Nuclear Instruments and Methods in Physics Research B7/8 (1985) 219-227 North-Holland,
SURFACE MICROANALYTICAL STUDIES OF NITROGEN ION-IMPLANTED STEEL * Charles G. DODD Connecticut
Scott M. BAUMANN
Charles Evans & Associates,
Keith 0. LEGG Ion Technology, Inc.,
CT 06497, USA
and James C. NORBERG
San Mateo, CA 94402, USA
GA 30357, USA
Five types of industrial steels, 1018, 52100, M-2, 44OC, and 304 were ion implanted with nitrogen and subjected to surface microanalysis by three independent surface techniques: AES, RBS, and SIMS. The results provided understanding for earlier observations of the properties of various types of steel after nitrogen implantation. The steels that retained the most nitrogen and that have been reported to benefit the most in improved tribological properties from ion implantation were ferritic carbon and austenitic stainless steels, such as soft 1018 and 304, respectively. Heat-treated martensitic carbon steels such as 52100 and M-2 tool steel were found to retain the least nitrogen, and they have been reported to benefit less from nitrogen implantation; however, the interaction of transition metal carbides in M-2 with nitrogen has not been clarified. The data showed that 440C steel retained as much nitrogen as 1018 and 304, but treatment benefits may be limited to improvements in properties related to toughness and impact resistance.
1. Introduction Beneficial effects of nitrogen implantation on ttibological properties of steel were first reported in 1973 [l]. Representative subsequent work includes that of investigators at Harwell [2,3], the Naval Research Laboratory [4,5], and Sandia [6,7]. This paper attempts to interpret diverse results of nitrogen implantation studies by characterizing the near-surface implanted regions of five important commercial steel types, using three state-ofthe-art surface analytical techniques. The analytical techniques employed were Auger spectrometry (AES), Rutherford backscattering spectrometry (RBS), and secondary ion mass spectrometry (SIMS). The samples were characterized by elemental concentration depth profiles and measurement of the mean chemical compositions in successive surface microlayers. These kinds of data shed light on the environment and bonding of subsurface nitrogen and on the ability of different types of steel to receive and retain nitrogen. One objective was to acquire the ability to prescribe an optimum surface treatment for a given steel in a specific application. A perplexing problem for practical workers in the field has been why a treatment that gave good results when applied to one type of steel was not effective with * Work supported by NSF-SBIR Grant No. DMR-8360256. 0168-583X/85/$03.30 0 Elsevier Science Publishers (North-Holland Physics Publishing Division)
another steel. Nitrogen implantation was found to be beneficial for surface modification treatments of some carbon steels and for austenitic stainless. On the other hand, nitrogen has not been found to be generally useful for all tool steels nor for some hardened martensitic steels. Some of the answers to these questions are more obvious than others. One might logically assume it to be unlikely that a 52100 martensitic through-hardened bearing steel could be made harder and more wear-resistant by nitrogen implantation, and this has been found to be the case [6,8]. However, the hardness of a steel has been found not always to be an indicator of its wear resistance properties [6,9]. The latter may be aided by stabilizing steel microstructures against deleterious work-hardening effects, by introducing residual stresses to combat fracture, and by hardening surfaces of soft steels, such as soft 1018, to permit subsurface deformation while resisting surface fracture [6,10]. Improvement of the surface wear properties of austenitic stainless may be accompanied by a reduction in surface hardness as a consequence of re-austenitization of martensite generated by previous surface work hardening. Studies of efforts to improve fatique life in various steels by nitrogen implantation have not been conclusive [ll]. The effects of nitrogen implantation on different tool steels containing large amounts of carbides are not clear. The martensitic matrix of these steels may be softened by partial transformation to austenite, but the IV. METALS Applications, Steels
C. G. Dodd et al. / Surface microanalytical studies of N-implanted
wear-resistant character of the contained transition metal carbides may be improved. Overall, the phenomena are complex. There is a need for more definitive and comprehensive surface characterization of these steels.
2.2. Ion implantation of steel specimens Ion implantation was done with the Georgia Inst. of Technology Acelerators Inc. implanter using a molecular nitrogen ion beam at 190 kV, equivalent to 95 keV atomic nitrogen, for which the projected range was calculated to be from 100 to 110 nm. Half of each specimen surface was masked to provide a control, unimplanted surface. The implantation dose was 4 X 10” ions/cm2, the specimen temperature during implantation 175”C, the target chamber pressure 6 X 10e6 Torr initially and 2 X low6 Torr during implantation, and the specimen ion beam current 3 pA/cm2.
2. Experiments 2.1. Preparation of steel specimens Specimens of AISI-SAE No. 1018 (UNS No. G10180) carbon steel were cut from soft rod stock. AISI-SAE No. E 52100 (UNS No. G52986) electric fbmace through-hardened martensitic steel specimens were cut from bearings. AISI Type M-2 high speed (molybdenum) tool steel specimens were cut from hard’ened bar stock. Specimens of AISI Type 440C (UNS No. S44004) martensitic stainless steel were cut from rods that had been hardened, quenched and tempered. AISI Type 304 (UNS No. S30400) austenitic stainless steel was cut directly from bar stock. The specimens were in the form of (approx.) 1 cm squares or circles, each about 2 mm thick. One surface of each specimen was prepared for ion implantation by grinding with silicon carbide paper and polishing with successively finer diamond powder down to 0.5 pm. The nominal chemical compositions of the five steels are shown in table 1.
2.3. Surface analyses of steel specimens Auger spectra and elemental depth profiles were obtained, using a JEOL JAMP-10 scanning Auger microprobe, by sputtering with argon ions at a mean, estimated sputtering rate of 5 nm/min. In general, Auger depth profiles provided a more intuitive and subjectively revealing picture of the subsurface region than RBS or SIMS. AES depth profiles for the five steels are shown ti fig. l-5. RBS measurements were made using a General Ionex Tandetron/MeV spectrometer employing a primary beam of 2.25 MeV He’+ ions. Experimental data were
.. .. ‘... . . . . . . .... .. . ..._.._......‘..
.-_-_-~---‘---.-~: _:~ 0
30 TIME (MIN. 1
Fig. 1. Auger depth profiles for soft 1018 steel. Note gaussian peak for nitrogen at point b; near-surface out-diffusion nitrogen peak at point a; carbon contamination layer on surface; oxides at steel interface; and dip in iron profile coincident with nitrogen peak at b.
C. G. Dodd et al. / Surjace microayzlytical
studies of N-implanted
Table 1 Approximate chemical compositions of steels (wt.%, bal. Fe) Steel
1018 52100 M-2 440C 304
0.17 1.0 0.93
. . ,,....
1.4 4.0 17 19
Estimated probable errors for compositions of surface carbon layers are L-50%, while those for subsurface layers are k 3%. Since one can measure only the thickness-density product from an RBS spectrum, the calculated thicknesses depended on the densities assumed to calculate them; hence the probable errors for layer thicknesses are k 25%. SIMS surface analyses were done using a Cameca IMS-3f Ion Microanalyzer with a Cs ion source for negative ion spectrometry. The depths of ion beamsputtered craters generated during an analysis were measured with a Sloan Instruments Dektak surface profilometer and the data used to convert ion beam sputtering rates to depths. The data were processed to generate plots of ion intensity versus depth, as displayed in fig. 7.
gathered at both near-normal and grazing detector angles (160°C). Computed (“theoretical”) spectra, based on assumed surface microlayer thicknesses and compositions, were then fitted to each experimental grazing angle spectrum using a data processing program. Fig. 6 shows a typical RBS plot, showing scattering yield vs scattered particle energy, or channel number. The theoretical curve is the dashed line, which is iteratively fitted to the experimental spectrum, from the iron edge down to lower energies. Computed spectra could be divided into as many as nine layers. The first layer corresponded to the outer surface layer, with successive layers at greater depths in the sample. Results of these computations are summarized in table 3, and thickness-composition products are shown in table 4.
Fig. 2. Auger depth profiles for through-hardened 52100 steel. Note skewed nitrogen profile inverse increase in iron profile with depth; oxides at steel interface; and carbon contamination
with maximum near steel surface layer on surface.
IV. METALS Applications, Steels
,....,,..............“..........’ . . . . . . . . . ..~~..~........‘...
12000 z ;; ioooo 6 I;: I+ 8000 6000
:::L!J~~ 09 ..' 0
Fig. 3. Auger depth profiles interface:
for hardened surface
M-2 tool steel. Note broad
peak and decreasing
depth; oxides atSteel
. . . . . . . . .._._. ,_......’
Fig. 4. Auger depth profiles for hardened and tempered 4tOC martensitic stainless steel. Note deep, semi-gaussian nitrogen peak with weak peak below steel surface; oxides at steel interface; the surface carbon contamination layer with a positive indication of the presence of silicon (present in other four steel specimens, also); bulk chromium content (the Auger energy window used for oxygen, 488-529 eV, included the 510 eV oxygen peak and the 529 eV chromium peak).
C:G. Dodd ef al. / Surface microanalyfical studies of N-implanted
. . . . .._...l.......__~.............~...
20 30 40 SPUTTERING TIME tMIN. 1
Fig. 5. Auger depth pro&es for 304 austenitic stainless steel. Note the broad gaussian nitrogen chromium intensity and flat iron profile, suggesting nitrogen present as chromium nitrides; layer on surface; and insignificant intensity of oxygen at steel interface.
peak with a coincident slightly thinner carbon
inverse drop in contamination
12000 Q 0 ._ %
Fig. 6. Representative RBS grazing incidence spectrum for soft 1018 steel. Note the computer-generated “theoreticat” curve dashed line that represents the best fit for iterative microsurface layer thickness-composition assumptions; also note iron edge corresponding to steel surface and attenuated backscattered ion intensity (from ‘*Fe surface” to “Fe substrate”) corresponding to thickness of implanted nitrogen: and note carbon edge. IV. METALS Applications, Steels
C.G. Dodd et al. / Surface microanalytical studies of N-implanted steel
. . . . . . . . . . . . . . 521oc m-2
to! h : * . Cn C
: 0 Y >, .‘: u) c 0)
r ._ z ._
Fig. 7. SIMS computer-generated negative FeN ion depth profile plots for five steel samples. Note confirmations of features seen in AES plots: near-surface nitrogen out-diffusion peak for 1018 steel; near-Gaussian nitrogen peaks for 1018, 304 and 44OC steels; and broad, skewed nitrogen peaks for 52100 and M-2 steels. Depth scale from sputtered crater.
reference to table 2 in which are listed the most apparent characteristics of each profile. Additional features of the Auger depth profile plots were the following: (1) Some iron depth profile curves displayed a dip that coincided in position with that of the implantednitrogen depth profile peak.
3.1. Elemental depth profiles by Auger spectrometry Figs. 1, 2, 3, 4, and 5 are AES depth profiles for 1018, 52100, M-2, 44OC, and 304 steel specimens, respectively. Comparison of the data may be facilitated by
Table 2 Nitrogen
depth profile peak characteristics
a) Surface peak
Main profile peak Gaussian
1018 52100 M-2 44x 304
and layer compositions
Yes No No Yes b, Yes
110 Yes Yes
55 55 140 130
80-180 45-140 25-170 60-190 45-200
Yes (?) b, (?) b’ (?) b’ No
45 55 30-80 55 b’
‘) Depths and breadths were determined using RBS data and are reported in nm. b, On the 44OC profile (and 52100 and M-2 profiles) there is a suggestion of a near-surface
at 55 nm.
C.G. Dodd et al. / Surjace microanalytical studies of N-implanted steel (2) A slight rise in the oxygen profile at the steel surface of all specimens indicated the presence of surface oxide. (3) Peaks labelled “a, b, and c” in fig. 1 represent times during measurement of the nitrogen depth profile of 1018 steel when argon sputtering was stopped and a nitrogen Auger spectrum was measured at high resolution. This was done in an attempt to see if any evidence could be obtained for differences in the bonding of nitrogen at various depths. No differences could be detected in AES nitrogen peaks for any of the steel specimens. (4) Each of the nitrogen implanted steels had a significant surface layer of carbon contamination as well as a small amount of silicon. The carbon layer on the non-implanted, control portions of the samples was much thinner than that on the implanted portions. The silicon is believed to be a consequence of the use of DC 704 silicone pump oil in the implanter diffusion pump. Silicon is only shown in the depth profile for the 440C steel sample for clarity. 3.2. RBS spectra and surface layer composition-thickness computations A “theoretical” RBS plot at grazing incidence for 1018 steel is shown in fig. 6. The rest of the RBS data are presented in tables 3 and 4. Because the nitrogen scattering cross section was much smaller than those of the heavier alloying metals in M-2, 44OC, and 304, it was not possible to measure directly the dose of implanted nitrogen in these steels. Therefore, the concentrations were determined by the reduction in the concentrations of metals in the steels. Where metal concentrations were reduced to 80%, for example, the concentration of implanted nitrogen was assumed to be 20%. Thickness of carbon contamination on steel surfaces could not be measured directly because the Table 3 “Theoretical”
Table 4 Thickness-concentration layers
Thickness-concentration (nm X at. fraction)
1018 52100 M-2 44OC 304
33 23 18 33 36
for all surface
cross section of carbon is not large enough to obtain a measurable peak for this layer. Thicknesses of the carbon layers could be estimated by observing the shift of the metal peak, but this calculation was heavily dependent on the calibration of the RBS detector system. As a result, error bars of t-5058 were quoted for the carbon layers. An estimate of the relative amounts of nitrogen retained in the steels was obtained by adding products of thickness-concentration terms for all surface microlayers in each specimen. Results of these computations are shown in table 4. It will be noted that the products listed in table 4 are generally proportional to the degree to which a steel benefits from the implantation of nitrogen, except, to some extent, for 440C. 3.3. SIMS ion microanalyzer data Shown in fig. 7 are processed data plots of ion intensity versus depth for FeN negative ions for steel specimens 1018, 52100, M-2, 44OC, and 304. The nearsurface nitrogen out-diffusion peak found in the 1018 Auger depth profile was confirmed by the SIMS results. The subsurface depths and the approximate shape of the main peak also were confirmed. The FeN negative
Layer 2 ‘)
Layer 3 ‘)
Layer 4 ‘)
30/Fe 0.9, N 0.1; and Bulk/Fe 1.0.
lOO/Fe 0.78, N 0.15, W 0.017, MO 0.035, Cr 0.02
30/Fe 0.84, N 0.1, W 0.017, MO 0.035, Cr 0.02
Bulk/Fe 0.93, WO.017, MO 0.035, crQ.03
150/Fe 0.66, N 0.2, Cr 0.14, MO 0.004
30/Fe 0.74, N 0.1, Cr 0.16, MO 0.004
Bulk/Fe 0.82, Cr 0.18, MO 0.004
150/Fe 0.54, N 0.22, Ni 0.08, Cr 0.16
30/Fe 0.63, N 0.1, Ni 0.09, Cr 0.18
Bulk/Fe 0.7, Ni 0.1, Cr 0.2
‘) The figure before the slash (/) indicates fraction.
0.9, N 0.1
0.85, N 0.15
0.75, N 0.25
0.9, N 0.1
in nm. Other figures indicate
IV. METALS Applications, Steels
Dodd et al. / Surface microanalytical
ion plots for 52100 and M-2 steels confirmed the broad, skewed nitrogen peaks seen in AES depth profiles. The SIMS data for 440C and 304 steels confirmed all features of corresponding AES depth profile curves.
4. Discussion Near surface composition variations in composition with depth and the chemical state of implanted ions are important questions to be addressed by surface microanalysis. The combined use of AES, RBS, and SIMS in this work illustrates an integrated approach to the problem. The data will be discussed with regard to the above questions, and the possible consequences for the use of nitrogen implantation for the improvement of tribological properties. The near-surface out-diffusion peak on the nitrogen profile for 1018 steel, shown at point “a” in fig. 1, has been found to be characteristic of soft, ferritic carbon steels generally and for pure iron [10,12]. The dip in the iron profile that is coincident in depth with the peak in the nitrogen profile, a consequence of less iron per unit volume in the analyzed surface layer, is probably an indicator of, but not positive evidence for iron nitrides. Finally, the carbon contamination layer (produced in all oil-diffusion-pumped ion implantation equipment), the presence of oxide at the steel surface, and the surface silicon contamination are examples of information obtainable from the surface analytical approach described here. The less intense, broad, and skewed nitrogen depth profiles for 52100 and M-2 steels, seen in figs. 2 and 3, and the absence of corresponding dips in the iron profiles suggest escape of nitrogen from the steel subsurface and less bonding of nitrogen in the form of iron nitrides. The peak positions for nitrogen maxima in 52100 and M-2 are near that for the out-diffusion peak in 1018, as noted in table 2. It has been reported that an out-diffusion peak is seen in nitrogen-implanted M-2 steel when the steel is in the soft, unhardened state [lo]. Several investigators have reported that hardened, martensitic 52100 does not benefit tribologically from nitrogen implantation [8,13,14]. This is, at least, consistent with the observation that less nitrogen is retained by 52100 steel than by 1018 and 304. There is another factor that may play an important role in the nitrogen implantation treatment of martensitic steels. This is the strong austenitizing character of implanted nitrogen. It is considered likely that nitrogen implantation in steels such as 52100, M-2, and 440C promotes the martensite-to-austenite transformation, resulting in softening of the steel surface and degradation of tribological properties in martensitic steels. It is not clear whether or not this austenitization during implantation of nitrogen in martensitic steel affects the amount
studies of N-implanted
of retained nitrogen and the nitrogen depth profile. The hardened and tempered stainless 440C steel depth profile shown in fig. 4 is more puzzling. Corresponding to the large nitrogen peak there are only weak dips in the iron and chromium profiles, suggesting that nitrogen may not be bound primarily as nitrides in this steel. On the other hand the RBS data and the SIMS plots both suggest some nitrides of chromium, and, probably, iron nitrides, as found by Singer  for 304 steel. At the chromium concentration level in 304 the dip in the chromium profile is just beginning to be apparent, as seen in fig. 5. One further point of interest seen in fig. 4 and noted in table 2 is the suggestion of a near-surface out-diffusion peak, at the shoulder on the 440C nitrogen depth profile. This observation, together with the small dip in the iron profile coincident with the nitrogen profile peak, may indicate that less nitrogen was held as iron nitrides in 44OC, and that some out diffusion occurred either during or after implantation. It is interesting to note that all three of the martensitic steels analyzed, 52100, M-2, and 44OC, for which nitrogen implantation did not show significant tribological benefits [7,13,14], showed evidence of a slight peak near the surface. Stable chromium nitrides are likely to be present in the implanted 440C surface, thereby accounting for more nitrogen retained in 44OC than in 52100 or M-2. Significantly less nitrogen was retained as iron nitrides in hardened, martensitic steels such as 52100 and M-2 than in hardened and tempered 440C stainless steel, but the chromium content of the latter steel probably contributed to the difference. A dip in the chromium depth profile curve coincident with the nitrogen profile peak can be seen clearly in fig. 5 for 304 steel. This together with the RBS and SIMS results and the independent confirmation of chromium nitrides in 304 steel  suggests that much, if not all, nitrogen is bound thus in 304. Improvements in tribological properties shown by nitrogen implanted 304 steel are related to stabilization of the surface austenitic phase, and this may be the most important reason for recommending nitrogen implantation treatment for 304. It has been reported that further tribological property improvement can be attained by the implantation of both nitrogen and titanium in 304 [lo].
5. Conclusions This work has demonstrated the value of an integrated surface microanalytical approach to the characterization of nitrogen implanted steels. The independent confirmation of depth profile characteristics and calibration of sputtering rates have been especially valuable. Implanted nitrogen depth profiles measured by AES and SIMS have yielded clues concerning relative
C.G. Dodd et al. / Surface microanalyrical
amounts of retained nitrogen and the manner in which nitrogen is retained in steel. RBS data have yielded measures of implanted surface region microlayer thicknesses and composition, as seen in tables 3 and 4. The total relative amounts of nitrogen retained in surface layers, as reported in table 3, is related to the tribological benefits attained by nitrogen implantation. This generalization held for all steels studied except 44OC, where the concept is complicated by the presence of martensite and carbides. The authors wish to acknowledge the assistance of Hillary Legg in coordinating the specimen preparation process. The encouragement and advice of Charles Evans at all stages of the work is noted with sincere appreciation. Support of the National Science Foundation is acknowledged under Grant No. DMR-8360256, NSFSBIR.
studies of N-implanted
References [l] N.E.W. Hartley
et al., J. Mat. Sci. Lett. 8 (1973) 900.  N.E.W. Hartley, Thin Solid Films 54 (1979) 177. (31 G. Dearnaley, in: Ion Implantation Metallurgy, eds., C.M. Preece and J.K. Hirvonen (The Metallurgical Society of AIME, 1980) p. 1. (41 J.K. Hirvonen, J. Vat. Sci. Technol. 15 (1978) 1662. IS] I.L. Singer, in: Ion Implantation and Ion Ream Processing of Materials, eds., G.K. Hubler et al., Materials Research Society Symp. Proc., vol. 27 (North-Holland, Amsterdam, 1984) p. 585.  D.M. Follstaedt et al., op. cit. ref. , p. 655.  L.E. Pope et aI.;op. cit. ref. [S], p. 661.  I.L. Singer and R.A. Jeffries, J. Vat. Sci. Technol. Al (1983) p. 317.  WC. Oliver et al., op. cit. ref. (51, p. 603. [lo] I.L. Singer and R.A. Jeffties, op. cit. ref. (51, p. 667. [ll] W.W. Hu et al., op. cit. ref. , p. 92. 1121 W.M. Bone et al., J. Vat. Sci. Technol. A2 (1984) p. 788. 1131 S.A. DiIIich et al., op. cit. ref. , p. 637. 1141 N.E.W. Hartley and J.K. Hirvonen, Nucl. Instr. and Meth. 209/210 (1983) 933. (151 I.L. Singer and J.S. Murday, J. Vat. Sci. Technol. 17 (1980) 327.
IV. METALS Applications, Steels