma-

115

tion of films in the presence of UV radiation in-creases, the thickness of the film and the value ofE cannot be estimated by the Faraday's law meth-od. The magnitudes of field (E) in the presence ofUV radiation at a particular current density aretaken to the same as in the absence of UV radia-tion. The calculation of the field in the absence ofUV radiation was made by the usual Faraday'slaw method, when the electronic current is negli-gible. This has been described earlier".

It can be concluded that, in the presence of UVradiation, the reduced rate of formation of anodicNb205 films is due to the development of a posi-tive space charge near the solution! oxide inter-face. Increase in the breakdown voltage value inthe presence of UV radiation may be due to adecrease in the recombination constant (r) and/orincrease in the impact ionization constant (Em).

AcknowledgementS.M. thanks the CSIR, New Delhi for the award

of a research associateship.

References1 Hurlen T, Bentzen H & Hornkjol S, Electrochim Acta, 32

(1987) 1613.2 Torresi R M & Nart F C, Electrochim Acta, 33 (1988)

1015.3 Quarto F Di, Piazza S, D'Agostino R & Sunseri C, Elec-

trochimActa, 34 (1989) 321.4 Badawy W A, Felrbe A & Plieth W J, Electrochim Acta,

34 (1989) 1711.5 YoungL, Trans FaradaySoc, 50 (1954) 153.6 VermilyeaD A, J appl Phys, 26 (1955) 489.7 Vermilyea D A, J electrochem Soc, 104 (1957) 212.8 Nigam R K & Kapoor R, Indian J Chem, 16A (1978)

283.9 Nigam R K, Singh K C & Maken S, Can J Chem, 65

(1987) 512.10 Holtzberg F, Reisman A, Berry M & Barkenblit M, JAm

Chem Soc, 79 (1957) 2039.11 Ikonopisov S, Electrochim Acta, 22 (1977) 1077.]2 Ikonopisov S &. Girginov A C, R Acad Bulg Sci, 28

(1979)257.13 Burger F J & Young L, Electrolytic capacitor in progress

in dielectrics, Vol 5 edited by L B Birks (Heywood, Lon-don) 1962.

14 Vermilyea D A, Anodic films in advances in electro-chem-istry and electrochemical engineering, Vol 3 edited by PDelahey & C W Tobias, (lnterscience, New York) 1963,p.211.

15 Korinek G J, Electrochem Techn, 2 (1964) 360; 3 (1965)206.

16 Renne V T, Elektroliticheskie Kondensatory (Energia,Leningrad)·1969.

17 Malinenko V P, Odynets L L, Chekmasova S S & HaninaL Va, Elektrokhimiya, 7 (1971) 1846.

18 Odynets L L & Chekmasova S S, Elektrokhimiya; 9(1973)1177. '

19 Young L, Proc roy Soc London, 258A (1960) 496; 263A(1961) 395.