Sputter rates

Finally, it is difficult to caUbrate the depth scale in a depth profile. This situation is made more compHcated by different sputtering rates of materials. Despite these shortcomings, depth profiling by simultaneous ion sputtering/aes is commonly employed, because it is one of the few techniques that can provide information about buried interfaces, albeit in a destmctive manner. 

Every material sputters at a characteristic rate, which can lead to significant amb ity in the presentation of depth profile measurements by sputtering. Before an accurate profile can be provided, the relative sputtering rates of the components of a material must be independently known and included, even though the total depth of the profile is normally determined (e.g., by stylus profilometer). To first order, SNMS offers a solution to this amb ity, since a measure of the total number of atoms being sputtered from the surface is provided by summing all RSF- and... 

Table 3 Results of GOMS analyses for impurities in three high-purity semiconductor substrates. Lower detection limits are achieved for materiais with higher sputtering rates.

In both electron post-ionization techniques mass analysis is performed by means of a quadrupole mass analyzer (Sect. 3.1.2.2), and pulse counting by means of a dynode multiplier. In contrast with a magnetic sector field, a quadrupole enables swift switching between mass settings, thus enabling continuous data acquisition for many elements even at high sputter rates within thin layers. 

Eig. 3.54). ERISS enables static LEIS measurements (without noticeable damage), even at polymer surfaces (see below) for which the sputter rate even under LEIS conditions is very high. 

The quantification algorithm most commonly used in dc GD-OES depth profiling is based on the concept of emission yield [4.184], Ri] , according to the observation that the emitted light per sputtered mass unit (i. e. emission yield) is an almost matrix-independent constant for each element, if the source is operated under constant excitation conditions. In this approach the observed line intensity, /ijt, is described by the concentration, Ci, of element, i, in the sample, j, and by the sputtering rate g, ... 

The emission yield, Ra, defined as the radiation of the spectral line, k, of an element, i, emitted per unit sputtered mass must be determined independently for each spectral line. The quantities g, and Ry are derived from a variety of different standard bulk samples with different sputtering rates. In practice, both sputtering rates and excitation probability are influenced by the working conditions of the discharge. Systematic variation of the discharge voltage, L/g, and current, I, leads to the empirical intensity expression [4.185] ... 

In order to carry out depth profiling with AES, the sputtering rate must be determined. The sputtering rate is usually measured by determining the time required to sputter through a layer of known thickness. Anodized tantalum foils are convenient for this purpose since the oxide thickness can easily be controlled and since the interface between the metal and the oxide is relatively sharp [43]. 

If it is required that the surface of the sample remains undisturbed during analysis, SIMS must be carried out at very low surface removal rates, typically about 10 monolayer/s. The terms static and dynamic are used to divide the sputtering rate of the sample into regimes where only surface species are observed (static SIMS) or where surface and bulk species are observed (dynamic SIMS). The static limit is usually considered to be <10 ions/cm impinging on the sample surface. Under these conditions, only about 1/1000 atoms on the surface of the sample are struck by a primary ion. 

GDS instruments are viable alternatives to the traditional arc and spark-source spectroscopies for bulk metals analysis. Advantages of GDS over surface analysis methods such as AES, XPS and SIMS are that an ultrahigh vacuum is not needed and the sputtering rate is relatively high. In surface analysis, GD-OES, AES, XPS and SIMS will remain complementary techniques. GD-OES analysis is faster than AES (typically 10 s vs. 15 min). GD-OES is also 100 times more sensitive than... 

As a first approximation, a simple linear relation may be assumed to convert the time of sputtering to depth. Thus for a constant sputtering rate, the depth of erosion z will be given by ... 

They employed a FIB of 30 keV Ga ions which as focused to a spot with a diameter which could be varied between 0.05 and 1 pm. The beam current varied with focus size between 13 pA and 1.2 nA, and sputter rate typically increased with beam current from 0.005 to 0.5pm3s 1. The beam control was automated, so that the major part of the specimen preparation was performed automatically, only the final high-resolution operations being carried out by manual adjustment of the milling area. Their preparation scheme was as follows ... 

In order to get an idea about the depth resolution of the system and about sputtering rates, electrochemically prepared Ta2Os serves as a very suitable example. The thickness of the oxide layer can be easily varied by choosing a proper oxidation potential (16A/V). Tantalum oxide is not reduced during sputtering according to [26],... 

The time necessary for removing one monolayer during a SIMS experiment depends not only on the sputter yield, but also on the type of sample under study. We will make an estimate for two extremes. First, the surface of a metal contains about 1015 atoms/cm2. If we use an ion beam with a current density of 1 nA/cm2, then we need some 150 000 s - about 40 h - to remove one monolayer if the sputter yield is 1, and 4 h if the sputter rate is 10. However, if we are working with polymers we need significantly lower ion doses to remove a monolayer. It is believed [4] that one impact of a primary ion affects an area of about 10 nm2, which is equivalent to a circle of about 3.5 nm diameter. Hence if the sample consists, for example, of a monolayer film of polymer material, a dose of 10n ions/cm2 could in principle be sufficient to remove or alter all material on the surface. With a current density of 1 nA this takes about 1500 s or 25 min only. For adsorbates such as CO adsorbed on a metal surface, we estimate that the monolayer lifetime is at least a factor of 10 higher than that for polymer samples. Thus for static SIMS, one needs primary ion current densities on the order of 1 nA/cm2 or less, and one should be able to collect all spectra of one sample within a quarter of an hour. 

Transient Regime. Equation 7 is separable and may be integrated to obtain the time dependence of the oxide thickness. Assuming a finite sputtering rate, the result is... 

---