1
Investigation of the surface structure of zeolite A

2
The surface structure of zeolite A was investigated by FE-SEM, AFM, and HRTEM.

3
First, it was shown by FE-SEM and AFM that the surface of zeolite A terminated with the same structure.

4
Then, by combining the results with those of HRTEM, the terminal structure of zeolite A was identified as incomplete sodalite cages.

5
Because of their well-defined structures and large ion-exchange capabilities, zeolites are widely used as ion-exchangers, adsorbents and catalysts.1,2

6
In contrast to the formations of typical metal oxides, zeolite crystallization is affected by various factors (e.g. starting composition, pH, temperature, mixing procedures, agitation and synthesis time).

7
Thus, the crystallization mechanisms of zeolites are not well understood, and their elucidation remains a major topic in zeolite science.3

8
High-resolution transmission electron microscopy (HRTEM) can provide detailed information on the structures of zeolites and related materials;4–14 HRTEM images and selected area electron diffraction data have advantages over X-ray diffraction data since small specimens can be analyzed.6,7

9
It is also possible to reveal the terminal structure of zeolites directly from HRTEM images,8–14 which enables information on the crystal growth mechanisms of zeolite to be obtained.

10
However, there is a drawback as it is difficult to get a whole image of a sample by HRTEM, because to be observed by HRTEM the observation area must be very thin, and therefore, is usually limited.

11
Further, it is well known that zeolites with a low Si/Al ratio are very fragile under the irradiation of an electron beam; therefore, most of the terminal structures of zeolites remain unelucidated.

12
Zeolite A is one of the aluminosilicate zeolites, which contain three-dimensional pores within the structure, and is widely utilized in industrial fields as adsorbents, ion-exchangers (builders), and so on.

13
The crystal growth mechanism of zeolite A was recently investigated using AFM15–21 and a terminal structure was proposed from detailed analysis of the step height of the surface; however, direct determination of the terminal structure by HRTEM has not yet been reported.

14
Observation of zeolite A is quite difficult because it contains a maximum amount of Al in the structure (Si/Al = 1) and is easily damaged by electron irradiation.

15
To overcome the above noted drawbacks of HRTEM, we combined observations by three types of microscopes; a field emission scanning electron microscope (FE-SEM; Hitachi, S-900, horizontal resolution is 0.8 nm at 6 kV), an atomic force microscope22 (AFM; Nanoscope IIIa: Veeco Metrology Group, Digital Instruments) and an HRTEM (JEM 2010, JEOL).

16
In this study, it is first shown by FE-SEM and AFM that the surface of zeolite A characterized here is terminated with the same structure.

17
Then, the terminal structure of zeolite A is identified by HRTEM.

18
Finally, a key precursor for crystal growth is discussed.

19
A large single crystal of zeolite A was prepared using triethanolamine (TEA; Wako Pure Chemical Industries Ltd.; purity 98%).23,24

20
The aluminosilicate solution was prepared by dissolving sodium metasilicate nonahydrate (Na2SiO39H2O, Wako Pure Chemical Industries Ltd.), sodium aluminate (NaAlO2; Wako Pure Chemical Industries Ltd.; Al/NaOH is about 0.75), and TEA in distilled water.

21
2.86 g of Na2SiO39H2O and 3.72 g of TEA were added to 20.0 g of distilled water in a polyethylene bottle.

22
In another polyethylene bottle, 2.29 g of NaAlO2 and 3.72 g of TEA were added to 20.0 g of distilled water, and stirred for 10 min.

23
Both of the solutions were filtered through a 0.2 μm filter membrane, and the latter was then poured into the former, which had been stirred for 5 min.

24
All procedures were performed at room temperature.

25
Finally, the solution was placed in an oven at 353 K and heated for two weeks without stirring.

26
The products obtained were filtered, washed with distilled water, and dried at 323 K for 24 h.

27
For HRTEM observation, gold particles were attached to the surface of zeolite A to distinguish the original external surface from the surfaces cleaved during the sample preparation.

28
At first, 0.01 g of zeolite A and 0.1 ml of a solution of gold particles in ethanol (0.05 wt%, ca.

29
5 nm, Nippon Paint Co., Ltd.) were mixed.

30
The mixture was then dried in an oven at 323 K for 24 h (see ESI).

31
The crystals recovered were crushed, fixed on a Cu mesh grid and observed by HRTEM.

32
As zeolites are very sensitive to electron beam irradiation, low-dose HRTEM observation must be carried out, and so HRTEM photographs were obtained with negative films at ca.

33
1/30 electron beam density of ordinary observation conditions (0.05–0.1 A cm−2).

34
Acceleration voltage was 200 kV.

35
The terminal structure was simulated using a Crystal Kit and a Mac Tempas (Total Resolution).

36
The thickness of the sample and the focus condition calculated were 20 nm and −60 nm, respectively.

37
Fig. 1a shows an FE-SEM image of zeolite A synthesized in this study.

38
The crystals have a uniform cubic morphology in a size range of 5–50 μm, and the squared faces observed were assigned as {100}.

39
Fig. 1b shows a magnified FE-SEM image of zeolite A; steps can be clearly observed.

40
The image observed here supports the layer-by-layer growth model as reported by Agger et al.16

41
Fig. 1c and 1d show an AFM image of zeolite A and its cross-sectional profile, respectively.

42
As with the result by FE-SEM (Fig. 1b), steps without adhesion of amorphous matter can be clearly observed on the surface of zeolite A. A cross-sectional profile indicates that the steps observed are all ca.

43
1.23 nm high in this condition.

44
The framework structure of zeolite A, shown in Fig. 2a, consists of sodalite cages connected by double four membered rings (D4Rs).

45
A central void in the structure is called an α-cage.

46
In Fig. 2, four possible terminal structures of zeolite A surface are shown.

47
The surface can be terminated either with double four membered rings (b-1), complete sodalite cages (b-2), incomplete sodalite cages (b-3), or α-cages (b-4).

48
A step height of 1.23 nm corresponds to the characteristic length of the structure along 〈111〉 as reported by Agger16 and Sugiyama.19

49
From AFM observation, however, it is impossible to determine which of the four structures represents the real terminal surface at this stage; it can only be concluded that the steps observed by FE-SEM and AFM are 1.23 nm high, and that all steps are terminated with the same face.

50
Next, the cross-section of the surface was observed by HRTEM.

51
Large crystals must be pulverized for HRTEM observation; however, it is difficult to distinguish an original external surface from a cleaved surface.

52
Thus, gold particles were used as markers to determine the original surface (see Experimental section and ESI).

53
In this study, clear HRTEM images were successfully obtained although zeolite A is usually easily damaged by the electron irradiation (an HRTEM image of the internal structure is shown in the ESI).

54
However, it was quite difficult to get clear images of the external structure of zeolite A due to Fresnel fringes, since it is difficult to prepare a sample thin enough to be observed.

55
Fig. 3 shows the best image of zeolite A taken from the [100] direction.

56
Gold particles are observed on the surface, which guarantees that the observed surfaces are not cleaved ones (a low magnification image is shown in the ESI).

57
Fig. 4a, b, c and d shows enlarged images of Fig. 3.

58
Note that one image is duplicated in Fig. 4a, b, c and d for comparison.

59
Four different kinds of simulated terminal structures are shown in Fig. 4e, f, g and h, and the corresponding framework structures are illustrated in Fig. 4i, j, k and l, respectively.

60
Si and Al atoms are marked in the simulated images in Fig. 4e, f, g and h.

61
The four terminal structures are as follows;

62
4i: terminated with double four membered rings (D4Rs).

63
4j: terminated with complete sodalite cages (four membered rings (4Rs) are removed from the surface of Fig. 4i).

64
4k: terminated with incomplete sodalite cages (D4Rs are removed from the surface of Fig. 4i).

65
4l: terminated with α-cages.

66
By comparing the observed and simulated images, the models in Fig. 4i and j are rejected, since the shaded part on the surface is different from that of the HRTEM image (see arrows).

67
The model in Fig. 4l is also different from the HRTEM image, since the light and shade simulated is different from that in the HRTEM image (see arrows).

68
Thus, the model in Fig. 4k gives the best fit.

69
By comparing the best image observed with the simulated one, the terminal structure of zeolite A as synthesized in this study can be identified as incomplete sodalite cages, as shown in Fig. 4k.

70
This result supports one of the terminal structures proposed by Sugiyama et al19. although the result obtained is not always applicable to all zeolite A synthesized by various ways; therefore, there is room for improving the analysis using various methods including HRTEM, AFM and computational calculation.25

71
It also suggests that the precursors contributing to the crystal growth are smaller than sodalite cages, such as 4R or D4R.

72
In summary, the surface structure of zeolite A was investigated by the combined analysis of FE-SEM, AFM, and HRTEM.

73
By FE-SEM and AFM, it was shown that the surface of zeolite A characterized here was terminated with the same face.

74
Then, by HRTEM the terminal structure was revealed to be incomplete sodalite cages.