Direct impregnation method for preparing sulfated
zirconia supported on mesoporous silica
Chang-Lin Chen
a,b, Tao Li
a, Soofin Cheng
a, Hong-Pin Lin
a,c,
Chetan J. Bhongale
a, Chung-Yuan Mou
a,*a
Department of Chemistry, National Taiwan University, 1 Roosevelt Road, Section 4, Taipei 106, Taiwan
b
College of Chemical Engineering, Nanjing University of Technology, Nanjing 210009, China
cInstitute of Atomic and Molecular Science, Academia Sinica, Taipei 106, Taiwan
Received 1 October 2001; accepted 4 October 2001
Abstract
A new method has been developed to prepare sulfated zirconia (S–ZrO2) supported on mesoporous silica. With
direct exchange of metal containing precursors for the surfactants in the as-synthesized MCM-41 materials, the problem of fill-up of the mesoporous structure was avoided and high sulfur content was achieved. By using this method, the composite of S–ZrO2/MCM-41 with ZrO2content higher than 60 wt.% can be easily obtained without serious blockage
of the pore structure of MCM-41. Nevertheless, the pore size and pore volume of the resultant S–ZrO2/MCM-41
composites were found to vary markedly with the loading of ZrO2. The strong acidic character of the obtained
composites was examined by using them as catalysts in n-butane isomerization. Introduction of other metals such as aluminum as promoter into S–ZrO2/MCM-41 can be easily conducted by the direct impregnation method. Ó 2001
Published by Elsevier Science B.V.
Keywords: Direct impregnation; Sulfated zirconia; Mesoporous silica; Butane isomerization
1. Introduction
Mesoporous molecular sieves have potential applications as catalysts for processing large mole-cules, especially those encountered in petroleum refining and pharmaceutical industries, because of their tunable uniform pore structures (2–30 nm) and large surface areas (1000 m2/g) [1,2].
How-ever, the weak strength of acidity arising from the
amorphous nature of pore walls may greatly limit their uses [3,4].
Sulfated zirconia (S–ZrO2) has attracted
inten-sive attention as superacid due to its high catalytic activity in the isomerization of small hydrocarbon molecules, especially alkanes at low temperatures. S–ZrO2 materials have motivated several recent
reviews and ongoing studies of their applications [5–7]. Corma et al. [8] reported that higher sur-face areas of the original ZrO2 material
corre-sponded to higher acid strength after sulfation. This finding suggested that the strong acid sites were associated to sites of low coordination such as defects, corners or edges found at the surface
www.elsevier.com/locate/micromeso
*
Corresponding author. Fax: +886-2-23660954. E-mail address: [email protected] (C.-Y. Mou).
1387-1811/01/$ - see front matterÓ 2001 Published by Elsevier Science B.V. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 4 5 3 - X
of small particles. It is, however, difficult to in-crease the surface area of zirconia through con-ventional preparation methods. The relatively small surface area of S–ZrO2 may limit its
appli-cations.
A composite material, which can combine the advantages of mesoporous molecular sieve and S–ZrO2, should greatly expand the catalytic
ca-pabilities of the material, especially in applications as strong acid catalysts for reactions containing bulky molecules. Up to now, only a few papers reported on the preparation of supported S–ZrO2
on MCM-41 [9–12]. For example, Gao et al. [9] prepared S–ZrO2/MCM-41 by ‘‘two-step’’
im-pregnation methods. They found that the acidity of S–ZrO2/MCM-41 increased with the increase
in ZrO2 contents. However, with Gao’s method,
the MCM-41 framework was destroyed when ZrO2 contents were greater than 30%. Kawi et al.
[10] prepared S–ZrO2/MCM-41 by chemical
liq-uid deposition and hydrolysis of Zr(OPrn) 4. Their
method was tedious and sample of only one ZrO2
content was reported. We previously prepared S– ZrO2/MCM-41 by using a one-step incipient
wet-ness impregnation method with zirconium sulfate as the precursor, followed by thermal decomposi-tion [11,12]. With this method, S–ZrO2was highly
dispersed on MCM-41. In our previous method, calcined MCM-41 was used as starting material. However, it was difficult to maintain the crystalline structure of MCM-41 in strong acidic media such as the aqueous solution of zirconium sulfate. It was also difficult to obtain S–ZrO2/MCM-41 with
high ZrO2 content.
In this work, we take a new approach to im-pregnate catalyst on the surface of mesoporous materials. Instead of impregnating on calcined porous MCM-41 materials, we chose methanol as solvent to impregnate zirconium sulfate onto the as-synthesized surfactant/silicate composite. The ion exchange process is efficient enough to induce large amount of impregnation while the meso-structure is intact. Then we use solid-state disper-sion method to disperse the zirconia. Previously, Xie and Tang [13] reported that inorganic salts could be spontaneously dispersed on amor-phous materials such as Al2O3 and SiO2. In our
new method, the problem of pore fill-up can be
avoided because during impregnation–dispersion process the surfactant is still mostly inside the channels.
In the following, we first report the new method of impregnating zirconium sulfate on the as-syn-thesized surfactant-silicate composite. Then the effect of using solid-state dispersion method to prepare S–ZrO2/MCM-41 composites is examined.
The materials were characterized with XRD, TGA as well as surface area and pore size distribution. The catalytic activity of the resultant S–ZrO2/
MCM-41 composites was studied by carrying out n-butane isomerization reaction.
2. Experimental 2.1. Sample preparation
As-synthesized pure siliceous MCM-41 was prepared using the delayed neutralization pro-cesses reported by Lin et al. [14].
The S–ZrO2/MCM-41 composites were
pre-pared by the following steps: (1) direct impregna-tion of the as–synthesized MCM-41 with desired amount of zirconium sulfate/methanol solution at about 50 °C for 5 h, (2) controlled decomposi-tion of the template remained in MCM-41 in air at 400°C for 10 h, followed by solid-state dispersion of zirconium sulfate into MCM-41 and (3) de-composition of zirconium sulfate in air at 680 °C for 3 h.
2.2. Methods of characterization
The powder X-ray diffraction data was col-lected on a Scintag X1 diffractometer using Cu-Ka radiation (k¼ 0:154 nm). The surface area and pore size distribution measurements were carried out on a Micromeritics ASAP 2000 automatic adsorption instrument using nitrogen as the anal-ysis gas. The TGA data were obtained with a Du Pont 2000 TG analyzer. The sulfur contents in the calcined samples were determined by inductively coupled plasma atomic emission spectrometry (ICP-AES) using a Jarrel-Ash ICP 9000 instru-ment.
2.3. Infrared studies
Diffuse reflectance infrared Fourier transform (DRIFT) spectra of the S–ZrO2/MCM-41 sample
after adsorbing pyridine were recorded. DRIFT spectra of the samples were recorded using a BOMEM MB155 FT-IR/Raman spectrometer. The equipment was furnished with an in situ sample cell (Harrick). The sample was pre-heated at 300 °C for 3 h under 106 mbar vacuum
be-fore pyridine vapor was introduced at room tem-perature, followed by evacuation for 30 min. Spectra were acquired from room temperature to 500°C.
2.4. Catalytic studies
The composites of S–ZrO2/MCM-41 were
tes-ted as catalysts for n-butane isomerization in a fixed-bed continuous flow reactor. The reactor was operated at atmospheric pressure. Approximately 1.0 g of the catalyst was loaded into the reactor and then pretreated for 3 h in flowing dry air (60 ml min1) at 450°C. The reactor temperature was
then lowered to the reaction temperature of 250 °C. The feed gas n-butane/H2 mixtures (1:10 v/v)
flowed through the catalyst bed at n-butane weight hourly space velocity (WHSV) of 0.3 h1. The flow
rate was monitored using a Brooks mass flow controller. An on-line Shimadzu 14Bgas chro-matograph equipped with FID was used to ana-lyze the reaction products.
3. Results and discussion
3.1. Thermal decomposition analysis
It was found that the crystalline structure and surface area of MCM-41 were better retained when as-synthesized MCM-41 was used as starting material than the calcined sample in preparation of impregnated catalysts. That is because the space-filling template inside the pores would pre-vent the complete fill-up of the pore volume with the impregnated precursor. Therefore, in the cur-rent preparation method, as-synthesized MCM-41 was used instead of the calcined one. TG analysis
was carried out to determine the proper tempera-tures for sample preparation.
Fig. 1 shows the TGA profile taken in air of Zr(SO4)2impregnated on as-synthesized MCM-41,
in comparison to those of as-synthesized MCM-41 and Zr(SO4)2. For all the samples, we first
equili-brate them at 120 °C for 1 h to remove water. Curve (a) shows that most of the surfactant in as-synthesized MCM-41 decomposed in the tempera-ture range of 200–350 °C. Curve (b) shows that the decomposition of Zr(SO4)2is negligible before
500 °C, but around 600–680 °C most of Zr(SO4)2
decomposed.
Curve (c) is the TG profile of Zr(SO4)2
im-pregnated on as-synthesized MCM-41. There were several weight loss steps. In comparison to the other two profiles, the weight loss occurred at temperatures lower than 400°C is likely due to the decomposition of surfactant. The weight loss at temperature higher than 600 °C is due to the de-composition of Zr(SO4)2. In between these two
temperatures should be a continuous decom-position of surfactant followed by a gradual decomposition of Zr(SO4)2. The decomposition
of surfactant was apparently obstructed when Zr(SO4)2was present. Two reasons may contribute
to this phenomenon. One is that there is stronger interaction between sulfate and surfactants. The other is that some small Zr(SO4)2crystallites may
block the pore mouths of MCM-41 so that the
Fig. 1. TGA profiles analyzed in air (flow rate¼ 60 ml min1)
of (a) as-synthesized MCM-41, (b) Zr(SO4)2and (c) Zr(SO4)2
surfactant and its fragments cannot evaporate as easily as that in the pristine MCM-41. On the other hand, the shift of the decomposition tem-perature of Zr(SO4)2 toward lower temperature is
probably due to it small crystallite size. 3.2. X-ray diffraction
Fig. 2 shows the XRD patterns of Zr(SO4)2
impregnated on as-synthesized MCM-41 with 50 wt.% ZrO2 loading after different heating
treat-ment. Fig. 2(a) shows that the diffraction peaks corresponding to both the crystalline Zr(SO4)2and
MCM-41 structures were present after the
im-pregnated sample was dried at 100°C. This result indicates that most of the Zr(SO4)2 crystallites at
this stage are relatively large crystals and located on the outer surfaces of the MCM-41 particles. After the sample was heated at 400°C for 10 h, it is of interest to note that the characteristic peaks due to crystalline Zr(SO4)2 disappear completely,
as shown in Fig. 2(b). According to the results of TG analysis, at 400°C the surfactant in MCM-41 should have decomposed, but Zr(SO4)2 should
not. The disappearance of the XRD peaks of crystalline Zr(SO4)2 may be explained by that
Zr(SO4)2moves into the channels of MCM-41 and
disperses onto the interior surfaces as the surfac-tant has moved out. Recently, Xie et al. [15] stud-ied the dispersion of Zr(SO4)2 on silica gel. They
revealed that the apparent dispersion threshold of Zr(SO4)2 is ca. 0.26 g/100 m2 on silica surface.
The surface area of our parent MCM-41 is about 1010 m2/g. In 50 wt.% ZrO
2/MCM-41 sample, the
amount of Zr(SO4)2 started with is very close to
the apparent dispersion threshold reported in that paper.
Fig. 2(c) shows that the small peaks corre-sponding to tetragonal ZrO2phase appeared when
the sample was further heated at 650 °C for 3 h. The intensity of these peaks increases when the temperature of the second heating stage was in-creased to 680°C (Fig. 2(d)).
Fig. 3 shows the XRD profiles of the composi-tes of S–ZrO2/MCM-41 with different ZrO2
con-tents. The appearance of the diffraction peaks at low angle region indicates that the ordered hexa-gonal arrangement of MCM-41 was still retained. In other words, the pore structure of MCM-41 was not destroyed during the solid dispersion pro-cesses. However, the intensity does decrease with the increase in ZrO2loading.
Fig. 3 shows that in the high angle region only the diffraction peaks due to tetragonal ZrO2phase
were observed on samples of low ZrO2 contents.
On the other hand, diffraction peaks correspond-ing to both tetragonal and monoclinic ZrO2 were
observed for the samples of high ZrO2 contents.
These results are consistent with that observed on silica gel supported Zr(SO4)2[15]. Monoclinic ZrO2
phase was also observed when the Zr(SO4)2loading
was over the apparent dispersion threshold.
Fig. 2. XRD patterns of Zr(SO4)2supported on as-synthesized
MCM-41 with final composition of 50 wt.% ZrO2/MCM-41
after various heat treatments: (a) 100°C, 12 h; (b) 400 °C, 10 h; (c) 400°C, 10 h then 650 °C, 3 h and (d) 400 °C, 10 h then 680°C, 3 h.
It has been reported that sulfated ZrO2of
meta-stable tetragonal phase has higher catalytic activity than the monoclinic phase [16,17]. From this point of view, among the S–ZrO2/MCM-41 samples, the
one with about 50% ZrO2content should have the
highest catalytic activity because this sample has the largest amount of tetragonal ZrO2.
3.3. Physico-chemical properties of the samples Table 1 shows the sulfur content, BET surface area, and total pore volume of series S–ZrO2/
MCM-41 samples calcined at 680 °C. It can be seen that the sulfur content increases with the zirconia loading. In contrast, both the BET surface area and pore volume of the samples decrease as the zirconia content increases. Fig. 4 shows the pore size distributions in comparison to that of
parent MCM-41. The pore diameter was found to decrease gradually with the increase in zirconia loading. The pore diameter reduced from about 2.9 nm for the parent MCM-41 to about 2.5 nm for the sample of 30 wt.% ZrO2, and to 2.2 nm
for that of 50 wt.% ZrO2. It is also noticeable that
the pore size distributions remain rather narrow for the S–ZrO2/MCM-41 samples. These results
confirm that the S–ZrO2 is highly dispersed onto
the interior surfaces of the mesopores of MCM-41. The uniform decrease of pore size, volume and surface area is an indication that most of the loading of zirconia is inside the pores. However, at loading higher than 50% the decrease in pore volume slightly levels off and the sulfur content increases markedly as shown in Table 1. The 50
Fig. 3. XRD patterns of the composites of S–ZrO2/MCM-41
with different ZrO2contents calcined at 680°C for 3 h in air.
Table 1
Physico-chemical properties of the samples Wt.% ZrO2in S–ZrO2/MCM-41a BET S.A. (m2/g) Pore vol-ume (ml/g) Sulfur con-tent (wt.%) 30 667 0.57 1.34 40 571 0.47 1.61 50 446 0.41 1.81 60 338 0.41 2.83 70 301 0.34 4.95 a
All samples were calcined at 680°C for 3 h.
Fig. 4. The pore size distribution curves of MCM-41 and var-ious S–ZrO2/MCM-41 samples (calcined at 680°C for 3 h).
wt.% ZrO2 content on MCM-41 coincides with
the dispersion threshold of zirconium sulfate on MCM-41. This probably shows that further load-ing results in formation of zirconia particles on the external surface of MCM-41.
DRIFT spectrum of calcined MCM-41 and that of S–ZrO2/MCM-41 with 50 wt.% ZrO2 were
compared. The spectra shown in Fig. 5 were taken after the samples were heated at 300°C for 1 h to drive away the physical adsorbed water. MCM-41 gives a strong absorption peak at 1353 cm1,
which corresponds to Si–O stretching vibration. In absorption spectra, usually Si–O stretching ap-pears in the 1000–1200 cm1 region. In the
diffuse-reflectance spectra, the peak shifts toward higher frequencies due to differences in the specular re-flectance component [18]. DRIFT spectrum also shows a sharp peak of relatively weak intensity at 3742 cm1 in the O–H vibration region. The latter
is associated with the free SiO–H groups on the surface of MCM-41. The intensity of both peaks was markedly reduced when S–ZrO2 was
impreg-nated on MCM-41. These results indicate that the silica surface of MCM-41 is covered by a layer of S–ZrO2in S–ZrO2/MCM-41.
3.4. Acidity of S–ZrO2/MCM-41
Fig. 6 shows the DRIFT spectra of the pyridine adsorbed sample with 50% ZrO2loading. It can be
seen that the intensity of the 1445 and 1595 cm1
peaks decreases and that of the latter even disap-pears after heating the sample above 200°C. These two peaks according to Parry [19] are assigned to H-bonding pyridine. The remaining three peaks are attributed to pyridine adsorbed on Lewis acid sites. Accompanying with the disappearance of H-bonding pyridine, the SiO–H vibration peak at 3742 cm1 appears. On the other hand, the
char-acteristic peak of pyridinium ions at 1540 cm1
was not observed. In other words, the surface
Fig. 5. DRIFT spectra of (a) calcined MCM-41 and (b) S– ZrO2/MCM-41 (50 wt.% ZrO2) after the powder samples were
heated at 300°C for 1 h.
Fig. 6. DRIFT spectra of S–ZrO2/MCM-41 (50 wt.% ZrO2)
after adsorption of pyridine and evacuation at different tem-peratures.
SiO–H groups only coordinate to pyridine through H-bonding and S–ZrO2/MCM-41 has only Lewis
acid sites but not Br€oonsted acid sites. With the increase in desorption temperature from 200 to 500 °C, a slight decrease in the DRIFT peak in-tensity was observed. However, the preservation of the three peaks associated with Lewis acid sites upon 500°C heating of the sample implies that the strength of the Lewis acid sites on S–ZrO2
/MCM-41 catalyst is rather strong.
3.5. Catalytic activity of S–ZrO2/MCM-41 in
n-butane isomerization
Isomerization of n-butane to isobutane was chosen as a model reaction to test the strong acidity of S–ZrO2/MCM-41 prepared by direct
impregnation method. In the isomerization of n-butane over S–ZrO2/MCM-41, the selectivity to
isobutane was found to be higher than 95% with only minor amounts of methane, propane and pentane being formed. The variation of the con-version at 250°C with time on stream for S–ZrO2/
MCM-41 of different ZrO2loadings is given in Fig.
7. Generally, the initial conversion increases with the zirconia content on S–ZrO2/MCM-41.
How-ever, a maximum initial conversion was observed on the sample of 60 wt.% ZrO2loading. Moreover,
the catalytic activities decayed with time on stream. It is noticeable that after 0.5 h on stream, the activity of S–ZrO2/MCM-41 with 50 wt.%
ZrO2 is the one giving highest activity. The 50
wt.% ZrO2 content on MCM-41 coincides with
the dispersion threshold of zirconium sulfate on MCM-41. From the XRD results, it seems to be the threshold for formation of pure tetragonal zirconia phase on MCM-41.
When comparing the catalytic activities of S–ZrO2/MCM-41 in n-butane isomerization with
those of S–ZrO2 supported on other supports
re-ported in the literature, S–ZrO2/MCM-41 has
similar activity as that supported on SiO2[20], but
lower activity than those supported on Al2O3[20]
and Al2O3–ZrO2 [21]. This phenomenon indicates
that the different surface properties of the supports play a very important role. Al can act as a pro-moter to increase the catalytic activity. A small amount of Al was introduced into S–ZrO2
/MCM-41, and the sample is abbreviated as SZA/MCM-41. It was found that the catalytic activity in n-butane isomerization could be largely improved. The results are shown in Fig. 8. For comparison, the activity of S–ZrO2supported on silica gel (SZA/
SiO2) is also shown in Fig. 8. It is obvious that
SZA/MCM-41 has much higher catalytic activity. The high degree of dispersion of our catalyst has several advantages: (1) Its high loading of S– ZrO2 is favorable for future detailed molecular
studies of the catalytic mechanism [22]. (2) The effect of promoters can be studied more effectively with these catalysts of larger surface area. More studies on the molecular level of alkane isomeri-zation with this effective acid catalyst system in-corporated with aluminum will be reported.
4. Conclusions
In this article, we have successfully demon-strated that direct impregnation method is an ef-fective way to prepare the composites of S–ZrO2/
MCM-41. With this method, S–ZrO2/MCM-41 of
high ZrO2 loading can be easily obtained without
damaging the pore structure of MCM-41. The pore size and volume were changed significantly
Fig. 7. Conversions of n-butane versus time on stream over S– ZrO2/MCM-41 (calcined at 680°C for 3 h) of different ZrO2
with different ZrO2loadings. The catalytic studies
in isomerization of n-butane show that the resul-tant composites of S–ZrO2/MCM-41 have strong
acidic properties. The S–ZrO2/MCM-41 catalyst
with 50 wt.% ZrO2 loading gave the highest
ac-tivity after 0.5 h on stream.
The direct method of impregnating metal oxide catalyst into the internal surface of mesoporous materials is general. Previously, we have demon-strated its use in direct surface functionalization of mesoporous silica [23]. The surface reaction of functionalization strongly derived the direct exchange while maintaining the mesostructure. It was also applied to supporting TiO2
photocat-alysts on MCM-41 [24]. By dispersing the metal containing precursors on the as-synthesized MCM-41 materials, the problem of complete fill-up of the mesopores is avoided. One can thus prepared supported catalysts of rather high surface area and with narrowly distributed pore diameter.
Acknowledgements
The financial support from China Petroleum Corporation, Taiwan is gratefully acknowledged.
C.-L. Chen thanks the financial support given by the Education Commission of Jianshu Province, China (Project 00KJB530001).
References
[1] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Back, Nature 359 (1992) 710.
[2] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548. [3] A. Corma, V. Fornes, M.T. Navarro, J. Perez-Pariente,
J. Catal. 148 (1994) 569.
[4] A. Sayari, Chem. Mater. 8 (1996) 1840.
[5] X.M. Song, A. Sayari, Catal Rev. Sci. Eng. 38 (1996) 329.
[6] V. Adeeva, H.Y. Liu, B.Q. Xu, W.M.H. Sachtler, Top. Catal. 6 (1998) 61.
[7] G.D. Yadav, J.J. Nair, Micropor. Mesopor. Mater. 33 (1999) 1.
[8] A. Corma, M.I. Juan-Rajadell, J.M.L. Nieto, Appl. Catal. A 116 (1994) 151.
[9] T. Lei, W.-M. Hua, Y. Tang, Y.-H. Yue, Z. Gao, Chem. J. Chinese Univ. 21 (2000) 1240.
[10] Q.-H. Xia, K. Hidajat, S. Kawi, Chem. Commun. (2000) 2229.
[11] C.-L. Chen, H.-P. Lin, S.-T. Wong, S. Cheng, C.-Y. Mou, Proceedings of the 3rd seminar on Science and Technology Catalysis, Fukuoka, Japan, March 9–10, 2000, p. 95. [12] C.-L. Chen, S. Cheng, H.-P. Lin, S.-T. Wong, C.-Y. Mou,
Appl. Catal. A 215 (2001) 21.
[13] Y.C. Xie, Y.-Q. Tang, Adv. Catal. 37 (1990) 1.
[14] H.-P. Lin, S. Cheng, C.-Y. Mou, Micropor. Mater. 10 (1997) 111.
[15] Y.-Y. Huang, B.-Y. Zhao, Y.-C. Xie, Appl. Catal. A 173 (1998) 27.
[16] C. Morterra, G. Cerrato, F. Pinna, M. Signoretto, J. Catal. 157 (1995) 109.
[17] R.L. White, E.C. Sikabwe, M.A. Coelho, P.E. Resasco, J. Catal. 157 (1995) 755.
[18] P.W. Yang, H.H. Mantsch, F. Baudais, Appl. Spectr. 40 (1986) 974;
G. Kortium, W. Braun, G. Herzog, Angew. Chem. 2 (1967) 333.
[19] E.P. Parry, J. Catal. 3 (1963) 371.
[20] T. Lei, J.S. Xu, Y. Tang, W.M. Hua, Z. Gao, Appl. Catal. A 192 (2000) 181.
[21] Y. Xia, W. Hua, Z. Gao, Appl. Catal. A 185 (1999) 293.
[22] T.S. Yang, T.H. Chang, C.H. Lin, C.T. Yeh, J. Mol. Catal. A 159 (2000) 397.
[23] H.-P. Lin, L.-Y. Yang, C.-Y. Mou, S.-B. Liu, H.-K. Lee, New J. Chem. 24 (2000) 253.
[24] Y.-H. Hsien, C.-F. Chang, Y.-H. Chen, S. Cheng, Appl. Catal. B. 31 (2001) 241.
Fig. 8. Comparison of the catalytic activity for n-butane iso-merization over SZA/MCM-41 and SZA/SiO2(both catalysts
with the same Al and Zr contents were calcined at 720°C for 3 h in air).
