## Abstract

Thermo-mechanical and -optical properties of Y_{3}Al_{5}O_{12} (YAG), YVO_{4}, and GdVO_{4} were evaluated with less than 2% of evaluation error. Measured thermal expansion coefficient for YAG, [100]-YVO_{4}, [001]-YVO_{4}, [001]-GdVO_{4}, and [001]-GdVO_{4} were 6.13, 1.76, 8.24, 1.19, and 7.26 × 10^{−6}/K at room temperature. Temperature coefficients of refractive index for YAG, YVO_{4} in ordinary and extraordinary polarization, and GdVO_{4} in ordinary and extraordinary polarization at room temperature for the wavelength of 1.06 μm were 12.1, 15.5, 8.41, 15.2, and 9.92 × 10^{−6}/K, respectively.

© 2014 Optical Society of America

## 1. Introduction

Recent advance in giant micro-photonics has enabled drastic power-scaling in various photon sources based on microchip lasers [1]. Although many benefits have been realized by the miniaturization of highly bright solid-state lasers [2, 3], the excessive heat generated by densification of the excitation limits the averaged-power from miniature lasers. The temperature increase in laser gain media due to this heat causes not only thermal expansion but also influences on the optical gain and the heat capacity [4, 5]. As a result of high power pumping, lensing and birefringence induced by the severe temperature distribution in laser gain media degrades the laser performance of destructs laser gain media itself.

In order to extend the scalability in both of brightness and average power, heat management in laser cavity according to precise thermo-mechanical and -optical parameters of laser gain media is the most critical. However, there are significant discrepancies in the reported value of material parameters. Especially reported values of the linear thermal expansion coefficient, *α* and temperature coefficient of refractive index, *dn*∕*dT* are different more than one order. For examples, reported values of *dn*/*dT* at 1.06 μm for *π*-polarization in YVO_{4} varies from 0.3 to 13.5 × 10^{−6} K^{−1} [6, 7], and still now there seems to be no hope that this difference will converge in near future.

The variation of reported *α* is within a several-fold range to the utmost: for example, reported values of *α* in Y_{3}Al_{5}O_{12} (YAG) varies from 2.5 to 8.2 × 10 ^{−6} K^{−1} [8, 9]. Our research started from a consultation that Nd:YVO_{4} crystal in the laser component with high brightness output designed for industrial application by use of the previously reported *α* [10]. Therefore, our report should be useful not only for scientific filed but also many real applications in industrial field urgently.

Moreover, *α* includes another important problem. Even though many researchers believe that YAG crystal has anisotropic thermal expansion [11, 12], it should be wrong physically. Thermal expansion can be described as a coordinate transformation per unit temperature, thus *α* is a second order tensor. From Onsager reciprocal relations, YAG (cubic) and vanadate (tetragonal) have one and two independent components of thermal expansion coefficient, respectively. Authors consider that the reported anisotropy in thermal expansion of YAG was contributed by experimental error, which should be re-evaluated with high accuracy. One of objectives in this work is an experimental confirmation by evaluation of dependence on crystal axes in *α* of YAG.

We already provided the evaluation procedure for *α* and *dn*/*dT* of GdVO_{4} [13], however those were not proved by the evaluation of experimental errors. In this work, we tried to summarize these thermal properties of YAG, YVO_{4}, and GdVO_{4} with high accuracy based on the careful treatment of experimental errors.

## 2. Methods

#### 2.1 Thermal expansion coefficient

Thermal expansion was measured using a push-rod type dilatometer (DIL 402C, NETZSCH). Measurements were carried out under the dynamic helium atmosphere with gas flow rate was 50 ml/min, and heating rate is 4K/min within the range from 0 to 300 °C in temperature, *T*. The contact force of the push-rod was 0.25 N, and measured data was calibrated by an fused silica standard.

Measured samples were (111)-cut YAG single crystal (Scientific Materials Co.), (100)-cut and (001)-cut YVO_{4} single crystals (ITI Electro Optics Co.), and (100)-cut and (001)-cut GdVO_{4} single crystals (Shandong Newphotons Science and Technology Co., Ltd.) with the size of 8 mm in diameter and 25 mm in thickness.

#### 2.2 Refractive index

The absolute value of refractive index, *n* was calculated from the angle of minimum deviation of the triangular prisms that have bases of isosceles right triangles [14]. The height of prisms and the length of long side in bases were 10 mm. We evaluated undoped and 1.0at.% Nd^{3+}-doped YAG single crystal (Scientific Materials Co.), YVO_{4} single crystals (ITI Electro Optics Co.), and GdVO_{4} single crystals (Shandong Newphotons Science and Technology Co.,LTD). Bases of the YAG prism were parallel to (111)-plane, on the contrary, bases of vanadates were parallel to (001)-plane.

The minimum deviation angle, *A*_{min} of these prisms and prism angles,
*A*_{p} were measured within the wavelength range between 420 nm and
1600 nm, where light was analyzed by the spectrometer (V-30D, Shimadzu Rika Co.). Values of
*A*_{min} and *A*_{p} were detected by the CCD
camera (CCD-41R, Shimadzu Rika Co.) with the resolution of 30 seconds, as shown in Fig. 1.Temperature in the environment of experimental setup was 23 °C.

From these measured angles *n* can be calculated by

*δn*due to detection errors of angles

*δA*can be defined by

#### 2.3 The detection of the shift of the fringes in transmission

For evaluating temperature coefficient of refractive index, *dn/dT*, we measured
phase shifts of interference fringes in transmittance of laser materials. The transmission
*T*_{0} of thin laser material with a thickness of *L*
is ideally modulated to *T*_{r} by multiple reflections by

*R*and

*φ*are the Fresnel reflectance and the phase of modulation.

*φ*is dependent on wavelength,

*λ*and is given by 4

*πnL*/

*λ*. Although this modulation can become smaller due to imperfectness in flatness of the sample, phase shift

*Δφ*from the phase

*φ*= 4

*πnL*/

*λ*in this modulation can be easily detected, which is expressed bywhere

*ΔT*is the temperature change. Figure 2 shows the concept of multiple reflection expressed by Eq. (3) .

Experimental setup for measurement of *T*_{r} is shown in Fig. 3. The light emitted from halogen lamp (PHL-150, Mejiro
Precision Inc.) was collimated and focused to the sample that was sandwiched by temperature
controlled copper plates within the range from 15 °C to 65 °C. The temperature
was stabilized by a peltier device and its driver (LDT-5948, ILX Lightwave Co.) with the
resolution of 0.1 °C. Transmitted light from the sample was collected into a
monochrometer (TRIAX-550, HORIBA Jobin Yvon S.A.S.) and detected by an InGaAs array sensor
(IGA512-1-1, HORIBA Jobin Yvon S.A.S.) with the spectral resolution of 0.12 nm.
*Δφ* can be obtained by the least square fitting of
*T*_{r} to Eqs. (3-4).

#### 2.4 Temperature coefficient of refractive index

We evaluated the modulation in *T*_{r} of (111)-, (100)-, and (110)-cut 1.0at.% Nd:YAG plates (Fujian Castech Crystals, Inc), (100)-cut 1.0at.% Nd:YVO_{4} (ITI Electro Optics Co.), and (100)-cut 1.0at.% Nd:GdVO_{4} (Shandong Newphotons Science and Technology Co.,LTD) plates. The thickness of YAG and YVO_{4} samples were ca. 0.2 mm, and GdVO_{4} sample has thickness of ca. 0.1 mm. These thicknesses were measured by a micrometer (BMD-25DM, Mitutoyo) with the resolution of 1 μm.

Temperature coefficient of refractive index *dn/dT* can be estimated from measured *Δφ* and *α* by

*dn*/

*dT*is given from the differentiation of Eq. (5) by

*δφ*,

*δT*,

*δλ*,

*δL*,

*δα*, and

*δn*are the evaluation errors of

*Δφ*,

*ΔT*,

*λ*,

*α*and

*n*, respectively. Factors of “2” in Eq. (6) are come from the subtraction in determining of

*Δφ*and

*ΔT*.

## 3. Results

#### 3.1 Thermal expansion coefficient and refractive index

Figure 4 shows measured *α*,
and calculated *n* are shown in Fig.
5.The difference in *n* of YAG between doped sample and undoped sample is
within 0.0006, and markers for Nd-doped samples were situated at the same position as markers
for doped samples in Fig. 5. Therefore there is a
certain difference (3 times of *δn*) between *n* of
1.0at.% Nd:YAG and undoped YAG. On the contrary, differences in *n* of vanadates
are below 0.0002, and those are lower than the evaluation error.

#### 3.2 Interferometric fringes in transmission

Figure 6 shows *T _{r}* of
various samples depending on the temperature,

*T*. The ratio of the modulation depth in transmission,

*ΔT*is larger than 3% of transmission. We were able to evaluate

_{r}*φ*with the estimation error of less than 0.02 rad by use of Eq. (3). The additional peaks at 1064 nm in transmittance are considered to be due to the fluorescence from Nd

^{3+}excited by probe light.

#### 3.3 Phase shift in the fringes and temperature coefficient of refractive index

As shown in Eq. (4), the value of
*Δφ*/*φ* depends not on thickness but on
*α*, *dn/dT*, and *ΔT*.
*Δφ*/*φ* of various samples at several
important wavelengths are shown in Fig. 7, where 0.9
μm means 946 nm for Nd:YAG, 914 nm for Nd:YVO_{4}, and 912 nm for
Nd:GdVO_{4}. Similarly 1.1 μm means 1064 nm for Nd:YAG and Nd:YVO_{4},
and 1063 nm for Nd:GdVO_{4}. Also 1.3 μm means 1319 nm for Nd:YAG, 1342 nm for
Nd:YVO_{4}, and 1341 nm for Nd:GdVO_{4}. Lines in Fig. 7 are fitted by the least square method. *dn/dT*
estimated from measured *Δφ* and *α* at room
temperature are shown in Table 1. In this work room temperature means the temperature range between 15 °C
and 65 °C. The high linearity in our fringe-shifts of transmittance in Fig. 7 is an evidence of high temperature-controllability in
our measurements.

## 4. Discussions

#### 4.1 Sellmeier equations for YAG, YVO_{4}, and GdVO_{4} single crystals

In order to quantize the relationship between *n* and *λ*,
we can use Sellmeier equations. We can fit *n* of YAG crystals to Sellmeier
equations given by

*A*,

*B*,

*C*, and

*D*are fitting parameters. In the case of vanadates, we can useThese fitting parameters for refractive indices in Fig. 5 are summarized in Table 2, where

*λ*should be expressed with the unit of “

*λ*m.” In addition, the difference of

*n*between Nd-doped and undoped YAG is too small to produce a significant mistake in the evaluation of

*dn*/

*dT*.

#### 4.2 Thermal expansion coefficient along various planes in YAG single crystal

The correct measurement of *α* should provide a certain answer to the conflict between Onsager reciprocal relations and the common sense of anisotropic *α* of YAG in the solid-state laser researchers [11,12]. Authors consider that the reported anisotropy in thermal expansion of YAG should be re-evaluated with high accuracy by use of our evaluation procedure.

Similarly to Eq. 54), we can estimate *α* by
*Δφ*/*φ* and
*dn*/*dT* by

*α*along [100]- and [110]-axes in YAG crystal evaluated by use of Eq. (9) at various wavelengths are shown in Table 3.The standard deviation of measured

*α*is only 0.12 × 10

^{−6}/K, and this value is smaller than the evaluation error given by

*α*is not dependent on the wavelength of probe lights, we can reduce the estimation error in

*α*as follows:and

*α*> is the mean value of

*α*. The estimated

*α*and errors along each planes from Eqs. (11), (12) are also shown in Table 3, where differences in

*α*between crystal orientations of YAG single crystal is less than the evaluation error of 0.1 × 10

^{−6}/K ( = 1.6%) at the room temperature. Therefore, we can conclude that no anisotropy of thermal expansion in YAG crystal detected. The detailed evaluation of experimental errors are discussed in Appendix-1.

Figure 8 shows clearly that the dependence of
Δφ/φ on crystal direction is quite small. This is the reason why the
difference in calculated *α* along various orientations.

## 5. Summary

We evaluated thermo-mechanical and -optical properties of YAG, YVO_{4}, and GdVO_{4} by with less than 2% of evaluation error, and it was also revealed that thermal expansion of YAG is independent on the crystal orientation against conventional knowledge. We tabulated thermal parameters of YAG, YVO_{4}, and GdVO_{4}, and it will be the reliable data table for thermal design and heat management not only of microchip lasers but also of a plenty of solid-state lasers.

## Appendix-1 Evaluation errors for *α* and *dn/dT*

Three runs of measurements by the dilatometer brought no difference larger than 1.0 nm, and the evaluation error of measured thermal expansion was 10^{−9} m /10K / 0.025m = 4×10^{−9}/K. Thus the experimental error in α is below 1% due to the resolution of the dilatometer.

30 seconds of resolution of our spectrometer was enough large to bring no difference in reading of the value of angles, thus it was considered to be *δA*.

0.02 rad of *δφ* is the maximum of errors during least square fitting. Since small *δλ* [5] and *δn*, *δλ*/*λ* and *δn*/*n* can be ignored.

Major part of evaluation error is due to *Δφ* and *ΔT*, and those depend on the effective data interval. Under the assumption that *dn*/*dT* is constant within the measured temperature range, *ΔT* become 50°C.

and 5Parameters for error calculation in Eq. (6) and calculated errors are summarized in Table 4, respectively.

The superiority of the evaluation method for *dn*/*dT* by the
measurement of the temperature shift in the fringes on transmission is high reproducibility. In
1991, authors demonstrated the frequency tuning of Nd:YVO_{4} microchip laser [15], where the ratio of the oscillation frequency and the
frequency tuning range, *Δν*/*ν* is the same
as the ratio of the optical path of a microchip and it’s temperature deviation: this is
definitively equal to *Δφ*/*φ* in this work.
Figure 9 comparatively shows the
*Δν*/*ν* evaluated in 1991 and the
simulated value from *Δφ*/*φ* in this work
(because temperature range in these two experiments were different). The accordance with these
two lines in Fig. 9 directly proves the small
reproducible error of evaluation methods that use fringes in transmissions due to
multiple-reflections inside samples. From our error evaluations, we can conclude that
*α* and *dn*/*dT* reported in this work is
highly accurate compared with other traditional works.

## Appendix-2 Comparison to past reports on *α* and *dn/dT*

We summarize the comparison between this work to past reports on *α* and
*dn/dT* in Table 6.
*α* of YAG in this work is similar to [16], while *dn*/*dT* is almost the same as the maximum
of previously reported value [17].
*α* and *dn*/*dT* of YVO_{4} in
this work are close to [18] and [19], respectively. In the case of GdVO_{4}, *α*
in this work is proximate to [20] and
*dn*/*dT* is larger than any other reported value.

## Appendix-3 Relations between thermal effect and thermal parameters

As well as thermal conductivity *κ* [4], *α* and *dn/dT* of laser gain media are quite important thermal parameters for describing thermal effects in the laser cavity. For example, thermal lensing is severely depends on these parameters. Lens-effects inside a gain medium are caused by three kinds of changes of optical path: thermal expansion, refractive index change induced by thermal strain, and temperature dependent refractive index. The curvature of thermal lens is the summation of curvatures formed by these optical path changes. Optical path changes caused by thermal expansion and refractive index change induced by thermal strain are proportional to *α*, while optical path changes caused by temperature dependent refractive index are proportional to *dn/dT*. The focal length of thermal lens is the inverse of amount of these curvatures, and proportional to thermal conductivity. In the case of end-pumped solid-state rod lasers, the focal length of thermal lens *f* can be expressed by

*w*

_{p},

*η*

_{a},

*η*

_{h},

*P*,

*r*,

*l*, and

*C*are the radius of pump beam, the pump absorption efficiency, the thermal load, the induced pump power, the radius and length of rod, and photo-elastic coefficient for desired polarizations, respectively [29].

On the contrary, the evaluation of *α* and *dn/dT* from the pump-induced therml lensing of laser cavities requires the precise evaluations of mode-matching and thermal load which are dependent on the pump intensity. Therefore, in this case it should be important for the estimation of experimental errors to clarify the uncertainty against not only *w*_{p}, *η*_{a}, *η*_{h}, *P*, *r*, *l*, and *C* but also measured focal lens including focusing aberrations and astigmatism.

## Appendix-4 Derivation of Eq. (2)

The experimental error of the refractive index estimated from the minimum deviation angle is given by Eq. (2), which can be derived by the differentiation of Eq. (1). Differentiating by *A*_{p} gives

*A*

_{min}gives

*δn*by

## Acknowledgement

Authors thank to Prof. G. Aka for his kind help in refractive index measurement, Netzsch-Gerätebau GmbH for measurements of thermal expansion. This work was partially supported by Genesis Research Institute, Inc., and by the Special Coordination Funds for Promoting Science and Technology of the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

## References and links

**1. **T. Taira, “Domain-controlled laser ceramics toward giant micro-photonics,” Opt. Mater. Express **1**(5), 1040–1050 (2011). [CrossRef]

**2. **N. Pavel, M. Tsunekane, and T. Taira, “Composite, all-ceramics, high-peak power Nd:YAG/Cr^{4+}:YAG monolithic micro-laser with multiple-beam output for engine ignition,” Opt. Express **19**(10), 9378–9384 (2011). [CrossRef]

**3. **R. Bhandari and T. Taira, “Megawatt level UV output from [110] Cr⁴⁺:YAG passively Q-switched microchip laser,” Opt. Express **19**(23), 22510–22514 (2011). [CrossRef]

**4. **Y. Sato and T. Taira, “The studies of thermal conductivity in GdVO_{4}, YVO_{4}, and Y_{3}Al_{5}O_{12} measured by quasi-one-dimensional flash method,” Opt. Express **14**(22), 10528–10536 (2006). [CrossRef]

**5. **Y. Sato and T. Taira, “Temperature dependencies of stimulated emission cross section for Nd-doped solid-state laser materials,” Opt. Mater. Express **2**(8), 1076–1087 (2012). [CrossRef]

**6. **P. A. Loiko, K. V. Yumashev, V. N. Matrosov, and N. V. Kuleshov, “Dispersion and anisotropy of thermo-optic coefficients in tetragonal GdVO_{4} and YVO_{4} laser host crystals,” Appl. Opt. **52**(4), 698–705 (2013). [CrossRef]

**7. **Z. Huang, J. Feng, and W. Pan, “Theoretical investigations of the physical properties of zircon-type YVO_{4},” J. Solid State Chem. **185**, 42–48 (2012). [CrossRef]

**8. **X. Xu, Z. Zhao, H. Wang, P. Song, G. Zhou, J. Xu, and P. Deng, “Spectroscopic and thermal properties of Cr,Yb:YAG crystal,” J. Cryst. Growth **262**(1-4), 317–321 (2004). [CrossRef]

**9. **K. Wu, L. Hao, H. Zhang, H. Yu, Y. Wang, J. Wang, X. Tian, Z. Zhou, J. Liu, and R. I. Boughton, “Lu_{3}Ga_{5}O_{12} crystal: exploration of new laser host material for the ytterbium ion,” J. Opt. Soc. Am. B **29**(9), 2320 (2012). [CrossRef]

**10. **Private communications

**11. **A. A. Kaminskii, *Laser Crystals* (Springer-Verlag, Berlin, 1981), p. 320.

**12. **W. Koechner, *Solid-State Laser Engineering 6th ed.* (Springer Science + Business Media, Inc., New York, 2006), p. 55.

**13. **Y. Sato and T. Taira, “Thermo-optical and -mechanical parameters of Nd:GdVO_{4} and Nd:YVO_{4},” CLEO’07, paper JWA87 (2007).

**14. **M. Born and E. Wolf, *Principles of Optics, 5th ed.* (Pergamon Press, 1975) sec. 4.7.

**15. **T. Taira, A. Mukai, Y. Nozawa, and T. Kobayashi, “Single-mode oscillation of laser-diode-pumped Nd:YVO_{4} microchip lasers,” Opt. Lett. **16**(24), 1955–1957 (1991). [CrossRef]

**16. **R. L. Aggarwal, D. J. Ripin, J. R. Ochoa, and T. Y. Fan, “Measurement of thermo-optic properties of Y_{3}Al_{5}O_{12}, Lu_{3}Al_{5}O_{12}, YAlO_{3}, LiYF_{4}, LiLuF_{4}, BaY_{2}F_{8}, KGd(WO_{4})_{2}, and KY(WO_{4})_{2} laser crystals in the 80–300 K temperature range,” J. Appl. Phys. **98**, 103514 (2005). [CrossRef]

**17. **P. R. Stoddart, P. E. Ngoepe, P. M. Mjwara, J. D. Comins, and G. A. Saunders, “Hightemperature elastic constants of yttrium aluminum garne,” J. Appl. Phys. **73**(11), 7298–7301 (1993). [CrossRef]

**18. **D. Rand, D. Miller, D. J. Ripin, and T. Y. Fan, “Cryogenic Yb^{3+}-doped materials for pulsed solid-state laser applications,” Opt. Mater. Express **1**(3), 434–450 (2011). [CrossRef]

**19. **H. C. Liang, K. Y. Lin, Y. C. Lee, and Y. F. Chen, “Precise measurement of group refractive indices and temperature dependence of refractive index for Nd-doped yttrium orthovanadate by intracavity spontaneous mode locking,” Opt. Lett. **36**(19), 3741–3743 (2011). [CrossRef]

**20. **H. Zhang, J. Liu, J. Wang, C. Wang, L. Zhu, Z. Shao, X. Meng, X. Hu, and M. Jiang, “Characterization of the laser crystal Nd:GdVO_{4},” J. Opt. Soc. Am. B **19**(1), 18–27 (2002).

**21. **N. R. Reddy and K. S. Murthy, “Thermal expansion of yttrium vanadate,” J. Mater. Sci. Lett. **2**(4), 139–140 (1983). [CrossRef]

**22. **X. Peng, A. Asundi, Y. Chen, and Z. Xiong, “Study of the mechanical properties of Nd:YVO4 crystal by use of laser interferometry and finite-element analysis,” Appl. Opt. **40**(9), 1396–1403 (2001). [CrossRef]

**23. **C. V. V. Reddy, P. Kistaiah, and K. S. Murthy, “X-ray studies on the thermal expansion of gadolinium vanadate,” J. Phys. D Appl. Phys. **18**(6), L27–L30 (1985). [CrossRef]

**24. **J. Petit, B. Viana, P. Goldner, J.-P. Roger, and D. Fournier, “Thermomechanical properties of Yb^{3+} doped laser crystals: Experiment and modeling,” J. Appl. Phys. **108**(12), 123108 (2010). [CrossRef]

**25. **T. M. Pollak, W. Wing, R. J. Grasso, E. P. Chicklis, and H. Jenssen, “CW laser operation of Nd:YLF,” J. Quant. Electron. **18**(2), 159–163 (1982). [CrossRef]

**26. **D. E. Zelmon, J. J. Lee, K. M. Currin, J. M. Northridge, and D. Perlov, “Revisiting the optical properties of Nd doped yttrium orthovanadate,” Appl. Opt. **49**(4), 644–647 (2010). [CrossRef]

**27. **R. Soulard, A. Zinoviev, J. L. Doualan, E. Ivakin, O. Antipov, and R. Moncorgé, “Detailed characterization of pump-induced refractive index changes observed in Nd:YVO_{4}, Nd:GdVO_{4} and Nd:KGW,” Opt. Express **18**(2), 1553–1568 (2010). [CrossRef]

**28. **H. Y. Shen, X. L. Meng, G. Zhang, J. J. Qin, W. Liu, L. Zhu, C. H. Huang, L. X. Huang, and M. Wei, “Sellmeier’s equation and the expression of the thermal refractive-index coefficient for a Nd_{0.007}Gd_{0.993}VO_{4} crystal,” Appl. Opt. **43**(4), 955–960 (2004). [CrossRef]

**29. **W. Koechner, *Solid-State Laser Engineering 6th ed.* (Springer Science + Business Media, Inc., 2006), chap. 7.