The photoactivity of porous Tio2 anodized at a high voltage

The photoactivity of porous Tio2 anodized at a high voltage

Volume 100, number 3 CHEMICAL PHYSICS LETTERS THE PHOTOACTIVITY OF POROUS Ti02 ANODIZED Doug MILLER, Suzanne MAMICHE-AFARA, 9 September 1983 ...

2MB Sizes 2 Downloads 10 Views

Volume 100, number 3

CHEMICAL

PHYSICS LETTERS

THE PHOTOACTIVITY

OF POROUS Ti02 ANODIZED

Doug MILLER, Suzanne

MAMICHE-AFARA,

9 September

1983

AT A HIGH VOLTAGE

M J. DIGNAM and hi. MOSKOVITS

Department of Chemishy, Lash Miller Chemical Laboratories. University of Toronto, Toronto, Canada J1.S IA1 Rcccited

29 March 19S3: in final form 27 June 19S3

The photoactivity of Ti02 produced b> hi&\oltage anodizrltion as described by Macbenoir et al. was comparred to single-crystal TiOz. The rtnodic o.xide filmwas found to be highly porous. The photocurrent and quantum yield of the anodic sample is markedly larger on the lonr_wavelcn~th side of the action spectrum as compared to singlecr)stdl rutile. We ascribe this behavior either to the formation of ;lndtasc “pockets” in the film or to the incomplete o.xidation of Ti during the anodization process.

l_ introduction The photoelectrocheniicril and photocatrilytic properties of TiO, have been studied extensively [ 1-3-l] *. The photoactive TiOz is nomraliy placed in a photoelectrochemic~ cell with either an aqueous or nonaqueous electrolyte. where it fomrs the anode. Photocatalytic reactions have two general classifications. as suggested by hlatsumoto [ 1] : reactions in which light energy is stored (AC > 0), and reactions where the rate is photoassisted by the catalyst (2.G < 0). The reactions where AC > 0 are important lie in the ares of solar energy conversion. Examples of these are: the photolysis of water [3--91. photorcactions of NO [lo] and CO [I 11, photosynthesis of amino acids [ 121, photocleavage of hydrogen sulfide [ 131. and the photoo.tidation of carbon with water [ 14,151. The other chrss of reactions includes processes such as the photodecomposition of carboxylic acids [ 161. the photoo_xidation of cyanide [ 17,l S] ) sulfite [ 1S] , hydroquinones [ 191. and hydrocarbons [20]. the photo Koibe reaction [21--241. the photoreactions of dichromate [25], photoreactions of water with alcohols [2,3], and in the photooxidation of polyolefins and polyamides. In the second class of reactions, the TiOz acts as an energy absorber. The * A comprehensive series of papers on the study of TiOp can be found in ref. [ 35 ] _

236

stored ener,T is then imparted to the system, thereby enabling sunlight to be used as an activation energy substitute. The wide forbidden band of TiOZ (~3 eV) implies that ultraviolet photons are required for photoreactions on TiO,. Unfortunately, though, sunlight is low in UV photon content: hence, many of these reactions will be quite inefficient in photoprocesses depending upon sunlight as the energy source. In this paper, we report an improvement in the quantum efficiency of a TiO-, PEC, at wavelengths longer than 390 nm. The action spectra and current -voltage plots for the anodized TiO2 and rutile singlecrysnal TiOZ were obtained.

2_ Experimental Photosensitive TiOl tihns were prepared using the high-voltage anodization technique reported by Marchenoir et al. [29,30] _ The voltage employed was well above the sparkling potential of Ti (-100 V), BSsisting the production of a highly porous TiO?/Ti sample. The Ti foil was purchased from the A-D MacKay Co., thickness O.OOl”, purity 999%. Ti foil was cut into 10 cm X 1 cm strips, and was successively polished with time carborundum paper, polished with an alumina-ethanol slurry, and ultrasonically degreased in CH,CI, for 30 min.

0 009-2614/83/0000-0000/S

03.00 0 1983 North-Holland

Volume 100, number 3

CHEMICAL PHYSICS LETTERS

9 September 1983

The [email protected] system consisted of a variablevoltage (O-150 V, 8 A) power supply connected to the electrochemical cell consisting of a cylindrical Pb foil cathode surrounding the Ti foil anode draped over a copper rod. The electrolyte was 5.0 N HzS04 to which OS g of lithium nitrate was added to increase

the conductance of the electrolyte_ It was reported [31] that the addition of LiNO, does not affect the composition of the anodic oxide. The voltage applied was raised slowly, so that the current was never greater than 8 A, and the maximum voltage attained (150 V), was applied for 5 min. The current density through the film was 020 A/cm’. The anodized sample was rinsed with triply distilled water. Prior to the study of the photoelectric properties, the film was soaked in 0.1 N NaOH for several hours. The TiO2/Ti sample was cut into 2 cm X 0.5 cm strips which were supported

on a glass slide by epoxy

cement. The sample was masked using silicone rubber cement, leaving an illumination area of =O.lO cm2. The singlecrystal TiO, was polished with 1 pm dia-

mond paste. The polished crystal was annealed

in oxy-

gen at a temperature of 900°C. pure hydrogen was used to dope the crystal, the temperature being 500°C. Finally the TiO2 was etched in sulphuric acid/ammonium sulphate (one to one by volume)_ Ohmic contact to the back of the crystal was established with a gallium-indium alloy. The photosensitive characteristics of the TiO, fti

Fig. 1. Photocurrent versus bii for the anodized TiOs samples;Ti(l-a)l81,Ti(l-e)181 and Ti(2)182 refer to three different samples; and for singlecrystal TiOa.

suiting from the conversion of photon fhrx (photons/ cm?) to power flux (mW/cmQ The intensity of the light source is determined in mW/cm?. Fig. 1 represents current-voltage curves for samples Ti(l-a)lSl, Ti(l-c)lSl, Ti(2)182 and singlecrystal TiO,. Fig. 2 is the action spectra for the four TiO2 samples. The sample Ti(2)182 prepared in August 1982 exhibited a small dark current, which increased slightly in the region of oxygen evolution. The samples Ti(l-a)181 and Ti(l-c)lSl initially exhibited a significant dark current in the oxygen evolution region. However, the dark current settled to a low level after

were studied using a PAR-170 potentiostat. The reference electrode was a saturated calomel electrode (SCE). The photoresponse curves were obtained using a 150 1%’xenon lamp white light source. The bias applied to the system ranged from -1 _OV SCE to +1-O V SCE. The intensity at various wavelength settings of monochromator was measured, allowing quantum efiiciency to be calculated_

3. Results The quantum efficiency, the following equation:

9, was determined

using

Q= 1_2408jffi, where j, the current density is defined as the current/ illuminated area bA/cm2), and 12408/h is a term re-

rig. 2. Action spectra (quantum yield versus wavelength) for three different samples of anodized TiOa, compared to singlecrystal TiOz. The bias was held at +I_0 V SCE for all samples. 237

V o l u m e 100, n u m b e r 3

CHEMICAL PHYSICS L E T T E R S

9 S e p t e m b e r 1983

l--ig. 3. (a) Scanning electron micrograph o f anodized TiO2 formed at 150 V (sample (1-a)181). T h e scale is 30 gill b e t w e e n tics. (b) lligll-magnifk-alion scanning micrograph o f the same sample us ill (a): scale; 1 ~ m b e t w e e n tics.

tliirty

minutes

of immersion

Evidence for the porosity

In the electrolyte.

of the fJnx is provided

in the scanning electron micrographs (fig. 3). The hi& voltage teclmique creates a bigbly porous o.Gdc film as shown in fig. 3x The mean pore diameter is of the or-

der of 1000 A (fig. 3). An X-ray cbemicsl analysis

wws also performed.

Tbc

only element detected other than Ti and 0 was S. prcsumably absorbed from the sulfuric acid anodizing solution_ No lithium wx detected in the analysis.

4. Discussion The large initial dark current observed in the study of tbeTi(l-a)lSl and Ti(l-c)lSl samples may be due to substantial amounts of oxygen absorbed in the porous TiOz _ III the UV region. singlecrystal TiOZ 113s ;1 much grcatcr qumrunl cfiiciency t11m does anodircd TiO? (fig. 2). A good quantum efficiency in tlic UV region is not surprising as. tbc band-gap of TiO? corresponds to a cutoff \mvelcngtll of400 nm. Tllc action spectra for tbc anodized TiO, xc reproducible in tile region 360100 nm. Even tbougb there are apparent tlifferences in the pllotosensitivity between samples it is signikmt tlmt in tllc near-visible region, tbc quantlull cfficicncies become conlparable. In the blue part of‘ tile spcctrunl. tlw polous TiOz possesses a qumtunl cfficitncy superior to that of the rutile single tryst-xl. This may be due to 3 number of leasons of wbicb WC present two 3s most lihely. Tlw first involves tile fornmtion of matase, 3x1 allotrope of TiO-, along the sides of rhe pore during bigb-voltage anodiz& hlarcbennir et al. [29jO] suggest that the higil voltage applied to tlw Ti foil during anodiGlg, 3 voltage wbicb exceeds the sparking potential of the oxide, creates 3 hot likunent or “‘spark” wbicli is sufficiently hot to convert tlw rutile to anatase. The anatase-rutile transforni3tion occurs at temperatures between 700 and 1200°C. depending upon the method employed. 1lence. during slow annealing conin~only used in single+xystsl preparation there is sufficient time for Al the 3nat;Lse to convert to rutile which has 9 lower pllotoactivity than anatase. This may 1101 be the case during tlie “‘sparking” process and as a result pockets of the matase Ti02 will rem-tin, presumably lining the pores. These pockets may contribute to the estension

of the photoactivity of the PEC into the blue region of the visible spectrum. Alternatively the improvement of the quantum efficiency in the 300 xun region could be due to incomplete osidation of the titaniuni during medic o.sidation. Gautron et al. [34] llavesbown that non-stoicbiometric TiO, (i.e. TiOl_,) shows improved qumtum efficiency and improved near-UV--\-isible rcspnnse for ~1range ofs rangins from 2 X IO-” to 47 X 10v4 _with optimum efficiency at around 7 X 1 O-‘_ Altl~ougl~ this mxkmism is 3 possible contributor we prefer the first since the quantum yield measured in the 400 nin region of the spectrum ~3s quite rcproducible for 3 number ofsamples prepared under sonx~1bat vaving anodizing conditions in wliicb circunistances one would have espccted the degree of nonif present at 311. to vaq.

sloicbionietly.

5. Conclusion Tbc results

01-a preliminxy

study

indicate

tlmt

[email protected];lgc anodizarion of Ti produces 3 ‘Mew” type of~l~l~ot~elfctrodr wliicli li3s 31i improved qumtuiii dficirncy in tlic ncttr-visible region of the spectrum olcr sin$x-Tstltl rurlle. Tbc appxcnt increase in pbo~oefficicncy in rbe blue region may be due to either the incorporiition of amtasc in the prcdoniinantly rutilc medic oxide libn or IO oxidation during rbe lli$-volt3s~‘ rinodizalion.

Acknowltulgenient We rbanh liesrarcb

NSERC

Fund

and the donors

for Iinmicia]

of tlw Petralcunl

assistance.

References

I-S] S. Sllto and J.Xl_ \Vhlte, Clwn. f’hys. Ldters ]6] kto

72 (19SO)

Jnd J_hl. I\‘hitc, J. I’hgb. Chctm. 65 (1981)

592.

239

Volume 100, number 3

CHEMICAL PHYSICS LETTERS

[ 7 ] F.T. Wagner and G-A. Somorjai, J. Am. Chem. Sot. 102 (1980) 5494. [S] I;. Kogo, H. Yoneyama and H. Tanura, 3. Phys. Chem. 84 (1980) 1705. [9] M.V. Rao, R. Rajeshwar, V.R.P. Vemeker and J. Dubow, J. Phys. Chem. 84 (1980) 1987. [ 101 H. Yoneyama, H. Shiota and H. Tamura, Bull. Chem. Sot. Japan 54 (1981) 1308. [ 111 T_ Inoue. A. Fujishima. S. Konishi and K. Honda, Nature 277 (1979) 637. [ 12 j Ii. Rciche and AJ. Bard, J. Am. Chem. Sot. 101 (1979) 3127. [ 131 81. Grfitzei, Chem. Erg. News 59 (July 17.1981) 40. [ 141 T. Kawai and T. Sakata. Nature 282 (1979) 283. [ 151 S. Sato and J.M. White, J. Phys. Chem. 85 (1981) 3136. [ 161 B. Kraeutler and AJ. Bard, J. Am. Chem. Sot. 100 (1978) 5985. 1171 S.N. Frank and AJ. Bard, J. Am. Chcm. Sot. 99 (1977) 303. [lS] S.N. Frank and A.J. Bard, J. Phys. Chem. 81 (1977) 1484. [19] F.F. Fan and AJ. Bard. J. Am_ Chem. Sot. 102 (1980) 2592. [ 70 1 WV_\V. Dann, K.O. Willbourn, F.F. Fan and AJ. Bard, J. Phys. Chem. 84 (1980) 3207. ]2l] B. Kracutlcr and AJ. Bard. J. Am. Chem. Sot. 100 (1978) 2239. [22] B. Kraeutirr, CD. Jaqer and A-l. Bard, J. Am. Chcm. Sot. 100 (1978) 4903.

9 September

1983

[23 ] B. Kraeutler and AJ. Bard, Nouv. 1. Chim. 3 (1979) 31. 1241 I. Izumi, F.F. Fan and AJ. Bard, J. Phys. Chem. 85 (1981) 218. [2S] H. Yoneyama, Y. Yamashita and H. Tamura, Nature 282 (1979) 817. [26] L. Tang, D. Sallet and J. Lemaire, Macromolecules 15 (1982) 1437. [ 27 ] R. Arnaud and J _ Lemaire, in: Developments in polymer degadatron, Vol. 2, ed. V_ Grassie (Applied Science Publishers, London, 1981) p_ 159. [28] R. Amaud and J. Lemaire, in: Developments in polymer photochemistry, Vol. 2, ed. N. Allen (Applied Science Publishers_ London. 1981) p_ 135. [29] J-C. hiarchenoir, J. Gautron and J--P. Loup, hletaux Corrosion Ind. 619 (1977) 83. 1301 J-C. hlarchenoir. J.-P. Loup and J. hlasson, Thin Solid Films 66 (1980) 357. 1311 WTL. B&n, Formation of Porous Films onTitanium Alloys by Anodization Treatments, Air Force Material Lab, Wright-Patterson, OH (1979), hlicrofim. [32] PJ. Boddy, J. Electrochcm. Sot. 115 (1968) 199. [ 33 ] R J _H. Clark, The chemistry of titanium and vanadium (Elsevier, Amsterdam, 1968) p_ 271. 134 ] J. Gautron, P. Lemasson and J.-F_ M~ucco. Discussions Faraday Sot. 70 (1980) 81. [35] J. Electrochem. Sot. 115, No. 2 (1968); hletaus Corrosion lnd. 619 (1977); Discussions Faraday Sot. 70 (1980).