May 1, 1996 / Vol. 21, No. 9 / OPTICS LETTERS 641

Ion-beam manipulation of the photorefractive properties ofstrontium barium niobate planar waveguides

Elaine E. Robertson, Robert W. Eason, and Malgorzata Kaczmarek

Optoelectronics Research Centre and Department of Physics, University of Southampton, Southampton SO17 1BJ, UK

Peter J. Chandler and Xuefeng Huang

School of Mathematical and Physical Sciences, University of Sussex, Brighton BN1 9QH, UK

Received October 10, 1995

Photorefractive planar waveguides have been fabricated in cerium-doped strontium barium niobate(SrxBa12xNb2O6, SBN) single crystals by ion-beam implantation. The losses measured were as low as 0.1and 7.0 dB cm–1 for the TM and the TE modes, respectively. Subsequent two-beam-coupling experimentsperformed on the waveguides showed that, unlike for BaTiO3 and KNbO3 waveguides formed by ion-beamimplantation, the two-beam-coupling gain direction did not reverse. Also, the response time had been reducedby 2 orders of magnitude. 1996 Optical Society of America

An advantage of fabricating waveguides in optical ma-terials is that, because of the beam-confining proper-ties inherent in a waveguide, high light intensity ispreserved throughout the guide. In a photorefractivematerial this intensity enhancement is particularly ad-vantageous, as the photorefractive response time is de-pendent on the intensity. Hence, for a given inputpower, a substantial decrease in the response timewill be observed. A further advantage is that using awaveguide geometry provides compatibility with otherintegrated-optical devices. Photorefractive materialssuch as BaTiO3, KNbO3, and SrxBa12xNbO6 (SBN) allhave large electro-optic coefficients (r42 ­ 1640, r42 ­380, and r33 ­ 1340 pm V21 for BaTiO3, KNbO3, andSBN:75, respectively), resulting in strong photorefrac-tive responses. SBN is an interesting material, as itselectro-optic coefficient compares favorably with that ofBaTiO3, but it does not suffer the 6–9 ±C phase transi-tion as BaTiO3 does.

A variety of methods were used previously tofabricate both planar and channel waveguides inSBN: sulfur diffusion,2 ion-beam implantation,3 straintechniques,4 and electro-optical methods,5 the first twoproducing planar guides and the latter two, channel.In the case of the sulfur-diffused guides, losses werehigh and, although the guides were still found tobe electro-optic, no photorefractivity was observed.In the case of the ion-implanted guides, again nophotorefractivity was observed, although because theguide was implanted onto the z face of the materialthe photorefractive effect would be inherently moredifficult to detect. However, for ion-implanted BaTiO3and KNbO3 waveguides a strong photorefractive effectwas observed.6,7 To determine whether a guideis photorefractive, one usually performs two-beamcoupling with two mutually coherent extraordinarypolarized beams interfering within the waveguide.The interesting feature of ion-implanted BaTiO3and KNbO3 waveguides was that the gain directionin the waveguide was opposite that in the bulk,suggesting that the predominant charge carrierhad changed (in both cases from hole to electron).

0146-9592/96/090641-03$10.00/0

In both cases the authors suggested that a formof chemical reduction was occurring within thewaveguide layer, induced by the ion-implantationprocess. As the predominant charge carrier in SBNis the electron, before any ion-implantation experi-ments it was thought that no gain reversal in SBNguides would be seen, because the implant has areducing effect. Furthermore, it was thought possibleto reduce the response time dramatically within theguides, as the theoretical results of Klein et al.8 predictthat this occurs as the crystal is chemically reduced.We discuss our result for SBN waveguides in compari-son with this theory.

To fabricate the waveguides we mounted the crys-tals in a vacuum chamber and bombarded them with2.0-MeV H1 ions, which were directed nominally per-pendicular to the surface (i.e., parallel to the x axis)of the crystal. The ions penetrate the surface layerof the crystal, initially interacting electronically withthe crystal lattice, which produces little or no dam-age to the crystal. However, once the ions slow downthey undergo nuclear collisions, forming a damagedlayer that has a refractive index lower than that of thebulk. Both SBN:75 and SBN:61 crystals have beenstudied here, which have Curie temperatures of 57 and75 ±C, respectively. Unlike in the procedure adoptedin Ref. 3, the current used was kept deliberately lowsø0.05 mAd so as not to heat the surface of the crys-tal unduly, as this could cause adverse effects in thewaveguide layer. If the waveguide layer is heated be-yond the Curie point it depoles, and no photorefrac-tive effect is subsequently observed. This may explainwhy no report of photorefractivity was mentioned inRef. 3, as the current used was six times higher thanthe one reported here. All our implantations werecarried out at room temperature with initial doses of2 3 1015 ions cm–2 in the SBN:61 and 1 3 1016 ionscm–2 in the SBN:75. In the SBN:61, two further im-plants of 2 3 1015 ions cm–2 were subsequently per-formed into the existing guide, and a further implant of1 3 1016 ions cm–2 into the SBN:75. To determine theresulting guide depth, we examined the crystals under

1996 Optical Society of America

642 OPTICS LETTERS / Vol. 21, No. 9 / May 1, 1996

Table 1. Losses in Ion-Implanted SBN:61 (Assuminga Coupling Efficiency of 80%)

Dose TM loss TE loss(ions cm22 3 1015) (dB cm21) (dB cm21)

2 1.6 9.14 6.0 246 13.0 39

Table 2. Losses in Ion-Implanted SBN:75(Assuming a Coupling Efficiency of 80%)

Dose TM loss TE loss(ions cm22 3 1016) (dB cm21) (dB cm21)

1 0.1 7.02 0.5 18.0

the microscope. One can clearly see the waveguide bylooking at the z face, as the implanted region appearsdarker than the bulk crystal. Using 2.0-MeV H1 ions,we measured the depth to be s30 6 1d mm. After eachimplant losses were measured, and the photorefractiveresponse time was measured by two-beam coupling.

To measure the losses of the waveguides, we pre-cision end polished the crystals to obtain maximumcoupling efficiency. Light was then launched into thewaveguide through a 43 objective and collected at theexit of the waveguide by a 103 objective, which imagedthe light from the waveguide onto a calibrated power-meter. All losses were measured at 647 nm becauseof the high transmission of SBN at this wavelength.Each waveguide was 5 mm in length. Table 1 showsthe losses observed for the SBN:61 waveguide andTable 2, those of the SBN:75 guide for different doses.It is interesting to observe how the losses of both theTM mode and the TE mode for the SBN:61 increasedramatically as the implantation dose increases. Thelosses incurred in the SBN:75 showed the same trendbut to a lesser extent. This may be attributable to in-creasing conductivity in the waveguide layer, but, asyet, no clear explanation is available. The losses ob-tained for the 1 3 1016-ions cm22 dose are believedto be the lowest measured in any SBN planar wave-guide. For other ion-implanted waveguides such asLiNbO3, waveguide losses have been reduced by an-nealing at 200 ±C.9 This was not attempted in theSBN guides because of the low Curie temperatures asmentioned above. It was also found that the damagedlayer of the waveguide has a dramatically increasedconductivity, making crystal repoling difficult (for eachcrystal it was found that a maximum electric f ield ofø100 V cm21 could be applied, whereas the usual pol-ing f ield is ,1.5 kV cm21).

To determine whether the guides were photorefrac-tive, we carried out a series of two-beam-coupling ex-periments. For the SBN:61 all the fabricated guideshad retained their photorefractive properties. How-ever, for SBN:75, only the waveguide with the 1 3 1016-ions cm–2 dose still showed photorefractive behavior.The beam-coupling experiments showed that the gaindirection within the waveguides was the same as thatof the bulk, whereas in both BaTiO3 and KNbO3 wave-

guides the gain direction had reversed. This absenceof gain reversal within the SBN waveguides impliesthat, as predicted, electrons remain the predominantcharge carrier after the implantation, whereas gain re-versal would indicate a change in predominant chargecarrier from hole to electron or vice versa. The pho-torefractive response time was also measured for thebulk and the three SBN:61 waveguides implanted atdifferent doses as a function of the incident irradi-ance (Fig. 1). It can be seen clearly how the responsetime decreases as the implantation dose increases. Al-though a faster time would be expected in a waveguidethan in the bulk because of beam confinement, if theresponse times are compared at a common value of ir-radiance (unlike the results presented in Ref. 6) the re-duction in the response time with the ion-implantationdose is evident. If the response times are compared atthe same intensity for the different implantation doses(Fig. 2), again it is apparent that the implantation ishaving a profound effect on the intrinsic photorefrac-tive properties. In summary, the two important ex-perimental results that we have observed for SBN were

Fig. 1. Response time versus irradiance for bulk andwaveguide SBN:61.

Fig. 2. Response time versus ion-implantation dose inSBN:61.

May 1, 1996 / Vol. 21, No. 9 / OPTICS LETTERS 643

Fig. 3. Response time and beam-coupling gain in BaTiO3(after Ref. 8).

the absence of gain reversal in two-beam coupling andthe decrease in the response time.

Klein et al.8 have shown how the impurity-ion oxi-dation state ratio affects the photorefractive gain andresponse time. Figure 3 (plotted with the parametersfor BaTiO3 but still relevant to SBN) shows that thegain and response times vary as the donor/acceptorratio, or reduction ratio sXyX1d, increases. Lookingat the plot for the gain, G, against XyX1, we see that asthe crystal is reduced sX1 1 e2 ! Xd the predominantcharge carrier changes from hole to electron. In bothBaTiO3 and KNbO3 the gain direction reverses, imply-ing a change in predominant charge carrier from holeto electron. However, inasmuch as SBN is initiallyan electron conductor, the reduction process would notchange the dominant carrier; i.e., it remains an elec-tron conductor. One possible reason for this is thatduring the implantation process the H1 ions (and theHe1 ions for KNbO3 ) disrupt the lattice, creating oxy-gen vacancies. To compensate for this charge imbal-ance the reduction ratio changes, chemically reducingthe crystal. It is this reducing effect that causes thepredominant carrier to change; hence the gain direc-tion reverses in BaTiO3 and KNbO3 . To explain thedecrease in the response time with increasing implan-tation doses we have to revert to the plot in Fig. 3,which shows how the response time varies as the re-duction ratio increases. Once the crystal has been re-duced such that it is in the electron-conducting regime(as it is initially with SBN), as the reduction ratio in-creases, the response time decreases. As the implan-tation dose increases, the amount of reduction requiredfor restoration of the charge balance will increase;hence the reduction ratio will increase. It is this be-havior that we suggest causes the reduction in responsetime measured. In the case of the SBN:75 waveguidewith an implantation dose of 2 3 1016 ionsycm–2 nophotorefractive effect was observed. There are twopossible explanations for this. First, it is possible thatsuch a high dose damages the waveguide layer of thecrystal such that it is no longer photorefractive; sec-ond, it may be that the reduction ratio has been in-creased to such an extent that the photorefractive gain

is negligible (as predicted in Fig. 3 with the plot of G

against XyX1). These reduction effects have been ob-served in other photorefractive materials (BaTiO3 andLiNbO3) when crystals have been heated in reducingatmospheres.8,10

From Fig. 3 it can be seen how the response time hasbeen reduced by more than 2 orders of magnitude inthe waveguides. This is an interesting result becausephotorefractive waveguide devices with considerablyfaster response times than their bulk counterparts,combined with the reduction in response time that isdue to the waveguiding properties, can be designed.A further advantage of this reduction effect is that thephotorefractive properties can be manipulated, withaccuracy, by ion-beam implantation.

In conclusion, we have produced planar waveguidesin SBN:61 and SBN:75 by implanting bulk crystalswith H1 ions. The losses measured were as low as0.1 dB for the TM mode, which we believe to belowest loss reported to date in SBN waveguides. Theresulting guides were shown to have preserved theirphotorefractive properties with the gain direction inthe waveguide the same as that of the bulk. Wereduced the photorefractive response time by 2 ordersof magnitude by increasing the implantation dosebecause of the chemically reducing properties of theimplant. This is believed to be the first report of thephotorefractive effect in SBN waveguides and of theiroptimization through ion-beam implantation. Futureresearch will involve looking at the effect of sequentialion implantations on losses, as currently it is unknownwhether the source of the increasing losses is due to thedose or to the sequential nature of the implant.

We thank P. Townsend, University of Sussex,and P. G. R. Smith, University of Southampton, foruseful discussion and D. P. Shepherd, University ofSouthampton, for help in the polishing of the crystals.

References

1. A. Yariv, Optical Electronics, 4th ed. (Saunders,Philadelphia, Pa., 1991).

2. O. Eknoyan, C. H. Bulmer, H. F. Taylor, W. K. Burns,A. S. Greenblatt, L. A. Beach, and R. R. Neurgaonkar,Appl. Phys. Lett. 48, 13 (1986).

3. D. Kip, S. Aulkemeyer, and P. Moretti, Opt. Lett. 20,1256 (1995).

4. J. M. Marx, Z. Tang, O. Eknoyan, H. F. Taylor, and R.R. Neurgaonkar, Appl. Phys. Lett. 66, 274 (1995).

5. N. H. Zhu and Z. Q. Wang, Opt. Quantum Electron. 25,517 (1993).

6. K. E. Y. Youden, S. W. James, R. W. Eason, P. J.Chandler, L. Zhang, and P. D. Townsend, Opt. Lett.17, 1511 (1992).

7. M. Zha, D. Fluck, P. Gunter, M. Fleuster, and Ch.Buchal, Opt. Lett. 18, 577 (1993).

8. P. Gunter, Electro-Optic and Photorefractive Materials(Springer-Verlag, Berlin, 1986), p. 266.

9. G. L. Destefanis, J. P. Gailliard, E. L. Ligeon, S. Valette,B. W. Farmery, P. D. Townsend, and A. Perez, J. Appl.Phys. 50, 7898 (1979).

10. J. P. de Sandro, (Optoelectronics Research Centre,University of Southhampton, Highf ield, SouthhamptonS017 1BJ, UK (personal communication).