Cheng-Dun
Li
a,
Ting-An
Lin
a,
Po-Hsiung
Chen
b,
Tsai-Sheng
Gau
bc,
Burn-Jeng
Lin
bc,
Po-Wen
Chiu
*b and
Jui-Hsiung
Liu
*a
aDepartment of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan. E-mail: rsliu@mx.nthu.edu.tw
bTSMC-NTHU Joint Research Center, National Tsing Hua University, Hsinchu 30013, Taiwan
cCollege of Semiconductor Research, National Tsing Hua University, Hsinchu 30013, Taiwan
First published on 9th April 2024
This work reports the synthesis and characterization of a novel pentameric tin chloro cluster, (vinylSn)3Sn2Cl5O2(OH)2(t-BuCO2)6 (1), and explores its application as an efficient negative-tone photoresist in a 1:2 weight ratio blend with [(n-BuSn)12O14(OH)6](BF4)2 (2). Through e-beam lithography, a small high-resolution pattern (HP = 20 nm) is achieved for the blend photoresist (3) at a dose of 2080 μC cm−2. Additionally, EUV lithography demonstrates the development of a high-resolution pattern (HP = 16 nm) at an EUV dose of 70 mJ cm−2. Mechanistic studies by reflective FTIR indicate a significant decomposition of Sn–carbon and SnO2(t-Bu) moieties starting at J = 35 mJ cm−2, which is accompanied by growth of the Sn–O absorption intensity. A collapse of the cluster frameworks of clusters (1) and (2) is observed at J > 70 mJ cm−2. High-resolution X-ray photoelectron spectroscopy (HRXPS) reveals that low EUV light predominantly decomposes Sn–butyl and Sn–Cl bonds. As EUV doses increase, primary photolytic reactions involve cleavage of Sn–butyl, Sn–O2CBut, and Sn–vinyl bonds. Notably, the photolytic decomposition of Sn–Cl bonds is distinctive, with only two out of five bonds being cleaved, even at high EUV doses, resulting in a break in film growth at J = 27–35 mJ cm−2 in the EUV contrast curve. Moreover, HRXPS analysis suggests that radical propagation on the vinyltin end of the blend is unlikely, providing concise mechanistic insights into the photochemical processes governing the behavior of this advanced photoresist.
While metal clusters incorporating multiple chloro ligands are well-documented,26 their application as negative-tone photoresists remains unexplored. Polynuclear chloro clusters present potential advantages, including the susceptibility of weak metal-chloride bonds to cleavage by EUV light, compact molecular packing facilitated by small chloro sizes, and enhanced EUV absorption efficiency compared to oxygen and carbon-based ligands. However, the utilization of metal chloro clusters as EUV photoresists is challenged by the reactivity of M–Cl with surface Si–OH, resulting in the formation of less soluble Si–O–M material.
In this work, we report the synthesis of a novel pentameric tin cluster, (vinylSn)3Sn2Cl5O2(OH)2(t-BuCO2)6 (1), featuring five Sn–Cl ligands arranged in a highly irregular geometry. While cluster (1) alone is an unsuitable photoresist due to incomplete film removal after a PAB (post-apply bake) treatment, its application in an EUV photoresist is realized through strategic blending. By combining it with another cluster, [(n-BuSn)12O14(OH)6](BF4)2 (2), in a weight ratio of 1:2, the resulting blend emerges as an efficient EUV photoresist, enabling the fabrication of a small half-pitch (HP) pattern of 16 nm at a dose J = 70 mJ cm−2. This performance surpasses that of its major component (2), which necessitates higher energy doses with J > 200 mJ cm−2 (see Fig. 8, vide infra) to achieve a comparable HP = 14–15 nm pattern. Characterization studies, including Fourier-transform infrared spectroscopy (FTIR) and high-resolution X-ray photoelectron spectroscopy (HRXPS), are employed to elucidate the aggregation mechanism of this novel photoresist blend.
Fig. 1 shows the thermogravimetric analysis (TGA) of cluster (1) to study its thermal decomposition process. The cluster is stable below 160 °C, but gradually loses weight from 160 °C to 410 °C at three stages according to the changes in curve slopes. The first stage (160–210 °C) corresponds to a small weight loss (12%), probably due to the loss of water or HCl. The former arises from a dehydration of two Sn–OH groups and the latter occurs with the reaction between the Sn–Cl and Sn–OH bonds. Small decomposition of Sn–carboxylate ligands is also likely to occur here. The second stage (210–312 °C) shows a large weight loss (ca. 35%), probably due to the thermal decomposition of Sn–butyl, Sn–vinyl and Sn–carboxylate ligands. A final stage (320–410 °C) is notable, involving a 17 wt% loss, due to unknown reactions. Loss of volatile tin compounds like vinylSnCl3, SnCl4 or SnCl2 likely occurs in this stage because the remaining weight is only 38.9%, far smaller than the expected 49.2% value for SnO2. Notably, cluster (1) has the Sn element alone ca. 39.6%, showing a large loss of tin species at the end of the TGA curve. The composition of this residue is further studied by XPS study (vide infra, see Fig. 11) that shows the following composition for this residue: Sn:C:O:Cl = 5:4.9:15.2:0.5; herein cluster (1) has the following element ratio: Sn:C:O:Cl = 5:36:15:5. Sufficient oxygen content remains in the final residues. Residues with a formula such as SnClxO2–0.5x are expected to have residual weight >49.2%. With this XPS study, the residue seems to have SnO2 as the main component together with a small proportion of Sn–Cl, and Sn–carboxylate residues. However, a loss of volatile tin content is obvious because the remaining weight ca. 38.9% is too small.
Our primary focus lies in harnessing the potential of cluster 1 as both an e-beam and EUV photoresist. However, employing this cluster alone results in the formation of an insoluble film, even with a mild PAB process (60 °C, 60 s). The film persists even after treatment with various solvents such as acetone, 2-heptanone, toluene, 4-methyl-1-pentanol, and isopropanol. Although a 1% aqueous solution of tetramethylammonium hydroxide (TMAH) is effective in removing the PAB-treated film, it also removes the e-beam exposure photoresist. To address these challenges, we explore the use of cluster 1 in a blend, combining it with [(n-BuSn)12O14(OH)6]X2 (X = BF4, 2).27 The selection of the 12-tin cluster (2) is based on its ability to form a smooth film surface, in contrast to its hydroxide analogue (X = OH), which has been reported to exhibit microcrystalline morphology.22,28 Additionally, we are aware that the 12-tin cluster (X = BF4, 2) is a very efficient photoresist to develop very high resolution EUV patterns with HP = 12–16 nm under high EUV doses (>200 mJ cm−2). Some representative patterns are shown in Fig. 8 (vide infra) and S3.† In our investigation of this new blend, we explore weight ratios with (1)/(2) < 0.5, as an increased proportion of cluster (1) in the blend results in incomplete removal of thin films with developers after the PAB treatment (Scheme 2).
Thin films of clusters (1), (2), and their blend photoresist (3) were characterized using optical microscopy (OM) and atomic force microscopy (AFM), and the results are illustrated in Fig. 2. The films were prepared at a concentration of 1.5 wt% in 4-methyl-2-pentanol before spin coating. A PAB process was conducted at 80 °C for 60 s, followed by cooling to room temperature for 24 hours. With the thickness ranging from 15.8 to 25.0 nm, these photoresists exhibited a uniform and featureless appearance, without visible microcrystalline structures observed over the domain of 500 μm × 600 μm, as depicted in the OM images. The surface topography of the films was examined at a higher resolution, revealing remarkably smooth surfaces with roughness values of approximately 1.08 nm, 1.74 nm, and 0.93 nm for photoresists (1), (2), and (3), respectively, over the domain of 5 μm × 5 μm. To assess the solubility of these films, various developers were employed. Notably, photoresist (1) exhibited resistance to complete removal by common solvents, with a residual thickness of 5–6 nm even after two minutes of developer cleaning.
Fig. 2 Surface characterization of photoresists (1), (2) and (3) by OPM and AFM; the films are treated by a PAB process (80 °C, 60 s). |
E-beam exposure contrast curves wereobtained to assess the photosensitivity of the thin films of photoresists 1, 2, and 3, following PAB treatment (80 °C, 60 s). No post-exposure bake (PEB) treatment was applied to these photoresists. These curves illustrate the remaining fraction of the exposed resist, not removed by the developer after exposure, as a function of e-beam doses. In Fig. 3, the contrast curve of cluster (1) is unattainable due to a PAB process that reduces solubility of the unexposed photoresist with the developer 2-heptanone. In contrast, the contrast curves of photoresists (2) and (3) exhibit typical patterns for negative-tone photoresists. The critical energy is determined to be 1600 μC cm−2 for the 12-tin cluster (2), while this energy is drastically reduced to 500 μC cm−2 for the blend photoresist (3). Remarkably, the remaining height after 500 μC cm−2 is 17.8 nm for blend (3), representing a modest 15% loss in height compared to its initial height of 21.0 nm. The e-beam photosensitivity of blend material 3 demonstrates that it could be a suitable photoresist due to its low critical energy and a sharp curve slope. We also compared the photosensitivity between 12-tin cluster (2) and the well-known [(n-BuSn)12O14(OH)6](OH)2 (2′)25; the latter has a small critical energy J = 1360 μC cm−2, but with a small curve slope. In Fig. 3, the height of the exposed photoresist (2′) increases very slowly over a wide range of EUV doses (J = 550–1360 μC cm−2); this feature is unlikely to produce a high-resolution pattern. Accordingly, we selected 12-tin cluster 2 as the main component to blend with cluster (1).
E-beam lithographic patterns were generated for photoresist (3) under both no PEB and PEB conditions (80 °C, 60 s), with the corresponding images shown in Fig. 4 and 5, respectively. Photoresist 3 underwent e-beam exposure with doses ranging from 450 to 2720 μC cm−2 across various resolutions (HP = 70–20 nm). In the absence of PEB, high e-beam doses exceeding 2000 μC cm−2 were necessary to resolve a pattern with HP = 30 nm. Fig. 4 illustrates that the line/space ratios decrease with increasing resolutions. Specifically, the L/S ratio is 0.98 for HP = 71 nm, 0.79 for HP = 55 nm, and 0.76 for HP = 42 nm. This trend indicates inadequate EUV exposure for small half-pitch patterns (HP = 29 nm), resulting in top loss of exposed photoresists. Additional e-beam patterns, featuring various e-beam doses and HP = 20–70 nm patterns, are provided in the ESI (ESI, Fig. S1).†
Fig. 4 SEM images of e-beam lithography. PAB: 80 °C, 60 s, no PEB, initial thickness 21.0 nm, developer: 2-heptanone (60 s). |
The efficiency of e-beam lithographic studies was further improved through an additional PEB process at 80 °C (60 s). As shown in Fig. 5, a dose of 1760 μC cm−2 can achieve resolution for a HP = 31 nm pattern. In the case of HP = 20 nm, the exposed photoresist 3 exhibits a curved shape in addition to broken lines influenced by the developer (2-heptanone). However, this issue was addressed with higher e-beam doses (J = 2080 μC cm−2), providing a well-defined pattern for HP = 20 nm. The majority of small patterns (HP = 20–31 nm) exhibited an idealized L/S = 1, underscoring the high quality of photoresist (3). A gradual increase in the L/S ratios from a larger HP of 52 nm (L/S = 0.88) to a smaller HP of 20 nm (L/S = 1.1) is a typical phenomenon observed in mature photoresists, as photoresist blurring is more noticeable for smaller pitches. Under this PEB condition, additional e-beam patterns, featuring various e-beam doses and HP = 20–70 nm patterns, are presented in Fig. S2 (see the ESI†).
Fig. 5 SEM images of e-beam lithographic patterns with PEB (80 °C, 60 s) initial thickness 21 nm, developer: 2-heptanone (60 s). |
The contrast curve of photoresist (3) was measured under different EUV doses; pattern development and solvent cleaning are the same as those in the e-beam study. EUV-exposure experiments were performed at the Swiss Paul Scherrer Institute Center (PSI) with a standard of error ±1 mJ cm−2, and the thickness was measured at α-stepper with an error of ±0.1 nm. PAB was conducted at 80 °C (60 s) before EUV exposure. The curve in Fig. 6 is typical of a negative-tone photoresist. Exposed film thickness starts to increase at J = 18 mJ cm−2 showing high photosensitivity. Film thickness increases gradually until a maximum height ca. 17.8 nm at J = 53 mJ cm−2. There is a break in film growth in the J = 27–36 mJ cm−2 interval, showing similar solubility of the film toward the developer. This break seems to be real because there are four EUV doses at this break. Notably, our subsequent HRXPS indicates that high EUV doses can remove carboxylate and vinyl ligands more effectively at J > 36 doses, rendering the film less soluble (vide infra, Fig. 12 and Table 1). At J = 30 mJ cm−2, HRXPS indicates a sudden cessation of the Sn–Cl bond cleavage at J = 30 mJ cm−2 even though three Sn–Cl bonds are still present. This is probably the main reason for a break in film growth in the 27–36 mJ cm−2 interval. Similar to its e-beam lithography, a small loss of 15% of thickness is observed relative to its initial height of 21.0 nm.
Fig. 6 Contrast curve of photoresist (3) using EUV doses. Initial thickness 21.0 nm, developer: 2-heptanone (60 s). |
Dose mJ cm−2 | Sn | C | O | Cl |
---|---|---|---|---|
Unexposed | 19 ± 1 | 93 ± 1 | 40 ± 1 | 5.0 ± 0.1 |
30 | 19 ± 1 | 79 ± 1 | 42 ± 1 | 2.8 ± 0.1 |
60 | 19 ± 1 | 75 ± 1 | 42 ± 1 | 2.9 ± 0.1 |
90 | 19 ± 1 | 73 ± 1 | 42 ± 1 | 2.9 ± 0.1 |
120 | 19 ± 1 | 62 ± 1 | 43 ± 1 | 3.0 ± 0.1 |
150 | 19 ± 1 | 58 ± 1 | 43 ± 1 | 2.9 ± 0.1 |
Our next task is to develop the EUV lithographic pattern of photoresist (3) using 2-heptanone as the developer (60 s). An extra PEB treatment (80 °C, 60 s) is applied to save EUV doses. The interference mask at the Swiss PSI includes various HP patterns in the 16–50 nm range. Under small doses (63–77 mJ cm−2), as shown in Fig. 7, the SEM images of all developed EUV patterns are well-defined to enable the measurement of line/space (L/S) parameters. However, few small taints were found on the small HP = 16 or 18 nm pattern, due to inappropriate operation on unexposed photoresist cleaning. This minor problem can be solved with a change to a more soluble solvent such as propylene glycol methyl ether (PGME). The respective L/S values are as follows: 1.06, 1.00, 0.95, 0.96 and 1.05 and 1.07 for HP = 50, 35, 25, 22, 18, and 16 nm, very close to an idealized ratio (L/S = 1). As shown in Fig. S3† and 8, the development of a HP = 15 nm pattern for 12-Sn cluster (2) needs EUV doses J > 200 mJ cm−2, but the energy dose is further reduced to 60–80 mJ cm−2 using the blend photoresist (3) comprising a (1)/(2) = 1:2 composition. Other EUV SEM images with different doses and half-pitch patterns are provided in the ESI (see Fig. S4).†
Fig. 7 SEM images of EUV patterns for photoresist (3). Initial thickness 22 nm. PAB (80 °C, 60 s) and PEB (80 °C, 60 s), developer 2-heptanone (60 s). |
We selected (n-BuSn)12O14(OH)6(BF4)2 (2) as the main component to blend with cluster 1 because this cluster can reach small half-pitch EUV patterns as depicted in Fig. 8 which shows the EUV-pattern of (n-BuSn)12O14(OH)6(BF4)2 (2) which shows superior performance to achieve small half-pitch patterns HP = 13–15 nm, but the corresponding EUV doses are relatively high J = 205–216 mJ cm−2. In these patterns, there are no broken lines or scums for the exposed photoresists, and the line/space (L/S) ratios are estimated to be 1.16, 1.20 and 1.20 for the HP = 15, 14 and 13 nm patterns, respectively. These L/S values are notable because the domains (HP < 15 nm) are very small for common photoresists to form a satisfactory pattern.29 Although the EUV patterning of (n-BuSn)12O14(OH)6(OH)2 (2′) has been previously reported,25 the resolution was reported to reach to HP = 30 nm with J = 68 mJ cm−2.25 It is unclear whether this cluster can enable high resolutions with HP < 16 nm.
Fig. 8 SEM images of EUV patterns for photoresist (2). Initial thickness 22 nm. PAB (80 °C, 60 s) and PEB (160 °C, 60 s), developer 2-heptanone (60 s). |
We prepared two samples for the etching resistance test: the first sample is using our blend sample (3) and the second sample is utilizing photoresist (BuSn)12O14(OH)6(OH)2 (2′). Both samples were coated on a ∼100 nm thick SiO2 layer and exposed to an e-beam (500 μC cm−2 dose) to obtain a 300 × 300 μm2 square pattern for subsequent etching tests. With exposure to a mixed gas of Cl2/CHF3/He for 5 s, the loss of height is used to measure their etching rates; their specific data are shown in Scheme 3. The etching rate of our photoresist (3) is 1.3 times higher than that of (BuSn)12O14(OH)6(OH)2 (2′). This leads to the selectivity (OX/PR) exhibiting an acceptable result (sel. = 0.63), which is not significantly different from that of photoresist (2′) (sel. = 0.85). These findings suggest that photoresist (3) shows promising potential in terms of etching resistance. All thickness data were measured using alpha-step and ellipsometer techniques.
IR spectra of photoresists (1)–(3) were studied, and the samples were prepared as KBr pellets. Fig. 9 shows the transmission mode, and there is no new peak for the blend photoresist (3), whose the spectrum is just an overlap of the spectra of two constituents (1) and (2). The absorption peaks for the sp3-hybridized υ(C–H), υ(O2CBut) and υ(Sn–O) stretching modes are labeled for photoresist (3). Notably, the υ(C–H) and υ(CC) stretching bands of the Sn–vinyl groups are too weak to observe. The blend material (3) was further baked at 80 °C for 60 s, and its KBr pellet shows an absorption spectrum that is essentially identical to the one without this PAB process. For photoresist (3), the three Sn–O bands at 500–700 cm−1 contributed mainly to 12-tin cluster (2).
Fig. 9 FTIR spectra of photoresists (1)–(3) as KBr pellets, cluster 1 (red) cluster 2 (black), blend 3 (blue) and blend 3 (after 80 °C, 60 s). |
Reflective IR spectra were conducted on the effects of EUV doses on the chemical changes in thin films. A thick film ca. 1.5 μm was prepared and the EUV-exposed spot is difficult to locate. Accordingly, the film was briefly rinsed in 2-heptanone for 10 s to locate the exposed spots. We do not exclude the possibility that this exposed spot has been detached from the substrate because the bottom part is still unexposed and removed with 2-heptanone. As shown in Fig. 10, a dose of 35 mJ cm−2 led to a sharp decrease in the absorption intensity of the υ(OH), υ(C–H) and υ(O2CBut) bands centered at 3201, 2920–2980, and 1423 cm−1 whereas inorganic bands at 500–900 cm−1 exhibit strong IR intensity. This process likely involves a EUV-activated decarboxylation, a dehydration reaction or a Sn–butyl bond cleavage. The relative rates of the carbon loss will be further studied by HRXPS studies. At this J = 35 mJ cm−2 dose, a small decrease in the υ(Sn–O) intensity at 680 nm−1 is noted, whereas no new υ(Sn–O) band is observed in this region. The framework of 12-tin cluster (2) in this blend material (3) seems to survive. However, at J = 70 mJ cm−2, most organic residues have been removed by EUV light and only two inorganic bands appear on the film. The band at 1100–900 cm−1 is likely due to the absorption of υ(Sn–O–Si) whereas a new band at 769 cm−1 is assignable to the υ(Sn–O–Sn) absorption of SnO2-like species.30–34 At this high dose, a collapse of the cluster structures of components (1) and (2) is sure to occur because their common band in the 680 nm−1 region becomes very weak in absorption intensity; formation of a new broad band at 769 cm−1, due to SnO-like species, is observed. These FTIR studies are not able to show the reactivity of Sn–vinyl and Sn–Cl bonds under a EUV dose of J = 30 mJ cm−2, due to their invisible absorption.
Fig. 10 FTIR absorption of photoresist (3) at different EUV doses. From bottom to top: 0, 35, 70 and 100 mJ cm−2. The film thickness is measured to be 1.5 μm. |
For photoresist 1, the effect of PAB on surface composition was examined using HRXPS (high resolution X-ray photoelectron spectroscopy) as a tool. A thin film of cluster (1) was prepared in ca. 25.0 nm thickness through spin-casting on a silicon wafer; the film was air dried at room temperature for 48 h. Another film was subjected to a PAB treatment (80 °C, 120 s). For these samples, five Sn atoms are used as the internal standard whereas the composition changes of other elements are calculated based on their intensity ratios with respect to the Sn absorption. As shown in Fig. 11, no composition change was observed among Sn, C, O and Cl elements. The Sn–Cl bonds of photoresist 1 seem to be robust toward the PAB process. This assessment is valid only on the surface because these PAB-treatment film still remains at 5–6 nm thickness after cleaning with the developer 2-heptanone (60 s). The Sn, C, O and Cl composition of cluster (1) for the TGA residue was also examined by HRXPS. The sample was treated by the same TGA operation using stepwise heating from 30 °C to 450 °C in air. The residues were pressed into a neat solid film ca. 10–12 μm thickness. If Sn content is set at 5 atoms, the loss of carbon and Cl is up to 87% and 90%, respectively. Our TGA study in Fig. 1 also suggests a loss of volatile Sn compounds because the residues are too low in the remaining weight. We also conducted the carbon analysis of the C (1s) carbon component on the TGA residues at 450 °C (see Fig. S5†), and a further band shape fitting indicates that the dominant carbon component is attributed to the sp3-hybridized t-butyl group.
Fig. 11 HRXPS of photoresist (1) without PAB (black, bottom), andwith PAB (80 °C, 120 s, red and middle) and the TGA residue after heating at 450 °C (blue line and top). |
We investigated the effect of the EUV doses on the element compositions of the film surface using HRXPS as a probe. The film is subjected to a PAB treatment before EUV exposure. This PAB-treated sample is expected to have the same element composition as its fresh thin film because no chemical reactions are shown by IR and XPS study. After EUV exposure at different doses, these samples are exposed to air before storage in a nitrogen-filled container. Fig. 12 shows the HRXPS spectra of the thin films at J = 30–200 mJ cm−2, which show well-defined absorptions of tin (Sn), carbon (C), oxygen (O) and chlorine (Cl) in the unexposed film. With increasing EUV doses, the absorption intensities of C (1s), O (1s) and Cl (2s) peaks become weak referring to the Sn absorption intensity whereas the boron (B) and fluoride (F) peaks are too weak to see the change. Silicon peaks start to appear at high energy doses (J > 90 mJ cm−1−2), due to the formation of inorganic Si–O–Sn species; the corresponding IR absorption is also detectable (950–1050 nm−1) in Fig. 9. With a weight ratio of (1)/(2) = 1:2, the molar ratio of (1)/(2) is 1:1.2. With their respective formulae, the element ratios of Sn:C:O:Cl are calculated to be 19:93:40:5.0; herein, chloride atoms are fixed to be five. The initial 19 Sn atoms are used as a standard at different EUV doses. Table 1 summarizes the element compositions at different doses. At J = 30 mJ cm−2, the carbon content decreased due to the loss of 14 atoms and the chlorine content is lost by 2.2 atoms. In contrast, we observed an increase of the oxygen content with 2 atoms. This process indicates photolytic decompositions of Sn–carbon, Sn–chloride and Sn–carboxylate ligands, forming new Sn–OH or Sn–O–Sn species. With increasing doses from J = 40 mJ cm−2 to 90 mJ cm−2, the carbon loss becomes slow and the oxygen increase is also slow. But there is a rapid loss of carbon atoms with high EUV doses, J = 90–150 mJ cm−2 whereas the oxygen increase is also relatively large. Our subsequent O (1s) component analysis with HRXPS studies indicate that decomposition of Sn–carboxylate ligands is more rapid at high EUV doses (vide infra, Table 3). Notably, the chlorine contents are kept at 2.9–3.0 atoms at J = 30–90 mJ cm−2, and no further decrease for higher EUV doses; this observation well matches two-stage film growth in the EUV contrast curve (see Fig. 1). There are three different processes corresponding to the loss of carbon atoms, including the EUV-activated decomposition of Sn–butyl, Sn–vinyl and Sn–carboxylate groups. Further analysis of their relative contents relies on HRXPS analysis of the carbon C (1s) component.
Fig. 12 HRXPS of photoresist (3) under different EUV doses (J = 0, 30, 60, 90, 120 and 150 mJ cm−2). |
HRXPS band shape fittings were conducted on the C (1s) and O (1s) absorption peaks to characterize the process of carbon and oxygen change. Band shape fittings are shown in Fig. 13 that also includes their quantitative analysis. Our photoresist 3 contains sp2-and sp3-hybridized C–H carbons in addition to the carboxylate C–O carbons. Therefore, only these three carbon components are added together to form the observed C (1s) absorptions at different doses using the software in HRXPS study. In the O (1s) analysis, only two Sn–O and RCO2–Sn components are considered as the basis sets. Three components were found for the observed C (1s) peaks, including the sp3 and sp2 C–C carbons and CO2 components centered at 284.0, 284.7 and 287.5 eV, respectively. The sizes of the area in each component are used to calculate their percentages in Table 2. The data in the bracket are the expected values from a blend composition with [1]/[2] = 1.1:1.9. The discrepancy is not surprising because the X-ray absorption coefficients are not the same for such different C–H bond modes. Notably, the relative ratios are significantly changed at J = 30 mJ cm−2, in which the sp3-carbon contents are greatly decreased whereas the sp2-carbon and carboxylate CO2 carbon components have increasing contents. Herein, the main component of sp3-carbon decomposition arises from Sn–n-Bu, instead of Sn–t-Bu. The Sn–n-Bu groups are cleaved in greater amounts than those for the SnO2C and Sn–CHCH2 carbons. This trend remains unchanged for increasing doses of J = 60–90 mJ cm−2. This trend is rational because there are 12 n-BuSn groups on the tin cluster cage (2) that is known to be very photosensitive due to the high Sn density. On the other hand, radical propagation on the vinyltin moiety is not evident because we do not see a loss of the sp2-carbon contents together with an increase in the sp3-carbon content. In Table 1, a quick loss of carbon atoms is noted after J = 90–150 mJ cm−2. In such a high dose stage, the Sn–O2C content is increased gradually whereas Sn–CHCH2 content remains approximately the same. Therefore, the SnCHCH2 bonds are cleaved in greater proportion than that for the Sn–O2C moieties amid these high EUV doses. Among these carbon components, EUV light decomposes the Sn–O2CBut and Sn–vinyl bonds significantly only in high doses whereas the Sn–Cl and Sn–n-Bu bond cleavage can readily occur with low EUV doses.
Fig. 13 Band shape fitting of the C (1s) components of HRXPS for photoresist (3) under different EUV doses (J = 0, 30, 60, 90, 120, 150 cm−2). |
Dose (mJ cm−2) | COO | CHCH2 | n-Bu/t-Bu |
---|---|---|---|
a The number in the bracket is the expected value calculated from a weight ratio of [1]/[2] = 1.1:1.9. | |||
unexposed | 4.8 ± 0.1 (6.5 ± 0.1)a | 14 ± 1 (7.0 ± 0.1) | 81 ± 1 (87 ± 1) |
30 | 5.6 ± 0.1 | 22.±1 | 72 ± 1 |
60 | 5.8 ± 0.1 | 24 ± 1 | 71 ± 1 |
90 | 5.9 ± 0.1 | 23 ± 1 | 71 ± 1 |
120 | 6.2 ± 0.1 | 23 ± 1 | 70 ± 1 |
150 | 6.2 ± 0.1 | 23 ± 1 | 71 ± 1 |
In the O (1s) peak analysis, the component centered at 530.6 eV in Fig. 14 is assignable to inorganic Sn–O species including Sn2OH, Sn2O and Sn3O;16,30–32 the second component centered at 529 eV is assigned to the t-BuCO2Sn oxygen. Previously, Zhang reported32 that the HRXPS spectra of the [(n-BuSn)12O14(OH)6]+2 core could be resolved into two components such as Sn3O and Sn2OH in this 530.0–531.2 eV region. But the EUV light J > 35 mJ cm−2 in our IR study (Fig. 10) shows the collapse of this 12-tin cage (2), possibly forming new species such as Sn2O, Sn4O and Sn6O as in the SnO2 solid phase.33,34 The use of one single Sn–O component at 530.6 eV is convenient for the band fitting as other Sn–O bands of cluster (1) are also considered. In this study, we prepared the blend (3) containing a weight ratio of (1)/(2) = 1.1:1.9, and the ratio of the two components, Sn–O: t-BuCO2, is calculated to be 65:35; the experimental value is 58:42 as shown in Table 3. A PAB process can affect this ratio, but to a small extent (see Fig. 11); this discrepancy is probably due to the different X-ray absorption coefficients of two O (1s) components. As shown in Fig. 14, the increase in the Sn–O content is contributed from the Sn–X bond cleavage (X= Cl and C), further forming the Sn–O bonds of various types in the EUV-exposure/air exposure sequence. However, this oxygen component analysis can detect the degree of carboxylate decompositions at different doses. In the range of J = 30–60 mJ cm−2, as shown in Table 3, this photolytic process takes place at a slow and steady rate because a large proportion of t-BuCO2Sn comparison with different dose 30–60 mJ condition have great proportions of O atom than 90–150 mJ. The slow rate of such a t-BuCO2Sn decomposition is also inferred from the C (1s) analysis in Table 2 and Table 3. At J = 30–60 mJ. cm−2, the increase in the Sn–O content is also relatively small, ca. 61%. But with high doses J = 90–150 mJ cm−2, the SnO content is increased rapidly, from. 61% to 68% whereas the t-BuCO2Sn content decreases relatively quickly, from 39% to 33%. Accordingly, high EUV doses are required to decompose the Sn–O2CBut bonds.
Fig. 14 Band shape fitting of the two O (1s) bands of HRXPS of photoresist 3 under different EUV doses. J = 0. 30, 60, 90, 120, 150 and 180 mJ cm−2. |
The main reactions at different EUV doses are summarized in Scheme 4. With IR and HRXPS studies, we postulate that the photolytic reaction at the low EUV doses at J = 30 mJ cm−2, involves a quick removal of sp3-hybridized carbons (Table 2), mainly from the n-butyl groups of 12-Sn clusters. The mechanism of this process has been elucidated by Brouwer and coworkers,35–37 involving a direct Sn–butyl bond cleavage, generating Sn(II)-species for a subsequent aggregation. The Sn–Cl, Sn–vinyl and Sn–O2CBut cleavages follow the reported mechanisms of tin carboxylate clusters.14–16 Although the photolytic decomposition of Sn–O2CBut and vinyltin groups also takes place for the blend (3), their rates are slow because a large proportion still remains at J = 60 mJ cm−2 as shown in Table 3. For the vinyltin groups, radical propagation at this functionality is not supportive. A summary of the three functionalities in EUV photochemistry is provided below. We postulate that at low doses of 30–60 mJ cm−2, HRXPS analysis on the C (1s) component in Table 2 indicates a quick cleavage of Sn–butyl and Sn–Cl bonds. At J = 60 mJ cm−2, 2–3 Sn–butyl bonds and two Sn–Cl bonds have been cleaved by EUV light. Photolytic decomposition of Sn–O2CBut and vinyltin groups becomes significant only at high EUV doses (J = 90–150 mJ cm−2).
Photolytic decarboxylation of Sn(O2CBut) and Sn–vinyl moieties is relatively slow as compared to the Sn–Cl and Sn–butyl cleavage according to the C (1s) analysis in Table 2. A fast cleavage of Sn–butyl and Sn–Cl bonds is rational because the sp3-hydridized Sn–C and Sn–Cl bonds are generally weak. No further Sn–Cl bond cleavage is noted at high doses J = 60–150 mJ cm−2. One possible reason is to form a Sn–Cl–Sn bridge that will be very difficult to dissociate because of two coordinated Sn centers. Unfortunately, a band shape fitting of Cl (2s) or Cl (2p) is not satisfactory because the peaks are very small.
Our HRXPS studies can explain the two-stage rise in the EUV contrast curve in Fig. 6. A break in the J = 27–36 mJ cm−2 region is associated with a sudden cessation of the Sn–Cl bond cleavage so that only Sn–butyl bond cleavage is occurring at these doses. The main reactions in different E|UV doses are shown in Scheme 3.
FTIR spectra also provide mechanistic insight; one novel finding is the collapse of the cluster structures of 12-tin cage (2), and 5-tin-5-Cl carboxylate species (1) is also observed starting at J = 70 mJ cm−2, forming SnO2-like species. Furthermore, increasing EUV doses significantly decreases the IR absorption intensity that is accompanied by increasing intensity of Sn–O bonds. Only a small proportion of organic moieties remains at a high dose at J = 150 mJ cm−2.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4na00006d |
This journal is © The Royal Society of Chemistry 2024 |