Professional Documents
Culture Documents
PDF The Physics and Technology of Laser Resonators First Edition Hall Ebook Full Chapter
PDF The Physics and Technology of Laser Resonators First Edition Hall Ebook Full Chapter
https://textbookfull.com/product/the-physics-of-laser-plasmas-
and-applications-volume-1-physics-of-laser-matter-interaction-
hideaki-takabe/
https://textbookfull.com/product/optics-photonics-and-laser-
technology-paulo-a-ribeiro/
https://textbookfull.com/product/optics-photonics-and-laser-
technology-2017-paulo-ribeiro/
https://textbookfull.com/product/laser-experiments-for-chemistry-
and-physics-1st-edition-robert-n-compton/
Basics of Laser Physics For Students of Science and
Engineering 2nd Edition Karl F. Renk (Auth.)
https://textbookfull.com/product/basics-of-laser-physics-for-
students-of-science-and-engineering-2nd-edition-karl-f-renk-auth/
https://textbookfull.com/product/laser-technology-applications-
in-adhesion-and-related-areas-1st-edition-k-l-mittal/
https://textbookfull.com/product/the-physics-of-solar-energy-
conversion-first-edition-bisquert/
https://textbookfull.com/product/handbook-of-laser-technology-
and-applications-volume-1-2nd-edition-chunlei-guo-editor-subhash-
chandra-singh-editor/
https://textbookfull.com/product/advances-in-laser-materials-
processing-technology-research-and-applications-2nd-edition-j-r-
lawrence/
The Physics and Technology o f Laser Resonators
The Physics and Technology of Laser
Resonators
Edited by
This book contains information obtained from authentic and highly regarded sources. Reprinted material is
quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts
have been made to publish reliable data and information, but the author and the publisher cannot assume
responsibility for the validity of all materials or for the consequences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic,
mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and
recording, or in any information storage or retrieval system, without written permission from the publishers.
T radem ark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.
Preface xiii
Unstable Resonators 21
P E D YER
2.1 Introduction 21
2.2 Advantages 21
2.3 Basic Concepts 22
2.4 Geometrical Mode Analysis 23
2.5 Practical Resonators 26
2.6 Near and Far-Field Pattern 27
2.7 Experimental Studies 29
2.8 Wave Analysis 30
2.9 Design Procedure 33
vi
4.4 Applications 72
4.4.1 Ring Laser Gyroscope 72
4.4.2 Nonplanar Solid State Ring Oscillator 75
4.4.3 O ther Applications 76
4.5 References 78
INDEX 247
LIST OF CONTRIBUTORS
K M ABRAMSM Optoelectronics and Laser Engineering, Department of
Physics, H eriot-W att University, Edinburgh, Scotland
On leave from The Technical University of Wroclaw, Poland
The topics covered in this book are centra! to the design, construction and
application of lasers. Most of the unique characteristics of lasers as optical
sources are linked to the properties of the laser resonator. This applies as
much to microscopic semiconductor lasers as to the high power industrial
material processing lasers whose volume may be more than 1011 times larger.
The idea for the book has grown from a 1-day tutorial-style meeting
organised under the auspices of the Quantum Electronics Group of the UK
Institute of Physics. Although the contents of the book are based on the
presentation given at this meeting considerable efforts have been made,
including substantial additional material, to shape the contributions into a
coherent whole.
The objectives have been to bring together in a single volume not only an
introduction to the underlying physics of laser resonators and laser beams, but
also material, not so readily available in text books, on the application of
basic principles to the design of resonators for most of the important types of
laser in common use today. Although, in some instances particular
technologies are described, the overall aim has been to emphasize design
principles rather than specific implementations.
Part One begins with a pivotal chapter describing the elementary principles of
laser beams and conventional 'free space' Gaussian resonators. It is followed
by four chapters which describe the design of certain variants on the basic
resonator theme. Thus, Chapter 2-5 cover 'Unstable Resonators, 'Waveguide
Resonators', 'Ring Resonators' and 'Multifold Resonators'. In each case, the
emphasis is on fundamental principles rather than technology. There has been
considerable interest recently in the use of laser mirrors which have spatial
variations in reflectivity as a means of controlling the transverse beam intensity
profile. This is covered in Chapter 6 describing 'Variable Reflectivity
Mirrors'. Recognizing that a number of important resonator types cannot be
described by simple Gaussian beam analysis, Chapter 7 has been included to
describe computer modelling of diffraction effects in 'Numerical Resonator
Calculations'. Many important applications depend on the availability of laser
sources of very narrow linewidths leading to the requirement for techniques for
'Frequency Stabilisation of Lasers' as described in Chapter 8.
Many lasers, particularly those operating at high average power produce
outputs which are not ideal TEM qq beams. Techniques for characterising the
'Propagation of Multimode Beams' and for 'Beam Analysis by Geometrical
Optics' are presented in the final two chapters of Part One.
The text is primarily intended for use in introductory courses in Lasers and
Optoelectronics at the level of final year undergraduate or first year graduate
students. It is also hoped that practising scientists and engineers and others
looking for an introduction to the field will benefit from the material and
design principles described.
We would like to thank a number of people, without whom this book could
not have been produced. Most importantly we thank the authors, whose
contributions make the book, for finding time in busy schedules to adapt and
expand their oral presentations to the needs of a text book. They have been
extremely patient in the face of our cajoling and editorial efforts. We
gratefully offer our thanks to colleagues and graduate students at H eriot-W att
University for responding so patiently and conscientiously to requests to
proof-read parts of the manuscript and particularly to Dr Howard Baker for
additional editorial assistance; to Mrs Lesley Lumsden for typing part of the
manuscript and Mrs Sherron Gilroy who prepared some of the illustrations.
Finally, our gratitude and admiration are due to Alan Colley for his creative
wordprocessing and text productions skills, his conscientious attention to detail,
amazing capacity for work and unflagging enthusiasm; he is mainly responsible
for the high quality of the camera-ready copy.
1 L aser R e s o n a to r s a n d G a u ssia n B e a m s
A C W A L K E R
1.1 INTRODUCTION
The first laser to be operated - the ruby laser in 1960 - and the vast
majority of other lasers developed since, consist of a gain medium placed
within an optical resonator (or cavity). Although the gain medium, which
gives each type of laser its name, determines the output power/energy and
ultimate tuning range, it is the resonator that controls the spatial quality and
spectral line width of any particular laser. As these two parameters are the
very factors that make lasers unique light sources, with high spatial and
temporal coherence, it follows that the design and construction of the
resonator deserves close attention.
Over the last 30 years or so, a wide range of resonators have been
developed and exploited within commercially available laser systems. These
include: free-space stable standing-wave and ring cavities, both constructed
with discrete mirrors; three dimensional guided-wave resonators formed out of
the gain medium itself or with external reflectors; and unstable or
non-confining cavities relying on mirrors of restricted transverse dimension or
spatially graded reflectivities. This chapter introduces the basics of laser
resonator design, in particular discussing what are called conventional stable
resonators and the spatial properties of the light inside, and emitted from,
such cavities. Subsequent chapters cover the physics of some of the more
complex resonators such as rings, unstable and waveguide resonators.
In general, a stable resonator, whether acoustic, electrical or optical, can
be defined as an extended environment within which the relevant type of wave
or oscillation can circulate in a self-consistent fashion. That is, every time
the excitation returns to a particular point in space, its phase and amplitude
are the same as the previous occasion. In this way, when driven for times
longer than this circulation time, constant amplitude constructive interference
occurs with consequent strong coupling between the driving mechanism and the
resonator. Thus, in the same way as with musical ietruments, where the
resonator (sound-box) efficiently couples the energy source (eg. vibrating
string) to free-space sound waves; a laser optical resonator provides a
mechanism by which the gain medium can be coupled to well defined
free-space optical modes.
The simplest, and most commonly used, laser resonator is the Fabry-Perot
2 Laser Resonators and Gaussian Beams
cavity, in which two mirrors oppose each other across the gain medium (See
Figure 1.1). Such a resonator provides physical confinement in one dimension
only and consequently light simply reflects back and forth along the axis
Fig 1.1 A resonant mode of an optical Fabry-Perot cavity. The curve represents the
instantaneous standing wave produced by the oscillatory electric field circulating
within the resonator. The two mirrors are separated by q half-wavelengths.
Phase
Thus: X =— (l.l)
<7
where q is an integer and n is the average refractive index over the length of
the cavity, L. As indicated in Figure 1.1, this corresponds to q
half-wavelengths fitting between the mirrors and is precisely the same
condition as for maximum transmission through a conventional Fabry-Perot
interferometer. Each value of q corresponds to a different longitudinal (or
"axial") mode of the cavity. In the absence of any additional
frequency-selective components inside the cavity, as many modes will oscillate
as are compatible with the gain line-width.
The separation between modes is constant in the frequency domain (p =
c/X) and is readily calculated using equation (1.1):
&v = v q+ l - vq
= £ z « + 1> - i k *
A C Walker 3
i.e. (1 .2)
F = 0.3)
1 - R e - 01*
where R is the geometric mean of the two mirror irradiance reflectivities and
a is the internal power loss coefficient. The mode width (FWHM) is then
given by:
5 v « Av/ F 0.4)
Amplitude
For the light flux within the cavity to build up from the initial low level of
spontaneous emission, the round trip net gain must be greater than unity. As
long as this is the case, the optical field grows exponentially. Clearly, such
an amplification cannot proceed indefinitely and in practice, the gain starts to
fall or saturate as the power-demand placed on the gain medium exceeds the
limits imposed on it by the rate of excitation (i.e. the pump power). As this
happens the resonator round-trip gain falls until it reaches unity. Any further
decrease in gain would result in decay of the optical field and recovery of the
gain. In a continuously pumped laser, the power circulating within the
resonator stabilisesat a level at which the saturated gain just balances the
internal losses. Thus the resonance condition that the amplitude be
reproduced each round-trip is, as with most types of resonator, automatically
satisfied.
It is clear from the above discussion that the transmission of the output
mirror is critical in determining the power output of the resonator. If this
transmission is very high then there is a large cavity loss factor and the build
up of internal irradiance, and hence the output power, is limited by the need
to maintain a large amplification factor in the gain medium. Conversely, if
the output mirror transmission is very low then the available output power is
shared between this low loss and other comparable sources of loss in the
4 Laser Resonators and Gaussian Beams
cavity, eg. absorption, scatter, reflections off internal optical surfaces, etc.
ABSORPTANCE, A
Fig 1.2 (a) Plot showing that to maximise the output irradiance I oup from a laser
resonator, an optimum output coupler reflectivity R, must be selected. Output
irradiance has been normalised to the saturation irradiance of the gain medium
I$ (b) Curves showing how the optimum reflectivity RQpt depends upon the
small-signal, double-pass gain factor G and the parasitic round-trip less factor
A. Both plots are based on the infinite plane-wave mean-field equation 1.7.
Writing the small signal gain factor for a double pass through the gain
medium as G t and assuming a saturation response appropriate for a
homogeneously broadened transition, then this relation becomes:
A C Walker 5
( 1 .6)
where I is the mean internal cavity irradiance and /§ the saturation irradiance
of the gain medium. Using Im t = -^I(ln R ) then:
lnC
0.7)
■]
For particular values of the gain and loss factor, G and A, equation (1.7)
indicates an optimum output coupler reflectivity, R, which maximises I oa f
Figure 1.2 shows examples of the dependence of the optimum R upon G and
A, deduced from equation (1.7).
The mean-field theory approach described above is only strictly valid for
low A, and R close to unity, when the spatial variation of irradiance along
the length of the gain medium is small. Otherwise the spatial dependence of
gain saturation, and the positions in the cavity of those components
contributing to parasitic losses, should properly be taken into account within a
more detailed theoretical description.
In general, a cavity of the type indicated in Figure 1.1, i.e. formed from two
plane mirrors, will not act as a stable resonator. Any beam of finite
cross-section within such a cavity will diverge due to normal diffraction
processes and consequently the transverse amplitude profile of the optical field
will not be reproduced on each round trip.
Within this book a variety of techniques for overcoming this divergence
are discussed, including: (i) transverse confinement of the optical field
(waveguide lasers); (ii) coupling out of the cavity all light beyond some radius
(unstable resonators); and (iii) the inclusion of focusing elements inside the
cavity of sufficient power to compensate diffraction spreading (stable
resonators). The latter approach, usually exploiting focussing cavity mirrors, is
one of the most common techniques and is the major topic of the remainder
of this chapter.
distance L along the z-axis, and in the absence of any optical elements, the
initial values of r1 and pj are transformed to r 2 = r1 + L p 1 and p 2 = pj (see
Figure 1.3b). This relationship between these four parameters can be
summarised in a vector/matrix equation of the type:
( 1 . 8)
B, (1.9)
G i * = G1
[Mil ft
In this simple example the matrix is the ray-transfer matrix
corresponding to propagation of the ray over some distance, L. Similar
ray-transfer matrices can be constructed for any optical transformation
including the action of refracting surfaces, thin lenses, spherical mirrors, etc.
(a)
1 0
Mm - 2 /R 1
Fig 1.3 (a) Definition of a ray vector, R (z), for a ray at position z
within an optical system. (b) Ray transfer matrix, for free-space
propagation over distance L. (c) Ray transfer matrix, Mfn, for a concave mirror
with radius of curvature R.
[ m j = [ \ 2 /R J] d -ii)
Note that this matrix does not account for the change in direction of the ray
on reflection, because in this analysis it is convenient to refer the angles, p,
to the general direction of propagation.
Ray transfer matrices can be used to determine the conditions under which the
simplest open resonator, formed by two spherical mirrors, are stable. At this
point it should be noted that for a specific ray starting at some arbitrary
position and angle, the resonator stability condition does not demand that the
ray return to this same position and angle on every round trip. It simply
requires that the light remain inside some finite boundary, such that all rays
contribute to a reproducible amplitude distribution (transverse mode). It is
therefore necessary to compute the effect on a ray of an arbitrary large
number of cavity round-trips and determine whether the components of the
ray vector will remain real and finite.
The general structure assumed for an open resonator is the common
arrangement of two spherical mirrors shown in Figure 1.4. (Other elements,
Fig 1.4 An optical cavity of length L formed by two spherical mirrors with
radii of curvature i?j and i?2 *
IA B in = 1i rrA s i n ( n 6 ) - s i n ( n - l ) I B s i n (n8)
lc Di nsin# L C s i n (n6) D s i n ( n 6 ) - s i n ( n - 1) ,] 0 . 1 3 )
For [M ^]n to be real, so too must be the term sin0. Given that A, B C
8 Laser Resonators and Gaussian Beams
k(A+D)2 < 1
r 2L 2L 21? 1
-1 < < 1
L ‘ R2 + RXR2 \
( 1 . 16)
This simple relation (1.16) specifies the conditions under which a cavity is
stable. It is sometimes written, using g1 = t - L / R 1 and g 2 = 1 - L I R ^ as 0 <
8182 < 1 *
1. PLAN E-CO N CA VE
R = 00 , L < R„ < co
Figure 1.5 shows the most common stable resonator geometries. By far the
most convenient and frequently used cavity is the plane-concave configuration.
One advantage of this arrangement is that the plane reflector can be the
partially transmitting mirror and consequently, as both its surfaces are easily
made flat, no geometric divergence/convergence is imposed on the output beam
- unlike the concave reflector where, when acting as the output coupler, the
outer surface has to be polished accurately convex to avoid negative lens
effects. A second advantage of the plane-concave cavity is that those
frequency selective elements (e.g. reflecting gratings, etalons) that are most
readily available, act as planar optical elements and can be conveniently
incorporated in this type of cavity as direct replacements for the flat mirror.
(In this case the output often has to be the curved mirror, with an
appropriately correcting outer surface).
The limitations on the values of R^ and R 2 that are specified in Figure
1.5 indicate the borderlines between stable and unstable configurations. These
A C Walker 9
borderline cavities are not normally used, since small manufacturing errors or
the presence of additional (weak) optical elements inside the cavity (including
lens effects in the gain medium itself) could make them unstable. Within the
unstable region, the resonator becomes highly lossy as the rays diverge to
infinity within a finite number of round trips. Thus for the plane-concave
resonator (Figure 1.5), the plane-plane limit, R t = R 2 = °°s should be avoided,
along with the geometry R 2 = L. The latter case corresponds to the
concentric cavity limit in the symmetric resonator, R t = R 2 = LZ2 (x = 0 in
Figure 1.5), as clearly the plane-concave cavities are simply folded equivalents
of the symmetric variety.
Cavities near the confocal limit R l = R 2 = L (y = 0 in Figure 1.5) are
only unstable if the symmetry is broken. For example, assuming 5 is small,
R l = L + 5, R 2 = L - 5 is unstable, while R l = L + 5, R 2 = L + 5 is
stable. This has no parallel in the plane-concave cavities as these are
inevitably always equivalent to symmetric resonators.
The "focal” symmetric cavity in which the focus ofeach mirror lies on
the surface of the opposing mirror, ie Ri = R 2 = 2L, is the only example of
this type which is not at a stability boundary. In fact it is a highly stable
configuration - well away from the marginal confocal (l?j = R 2 = LZ2) and
plane-plane (R^ = R 2 = °°) resonator geometries.
Besides the cavities discussed above their exist many other geometries that
satisfy the stability condition of equation (1.16). These include the asymmetric
concave mirror combinations and cavities in which one cavity mirror is convex
(R < 0). Such cavities are used relatively rarely.
The stability conditions derived in the previous section by ray analysis remain
equally valid when diffraction is included. A detailed diffraction analysis shows
that a number of transverse amplitude distributions - transverse modes - can
resonate stably in any one cavity.
The lowest order mode - the mode that is confined most strongly near
the optical axis - has a Gaussian amplitude distribution. That is the electric
field amplitude, E, varies in the transverse dimension, r, as:
E = E0 e x p ( - r 2/w 2 ) ( 1 .1 7 )
where E0 is the central amplitude (at r=0) and w is ameasure of the width
10 Laser Resonators and Gaussian Beams
v ( z ) = w o [1 + [ 7 % ) 2 V ( 1 - 18)
(iii) The divergence corresponds to the (initially plane) wave-fronts
becoming spherical (again with symmetry about z = 0, w = w0)
with radius of curvature given by:
R (z) = z [ l + [ ^ ] 2] (1.19)
From the above properties of Gaussian spherical waves we can deduce the size
of the lowest order transverse mode of any resonator. The relevant boundary
condition is that the wave-front curvature must match the curvature of the
mirror when reflecting from it. In this way it will simply retrace the same
pattern each time it crosses the cavity.
For the symmetric concave cavities, it is clear that the waist of the beam
- the narrowest, plane-wave part of the beam (z = 0, w = wQ) - must be
located at the centre and the mirrors, therefore, at z = ± L /2 (see Figure
1.7). Using equation (1.19), it can then be concluded that the mirror
curvature determines wQ through the relation:
2x1*$ 1 2i
XL
. 7T . (R1+R2 - 2 L ) 2
Thus it is apparent that the stability condition for a Gaussian mode within a
cavity is precisely consistent with that derived from geometric ray analysis -
equation (1.16).
When calculating the lowest order transverse mode geometry for a general
cavity, only the waist dimension - equation (1.21) - and the waist position are
required. The deviation of the latter from the resonator centre, simply
follows from the degree of mirror curvature asymmetry. Thusthe ratio ~ zx/ z 2
12 Laser Resonators and Gaussian Beams
7 = ~ (1.23)
z1 Ri - L
Finally, it can be deduced that the beam radii on each mirror are:
= rXRi i * r 71 l i (1.24)
i i J l rJ
rXRn i r l i
0.25)
I i J i Rt +R2 -L i
1.4.3 Fresnel Number and Diffraction Losses
Having determined the dimensions of the lowest order transverse modes, there
remains the question: how large should the mirrors be to avoid significant
losses resulting from light diffracting past their edges? Clearly, such a loss
factor can never be completely zero as all Gaussian-mode irradiance
distributions theoretically extend to infinity.
A useful dimensionless parameter in discussing this aspect of resonators is
the Fresnel number, N:
N = a 2/LX (1.26)
w (1.27)
[ ^ ] ! [ t^ ] ‘
For typical L /R values, the second bracket is of order unity. If N ~ 1, then
XL ~ a2 and hence a = wy?r.
The power loss for a gaussian beam of spot-radius w, reflecting from a
mirror of radius a = w/rr is just 0.2% (see section 1.5.3) and it can therefore
be concluded that the condition N ^ 1 is indeed generally sufficient to ensure
negligible diffraction losses.
Only when the plane-plane cavity condition is approached, L /R 0, does
the final term in equation (1.27) lead to significant round-trip diffraction
losses in an N = 1 cavity. In this regime, where the Gaussian mode volume
ultimately diverges to infinity at R 1 = R 2 = °°, the resonator modes are no
longer Gaussian but are instead determined dominantly by the mirror
A C Walker 13
higher phase velocities. Thus, when more than one transverse mode is
oscillating, the laser output is notsingle frequency (even if it is nominally
single longitudinal mode) and thus temporal coherence is degraded.
The degree by which the higher-order modes deviate in frequency from
the TEM qq^ axial mode is given by:
aA - ( L \ ) i (1.30)
1.5.1 The Paraxial Wave Equation and the Complex Beam Parameter.
where the electric field E is an oscillatory function of time and also varying
in space and where k=2%l\ is the wave vector. There are an infinite number
of possible solutions to this equation, dependent upon the boundary conditions
Another random document with
no related content on Scribd:
may be bought for 2 cents a pound or less, whereas an iron of
WILD HEATS.
After steel is melted, whether in a crucible, an open hearth, or a
Bessemer vessel, it boils with more or less violence. This boiling is
caused by ebullition of gases, and if steel be poured into moulds
while it is boiling the resulting ingot will be found to be honeycombed
to an extent that is governed by the degree of the boiling.
If a heat boils violently and persistently, it is said to be “wild,” and
if a wild heat be teemed the ingots will be honeycombed completely;
such ingots cannot be worked into thoroughly sound steel, and no
melter who has any regard for his work will teem a wild heat if he
knows it.
To stop the boiling is called “dead-melting,” “killing” the steel, so
that it shall be quiet in the furnace and in the moulds.
A crucible-steel maker who knows his business can, and he will,
always dead-melt his steel. It only requires a few minutes of
application of a heat a little above melting temperature, and this can
be applied by a skilled melter without burning his crucible or cutting
down his furnace; this is indeed about all of the art there is in
crucible-melting, the remaining operations being easy and simple.
Dead-melting in the Bessemer vessel is not possible by increase
of time; wild heats are managed differently, probably by adding
manganese or silicon, or both, but exactly how is not within the
author’s experience.
Dead-melting in the open hearth would appear at first sight to be
always possible, but there are more difficulties in the way than in the
case of crucible-melting.
The heat may be wild when the right carbon is reached, and then
the melter must use a little ferro-silicon, or silico-spiegel, or highly
silicious pig, or aluminum, and he must use good judgment so as not
to have his steel overdosed with any of these. From half an ounce to
an ounce of aluminum to a ton of steel is usually sufficient, and
although any considerable content of aluminum is injurious to steel
there is little danger of its being added, because of its cost, and
because a little too much aluminum will cause the ingots to pipe from
top to bottom.
Silicon seems to be the most kindly element to use, and it is
claimed that a content of silicon as high as 20 is not injurious; some
people claim that it is beneficial. That it does help materially in the
production of sound steel there can be no doubt, and if such steel
meets all of the requirements of the engineer and of practice it would
seem to be wise not to place the upper limit for silicon so low as to
prevent its sufficient use in securing soundness. But the author
cannot concede that as much as 20 silicon is necessary. In crucible
practice high silicon is not necessary; in “melting-iron,” or iron to be
melted, it means so much dirt, indicating careless workmanship; but
there will always be a little silicon present which the steel has
absorbed from the walls of the crucible during the operation of
melting. In high tool-steel silicon should be at the lowest minimum
that is attainable.
This discussion of wild heats may appear to be outside of the
scope of this work, and to belong exclusively to the art of
manufacturing steel, of which this book does not pretend to treat.
This is true so far that it is not recommended that the engineer shall
meddle in any way with the manufacturer in the management of his
work; on the other hand, it is vital to the engineer that he should
know about it, because wild steel may hammer or roll perfectly well,
it may appear to be sound, but the author cannot believe that it is
ever sound and reliable.
Again, it has a scientific interest; that wildness is due to too much
gas, and probably to carbon-gas, may be shown by an illustration.
It has its parallel in the rising of the iron in a puddling-furnace at
the close of the boil, a phenomenon with which every one is familiar
who has watched a heat being boiled or puddled. That all of the iron
does not run out of the puddling-furnace at this stage is owing to the
fact that there is not heat enough in the puddling-furnace to keep the
iron liquid after it has been decarbonized.
During the running of a basic open-hearth furnace an apparently
dead heat was tapped; before the steel reached the ladle there was
a sort of explosion; the steel was blown all over the shop, the men
had to run for their lives, and not one tenth of the steel reached the
ladle. The manager was rated roundly for carelessness in not having
dried his spout, and the incident closed. A few days later another
quiet heat was tapped and it ran into the ladle; about the time the
ladle was full the steel rose rapidly, like a beaten egg or whipped
cream, and ran out on to the floor, cutting the sides of the ladle, the
ladle-chains, and the crane-beams as it flowed. The men ran, and
there was no injury to the person.
Again the manager was blamed, this time for having a damp
ladle, and he was notified of an impending dismissal if such a thing
occurred again. He protested that he knew the ladle and the stopper
were red-hot, that he had examined them personally and carefully,
and knew he stated the truth.
There were several reasons for looking into the matter farther:
first, the man in charge was known to be truthful and careful, so that
there was no reason for doubting his word; second, if the vessel and
rod were red-hot, there could be no aqueous moisture there; and,
finally, such an ebullition from dampness was contrary to experience,
as a small quantity of water under a mass of molten iron, or slag,
results almost invariably in a violent explosion, like that of
gunpowder or dynamite.
Upon inquiry it was found that prior to both ebullitions there had
been a large hole in the furnace-bottom, requiring about a peck of
material to fill it in each case. Magnesite was used; the magnesite
was bought raw, and burned in the place. It is well known that it
takes a long time and high heat to drive carbonic acid out of
magnesite, and it was surmised that insufficient roasting might have
caused the trouble. Samples of burned and of raw magnesite were
sent to the laboratory, and the burned was found to contain about as
much carbonic acid as the raw magnesite. Then the case seemed
clear: This heavily charged magnesite was packed into the hole; the
heat was charged and melted. The magnesite held the carbonic acid
until near the close of the operation; then the intense heat of the
steel forced the release of the gas, which was at once absorbed by
the steel. Owing to the superincumbent weight of the steel the gas
was absorbed quietly, and when the weight was removed the gas
escaped, exactly as it does at the close of puddling or in the frothing
of yeast.
Whether the carbonic acid remained such, or whether it took up
an equivalent of carbon and became carbonic oxide, and then again
took up oxygen from the bath, and so kept on increasing in volume,
is not known.
The facts seem clear, and the collateral proof is that thorough
burning of the magnesite, and of any dolomite that was used,
prevented a recurrence of any such accidents.
Such ebullitions have occurred and caused the burning to death
of pitmen, and the statement of the above case may be of use to
melters in the future who have not met such an experience.
NITROGEN.
The real influence of nitrogen is not known to the author. Percy
shows that nitrogenized iron is hard, exceedingly friable, and causes
a brilliant, brassy lustre. He also says nitrogen is driven out at a
yellow heat; doubtless this is true of the excess of nitrogen, but it has
been shown in Chapter II that melting in a crucible will not drive the
nitrogen out of Bessemer steel.
When crucible-steel not made from Bessemer scrap and
Bessemer steel of equal analysis are compared in the tempered
condition, there is almost invariably a yellowish tinge over the fresh
Bessemer fracture which distinguishes it from the crucible-steel. The
Bessemer steel is also the weaker. These differences are believed to
be due to nitrogen.
Langley maintains his belief that oxygen is still the chief mischief-
maker; the author believes nitrogen to be the more potent of the two;
there is no known way to remove the nitrogen, and there the
question stands.
ELEMENTS OF DISINTEGRATION.
It has been stated time and again that these impurities are
elements of disintegration, and that it would be wise in every case to
restrict the quantities allowable within reasonable limits, giving the
steel-maker sufficient leeway to enable him to work efficiently and
economically, and at the same time to keep the quantities of these
impurities as low as possible.
On the other hand, able, successful, and conservative engineers
have claimed that if the steel-maker meets their physical
requirements as shown by prescribed tests they, the engineers,
should be satisfied; that they should not interfere with chemical
composition, as they had no fear of subsequent disintegrations.
This argument was answered by the statement that skilled steel-
workers could manipulate poor steel so as to bring it up to the
requirements; that the well-trained workers in the bridge-shops
would not abuse the steel; that the inherent deficiencies would not
be developed; the work would go out apparently satisfactory; and
that it might remain so for a long time, in the absence of unusual
shocks or strains, but that in an emergency such material might fail
because of deterioration where a purer material would have held on.
In the absence of proofs such statements have been met with a
smile of incredulity.
Fortunately some proofs are now at hand, and as the method of
getting them has been obtained, more will follow from time to time.
In Engineering, Jan. 17, 1896, Mr. Thomas Andrews, F.R.S.,
M.Inst.C.E., gives the following cases:
A fracture of a rail into many pieces, causing a serious accident.
A broken propeller-shaft which nearly caused a disastrous
accident.
Analysis of the rail:
Carbon 0.440
Silicon 0.040
Manganese 0.800
Sulphur 0.100
Phosphorus 0.064
It is clear that the sulphur is excessive, and that it was neutralized
so as to make the steel workable by an excess of manganese.
Of the propeller-shaft Mr. Andrews says chemical analysis of
outside and central portions of the shaft showed serious segregation.
“The percentage of combined carbon was nearly 50 per cent
greater in the inside of the shaft than on the outside; the manganese
was also in excess in the inside of the shaft; the phosphorus and
sulphur had also segregated in the interior of the shaft to nearly
three times the percentage of these elements found near the outside
of the shaft.”
Unfortunately Mr. Andrews does not give the analysis of the
shaft.
A number of micro-sections of the rail and of the shaft were made
and examined.
“Numerous micro-sulphur flaws were found, varying in size from
0.015 inch downward, interspersed or segregated in the
intercrystalline junctions of the ultimate crystals of the steel, and
being located in such a manner as to prevent metallic cohesion
between the facets of the crystals, thus inducing lines of internal
weakness liable to be acted upon by the stress and strain of actual
wear.”
The dimensions of these flaws in the rail varied from .0150 ×
.0012 to .0010 × .0004 parts of an inch.
In the shaft from .0160 × .0030 to .0020 × .0016 parts of an inch.
In the rail he found as many as 14 flaws in an area of only
0.00018 square inch, equal to nearly 60,000 flaws per square inch.
In the shaft he found as many as 34 flaws in an area of only
0.00018 square inch, equal to nearly 190,000 per square inch.
In speaking of the shaft he says: “In addition to blow-holes, air-
cavities, etc., the interior of the shaft was literally honeycombed with
micro-sulphide of iron flaws, which were meshed about and around
the primary crystals of the metal in every direction.” “The deleterious
effects of an excess of manganese in interfering with the normal
crystallization of the normal carbide of iron areas were also
perceptible.”
As the number of micro-sulphur flaws in the shaft were about
three times as many as in the rail, we may assume that the shaft
contained at least as large a percentage of sulphur as the rail, and,
owing to the general honeycombed structure, it would not be a far
guess to assume that the steel was teemed wild.
“The deleterious effect of these treacherous sulphur areas and
other microscopic flaws, with their prolonged ramifications spreading
along the intercrystalline spaces of the ultimate crystals of the metal
and destroying metallic cohesion, will be easily understood.”
“Constant vibration gradually loosens the metallic adherence of
the crystals, especially in areas where these micro-flaws exist.
Cankering by internal corrosion and disintegration is induced
whenever the terminations of any of the sulphide areas or other
flaws in any way become exposed at the surface of the metal, either
to the action of sea-water, or atmospheric or other oxidizing
influences. In many other ways, also, it will be seen how deleterious
is their presence.”
“Internal micro-flaws of various character are nevertheless almost
invariably present in masses of steel, and constitute sources of initial
weakness which not unfrequently produce those mysterious and
sudden fractures of steel axles, rails, tires, and shafts productive of
such calamitous results. A fracture once commencing at one of
these micro-flaws (started probably by some sudden shock or
vibration, or owing to the deterioration caused by fatigue in the
metal) runs straight through a steel forging on the line of least
resistance, in a similar manner to the fracture of glass or ice.”
It is understood that similar investigations are being carried out
on an extensive scale by Prof. Arnold; in the meantime the above
cases should satisfy any one that these impurities are elements of
disintegration, and that the less there are of them in any steel the
better for the steel.
It seems clear that if 10 sulphur will cause 60,000 flaws per
square inch, 01 sulphur ought not to cause more than one tenth of
that number; or, if an equal number, then they could only be one
tenth of the size.
The segregation found in the shaft is so excessive that it would
seem probable that there was a good deal of sin there also; but,
even if it were unavoidable segregation, the harm would have been
just so much the less if there had been less of total impurities
present to segregate.
ARSENIC.
Arsenic is known to be very harmful in tool-steel, and it is proper
to assume that it can do no good in structural steel. In any case
where the properties of steel do not come up to the standard to be
expected from the regular analysis examination should be made for
arsenic, antimony, copper, etc. These are not as universal
constituents of steel as silicon, phosphorus, sulphur, and
manganese, but they are present frequently, and in any appreciable
amount they are bad.
XI.
THEORIES OF HARDENING.
SEAMS.
In tool-steel there should be no seams at all. Some makers
declare that in high steel, seams are evidences of good quality; such
a statement is the veriest fraud; it is hard to get any high steel free
from seams, and therefore if the maker can get the user to believe
that a seam is a good thing he can enhance his profit; that is, he can
enhance it for a time until his fraud is understood.
Some seams are hard to see; when there is reason to suspect
one, a little filing across the line will show it in a distinct black line if it
is there. A file is an indispensable tool for an inspector, better than a
chisel or a grindstone.
In machinery and structural steel a few small seams may be
unobjectionable; too close inspection may lead to unnecessary cost
without a compensating gain; still every engineer should reserve the
right to determine what seams are allowable and what are not, for
his own safety.
Laps should not be tolerated in any work.
Torn cracks on edges or surface indicate burned steel or red-
short steel; they should not be allowed.
The grain of steel should be practically uniform, not too coarse,
not with brilliant lustre, nor with a dark india-ink tint. With an even
fine grain, a bright lustre may indicate a mild steel not worked badly.
Inspectors must learn by practice what is tolerable and what is not,
as it is impossible to lay down hard and fast rules; it is safe, however,
to say that a fairly fine grain of even texture, not much lustre, and no
india-ink shade, is indicative of good heating and proper working.
With these few general hints the subject must be left, for, like
tempering, inspecting is an art in itself, and it cannot be taught in a
book.
An expert inspector will see seams and pipes with his naked eye
that a novice could not detect with an ordinary magnifying-glass.
It may do no harm to the inspector to suggest to him that
amiability and good sense are the best ingredients to mix with sound
judgment.
If he will cultivate these, and learn to distinguish between a mere
blemish and a real defect, he will find his work made easy and
pleasant; and he will be far less likely to have bad work thrust at him
than he will if he makes it apparent that he regards himself as the
only honest man.