Specific Process Knowledge/Thin film deposition/Deposition of Gold/Adhesion layers: Difference between revisions

← Older edit Newer edit →

VisualWikitext

@@ Line 47: / Line 47: @@
 == Bilayer chemical composition and elemental distribution ==
+The knowledge of the chemical composition and elemental distribution of the
+bilayer systems is important to understand the type of interaction between
+the layers and their di�erent structure. To obtain such information, the
+bilayer systems were investigated using STEM-EDX, STEM-EELS and XPS
+depth pro�ling.
+For the chemical distribution of the elements in the 2-Ti/2-Au sample,
+the STEM-EDX measurements at the Ti/Au interface showed the presence
+of a continuous Ti layer below the Au layer (Fig. 5.9a). This result is in
+good agreement with the TEM micrograph of Fig. 5.5a.
+The analysis of the 2-Cr/2-Au sample showed instead the presence of
+Cr throughout the whole thickness of the Au layer (Fig. 5.9b). Since Cr
+and Au were deposited sequentially and not by a co-deposition process, this
+result suggests a strong inter-di�usion. The 2-Cr/20-Au sample was also
+investigated, to understand the degree of inter-di�usion for a thicker Au
+layer (Fig. 5.9c). In this case the inter-di�usion between the elements was
+incomplete, being present for a thickness of 2-3 nm. This suggests that the
+di�usion process is limited to such a thickness when Cr and Au are deposited
+at room temperature.
+To verify the chemical composition of the samples, in particular to investigate
+a possible presence of oxygen in the adhesion layer, STEM-EELS
+analysis was used. A line scan across the layer interfaces of the 2-Ti/2-Au
+sample shows the presence of a Ti core loss L3 edge at 460 eV and an L2
+edge at 465 eV. A SiO2 O-K edge is visible at 538 eV, while the O-K edge
+of O bounded to Ti is found at 532 eV (Fig. 5.10a).
+The same investigation was performed on the 2-Cr/2-Au sample (Fig.
+.10b). The analysis showed a Cr L3 edge at 585 eV and an L2 edge at
+eV. The SiO2 O-K edge is visible at 545 eV, while at 540 eV a weak
+OCr-K edge of O bounded to Cr is visible for a limited thickness below
+Au. Furthermore, the Cr edge presents a compositional tail along the scan
+direction, which con�rms di�usion into the Au layer. For the length of the
+tail there is no presence of OCr-K edge, indicating that Cr inside Au is in
+metallic form. This is in good agreement with the observed di�usion, which
+involves only metallic Cr.
+The source of oxidation of the adhesion layer could have three origins:
+i) oxidation due to migration of oxygen from the SiO2 substrate, ii) oxidation
+during the e-beam deposition process, or iii) oxidation due to substrate
+contamination with water and oxygen molecules. To investigate the oxidation
+origin, XPS depth pro�le analysis were performed on 2-Ti/20-Au and
+-Cr/20-Au as-deposited samples.
+The �rst analysis was performed to clarify if oxygen was originating from
+source i). To rule out possible oxygen migration from the oxygen-rich SiO2
+substrate, the metal thin-�lms were deposited on amorphous Si3N4. For
+both samples, after Ar ion milling of 20 nm of Au, the Au 4f signal intensity
+decreased and Cr 2p and Ti 2p signals started to appear together with the
+O 1s signal (Fig. 5.11a for Ti and 5.11b for Cr, respectively). For the Cr
+case, due to the lower spatial resolution of XPS with respect to STEM-EDX,
+it was not possible to verify the thickness of Cr-Au inter-di�usion in this
+case. The XPS electron escape depth at 1486.7 eV for Au is around 0.15
+nm [114], meaning that it is not possible to obtain information about interdi
+�usion for layers thinner than this value without collecting a signal from
+the underlying material.
+Peak �ts for the Ti and Cr signals were performed, and are reported in
+Fig. 5.12. The Ti 2p peak is a convolution of three components: a TiN
+doublet at 455.0 and 461.1 eV, a TiO2 doublet at 457.7 and 464.0 eV and
+a TiOx signal, which forms the descending background tail [115]. The �t
+gives the following information: 1) Ti has formed Ti-N bonds with the Si3N4
+substrate; 2) the adhesion layer was partially oxidized during deposition
+and not from source i). Metallic Ti could not be detected, but its presence
+cannot be excluded: metallic Ti is highly reactive with respect to oxygen
+and nitrogen, and the destructive sputtering process used for the depth
+pro�ling could have enhanced the mixing between Ti, O and N, catalyzing
+the reaction of the metallic Ti bound to Au to form an oxide or a nitride.
+The Cr 2p peak �t is formed by three components: a metallic Cr doublet
+at 574.4 and 583.6 eV, a Cr2O3 doublet at 576.3 and 585.6 eV and a CrO3
+doublet at 580 and 589.2 eV [115]. The result of the �t indicates a partial
+oxidation of Cr, as in the Ti case.
+The second analysis was performed to understand whether the oxygen
+was originating from source ii) or iii). A 2-Ti/20-Au/2-Ti/20-Au sandwich
+structure was deposited and analyzed. Both layers of Ti were partially oxidized:
+the Ti 2p peak signals (Fig. 5.13a) are present at the same depth
+together with the O 1s signals (Fig. 5.13b). The O 1s signal in the Ti layer
+in contact with the Si3N4 substrate has higher intensity than the one of the
+Ti layer between the Au layers. Hence, the Ti layer in contact with the substrate
+is more oxidized, which suggests that Ti reacted with water adsorbed
+on the substrate surface. The conclusion is that the oxygen originated from
+sources ii) and iii).
+Water is present in most thin-�lm deposition systems and it is very
+di�cult to remove because of its capacity to form hydrogen bonds with
+many materials. Ti reacts with water and oxygen and forms titanium oxides.
+Furthermore, the use of the same deposition conditions and of the same type
+of substrate with respect to Ti and a similar chemical reactivity of Ti and
+Cr suggest that also the Cr layer gets further oxidized by the water and
+oxygen molecules present on the substrate surface.
 == Adhesion layer effect on bilayer thin-film electrical resistivity ==