A geochemical investigation of "basalts" in southern Arizona

Item Type text; Thesis-Reproduction (electronic)

Authors Halva, Carroll Joe, 1938-

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 30/09/2021 09:11:23

Link to Item http://hdl.handle.net/10150/553953 A GEOCHEMICAL INVESTIGATION

OF "BASALTS" IN SOUTHERN ARIZONA

by

Carroll J. Halva

A Thesis Submitted to the Faculty of the

DEPARTMENT OF GEOLOGY

In Partial Fulfillment of the Requirements . For the Degree of

MASTER OF SCIENCE

In the Graduate College

THE UNIVERSITY OF ARIZONA

1961 STATEMENT BY AUTHOR

This thesis has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this thesis are allowable without special permission, provided that accurate acknowledgment of source is made. Requests of permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in their judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED:

APPROVAL BY THESIS DIRECTOR

This Thesis has been approved on the date shown below:

\ Paul E. Damon Professor of Geology ACKNOWLEDGEMENTS

I would like to thank several people who greatly aided and assisted in the technical work during this investigation and others who contributed their time and knowledge of the general geology of southern Arizona for their part in this report. This investigation was sponsored by a research assistantship in geochemistry which was supported in part by the Atomic Energy Commission under contract

A T (11-1)-689 and by the Research Corporation. This work was done under the direction of Dr. P. E. Damon.

X would like to thank Dr. P. E. Damon for his excellent and able guidance during this investigation which consisted in part of many helpful suggestions concerning the work along with technical advice in neutron activation and radiometric work. Professors

J. F. Lance and J. V. Anthony of the University of Arizona also contributed many suggestions which were of value in this investigation.

Professors M. H. Wittmeyer and T. W. Fern of the Nuclear

Engineering Department aided greatly in the initiation and successful completion of the neutron activation work. Their able assistantship and general interest in the program is greatly appreciated.

Dr. J. Fuchs of Arizona State University at Tempe made it possible to perform the emission spectrographic work by allowing the investigator to use his excellent laboratory facilities.

Mr. Carl Hedge of the geochemistry laboratories did most of the flame photometric work as well as aiding in the development of the E.D.T.A. method of analysis. The author is grateful for his aid and assistance. The members of the geochemistry laboratories also contributed greatly with timely suggestions and helpful crit­ icisms of this investigation. t a b l e o f c o n t e n t s

Page Chapter I - INTRODUCTION

1.1 Location of Area Under Investigation ...... 1

1.2 Purpose of the Investigation...... 1

1.3 Previous Investigations ...... 1

1.4 Present Investigation ...... 7

Chapter II - GEOLOGIC SETTING

2.1 General G e o l o g y ...... 11

2.2 Stratigraphy and Late Volcanic Sequence ...... 11

2.3 General Description of Sample Areas ...... 16

2.4 Mineralogy and Petrology of the B a s a l t s ...... 20

Chapter III - GEOCHEMICAL STUDY OF MAJOR ELEMENTS

3.1 Chemical Analysis of Major E l e m e n t s ...... 24

3.2 Radiometric Studies ...... 33

3.3 Location of Potassium in Bas a l t s ...... 42

Chapter IV - GEOCHEMICAL STUDY OF TRACE ELEMENTS

4.1 Emission Spectrographic Studies ...... 48

4.2 Neutron Activation Studies ...... 52

Chapter V - CONC L U S I O N S ...... 57

Appendix A - A SEMI-MICRO ANALYSIS OF SILICATE ROCKS FOR Ca, Mg, Fe and A1 EMPLOYING E.D.T. A...... 61

Appendix B - RADIOMETRIC ANALYSIS OF POTASSIUM IN SILICATE ROCKS ...... 67

Appendix C - NEUTRON ACTIVATION STUDIES OF B ASALTS...... 73

BIBLIOGRAPHY ...... 86 ILLUSTRATIONS Page

Fig. 1.1 Location of Area Under Study ...... 2

Fig. 1.2 Location of Samples in Southern A r i z o n a ...... 9

Fig. 2.1 Sequence Correlation of Late Volcanics in Southern Arizona ...... 13

Fig. 3.1 Calcium Oxide Content vs. Potassium Content for Southern Arizona "Basalts" ...... 32

Fig. 3.2 Potassium Content vs. Alpha Activity for Volcanics of Southern Arizona ...... 38

Fig. 3.3 Potassium Content vs. Alpha Activity for "Basalts" 41

Fig. 4.1 Neutron Activation Curves for Southern Arizona Volcanics...... 55

TABLES

I Chemical Analysis of Major Elements in Southern Arizona "Basalts" ...... 25

II Comparison of Data for Potassium Determination . . . 26

III Variations in CaO Content within a Single Flow . . . 26

IV Chemical Analysis of Basic Rocks ...... 28

V Radiometric Analysis of Volcanic Rocks ...... 34

VI Radiometric Variations within a Single F l o w ...... 36

VII Potassium Content of Plagioclase and Pyroxene in Basaltic Rocks ...... 46

VIII Spectrographic Data Obtained for Various Volcanics . 50

IX Artificial Produced Long Lived and Common Stable Isotopes of the Elements Commonly Found in Basalts . 76

X Average Compositions of Various Type Basalts .... 81

XI Sensitivity Limits for Elements Commonly Found in Basalts ...... 84 INTRODUCTION

1.1- Location of Area Under Investigation

This report concerns a study of "basalts” and some associ­ ated volcanic rocks in southeastern Arizona. This area includes

the counties of Pinal, Pima, Graham, Greenlee, Santa Cruz and

Cochise. The basaltic rocks under consideration are those of post-

Cretaceous age. These "basalts” outcrop quite widely over this area and samples have been collected from several localities.

Fig. 1.1 shows the outline of the area under consideration.

1.2- Purpose of the Investigation

The purpose of this investigation was two-fold. First,

an investigation to determine the chemical makeup of the "basalt"

flows has been conducted in order to understand more fully the

geochemical nature of the samples. Secondly, an attempt has been

made to combine this geochemical investigation with a geological

investigation in order to correlate or otherwise relate these

volcanics in time and space.

1.3- Previous Investigations

The "basalts" in southern Arizona, except in a few cases,

have not been thoroughly studied from a specific viewpoint. Much

1 MOHAVE

COCONINO NAVAJO APA CHE

Y A V A P A I

6 I LA MAR I C 0 P A YUMA

P I N A L G RA HAM

COCHISE MAP OF ARIZONA l"=50Mi SANTA CRUZ \ \ \ AREA UNDER STUDY i \ \ V

Fig. l.l Location of area under study.

2 3

of the previous literature, in which the basalts were studied, was

generally concerned with other topics and the basalts were only mentioned because of their proximity to the particular subject of

interest. This does not imply, however, that there is not a great

deal of material in the literature concerning the basaltic rocks.

Many studies of geologic interest in this area have mentioned

basalts as an accessory to the main line of study. Though alone

this material does not mean much, when several studies are

examined some characteristic features concerning basalts become

apparent.

Perhaps the earliest work concerning basalts in this area

was done by Tolman (1909). Tolman discusses the geology of

Turaamoc Hill. He determined that there were five separate basalts

in the Tumamoc sequence including three flows and two basaltic

intrusions. The basalt flows are the youngest igneous rocks

extruded or intruded into this area. These flows overlie an

older andesite and rhyolite flow in the same vicinity. His report

includes a section on folding and faulting as well as a petro­

graphic study of the various flows. He considers the sequence to

be Quaternary in age.

Smith (1927) describes a basalt intterbedded with valley

fill which he claims is Quaternary in age in his study of the

Commonwealth Mine. According to Smith the basalt is the youngest

rock present in this area and is underlain by two different series

of flows. The oldest series consists of flows of andesite while 4

the middle series consists of rhyolite, basalt, obsidian and

rhyolite in this order. A chemical analysis of two basalt sections was made by Smith and is presented below. It is of interest to

note the high potash content of the two basalts.

Oxide Turkey Creek Hills Section Gibbon Hills Sectii % Z

Si02 51.32 49.74 AI2O3 15.87 16.67 Fe2°3 3.87 6.24 FeO 8.37 3.08 MgO 6.88 5.22 CaO 8.21 9.84 k 2° 2.36 3.03 NaO 1.03 1.37 h 2o 1.08 2.61

Higdon (1935), in his study of the Sunshine mine, describes

a basalt composed mainly of feldspar with a glassy groundmass. The

basalt invades Cretaceous rocks and is thus post-Cretaceous in age.

Hasor (1937) describes an olivine basalt in the Tombstone

area which he suggests is a ring dike. No age relationships are

mentioned. The basalt is composed of nlagioclase, olivine and

augite.

Harshroan (1939), in his description of the Superior area,

makes note of the fact that the basalts are the latest igneous

rocks in this area. He assigns these basalts to the late Tertiary

or early Quaternary period for two reasons. One is the fact that

they are interbedded with the Gila Conglomerate, and the second

reason is because they fill north-south fault zones.

Popoff (1940) describes the basalts as being Tertiary to 5

Quaternary in age in the Rosemont mining area on the basis of cross-cutting relationships. The basalt is the youngest igneous rock in the area. The basalt overlies Cretaceous sediments and cuts through andesite of Miocene age in the Red Rock area. The basalt which he describes is quite amygdaloidal.

Jones (1941) in the Sycamore Ridge area, Johnson (1941) in the Helvetia Mining District, and Sopp (1940) in the Mountain Mine area all discuss the presence of basalt dikes of recent age. They do not, however, discuss their stratigraphic or age relationships to any extent.

Dikes of rhyolite and andesite porphyry cut basalts of

Tertiary age in the Putnam Wash area of Pinal county according to

Hillebrand (1953). The basalt in this area is mostly dike-like but appears to be related to volcanic flows in nearby areas.

Voelger (1953) has located pipes of porphyritic basalts, similar to those found at Sentinel Peak, in the Belota Ranch area.

These basalts are the youngest intrusives mentioned and are under­ lain in places by andesite. Evidence presented by Voelger indicates that the lavas were laid down after the lower Rillito beds but before the middle Rillito sequence. This is based on

the absence of volcanic pebbles of this porphyritic material in the lower Rillito but these pebbles are present in the middle

Rillito.

In the Aqua Fria River area the volcanics rest on an 6 unconformity of 20, 50 to 100 feet of relief. St. Clair (1957) has assigned these volcanics to the Tertiary period. The volcanics are olivine basalts, hornblende basalts and hornblendeandesites.

Basalt dikes also occur in this area composed of 36% plagioclase,

47% pyroxene, 9% olivine, 5% biotite and 3% iron oxide.

Bryner's (1959) work in the Comobabi Mountains apparently show the basalt dikes in this area to be the youngest rock of igneous origin. This is based upon crosscutting relationships.

Donald (1959) discusses basic dikes in the Baboquivaris and points out the considerable variations in their feldspar composition.

Colby (1958) has done extensive work in the Tucson Mountain

Red Beds. Here he has found basalt overlying sediments. Gas bubbles have appeared to flow through the sediments. The basalts are late Cretaceous or younger in age based on faulting relation­

ships. The basalts are intruded by latites in some areas.

Wargo (1959), in his study of the School House Quadrangle

in New Mexico, discusses the thick accumulations of volcanics in

this area. Wargo also has tried various methods of correlation

in an attempt to interrelate various flows from different local­

ities of his area. The basalts in this area are the youngest

volcanics present. They are relatively flat lying and have been

assigned by Wargo to the Plio-Pleistocene age.

Taylor (1959) discusses the possibility of correlation of 7 volcanics in Santa Cruz county and the Tucson Mountains. Radio- metric , spectrographic and petrographic studies have been coupled with a geologic study of this area in an attempt to correlate various lava flows. A general stratigraphic column has been drawn based on field evidence. Again, the basalts in this study seem to be the youngest igneous rocks present. Taylor also makes mention of the high potassium content of the Tumamoc Hill series of

"basalts."

Upon close examination of previous investigations several facts become apparent. The basalts are widespread in southern

Arizona and occur both as dikes and flows. The basalts are generally the youngest rocks in the igneous sequence, and usually overlie either andesite or sediments. The basalts are commonly interbedded with sediments and are generally later than major faulting. The relative thickness of the flows vary from locality to locality. No definite age can be assigned to the basalts, but, in most cases, they are probably Plio-Pleistocene.

1.4-Present Investigation

During this investigation samples of basaltic rock of post-Cretaceous age were collected from various localities in southern Arizona. Fig. 1.2 denotes the areas which were sampled.

Over one hundred samples were collected and particular attention was paid to the stratigraphy and petrology of the underlying forma­

tions as well as the "basalts." These samples were subjected to 8

Key to Figure 1.2

1 - Sawtooth Mountains

2 - Sasco Area

3 - Tortilita Mountains

4 - Galiuro Mountains

5 - White Mountains

6 - Tumamoc Hill

7 - Black Mesa

8 - San Xavier Mission

9 - Benson Highway "Basalt"

10 - Saguaro National Monument

11 - Santa Catalina Wash

12 - Reddington Pass

13 - Mineta Beds

14 - Tombstone Basalt

15 - San Bemadino Area

16 - Atascosa Mountains

17 - Pajarito Mountains

18 - Santa Rita Mountains SOUTHEAST ARIZONA

PINAL GRAHAM GREENLEE

PIMA COCHISE

JSANTA SCALE '« CRUZ

Fi-:. 1.2 L-cation of samples in Southern Arizona 10 alpha and beta counting, chemical analysis, emission spectographic analysis, neutron activation studies, hand specimen description and petrographic work. GEOLOGIC SETTING

2.1- General Geology

Host of the area under investigation is considered a part of the Basin and Range Province of the Western United States. The

Basin and Range Province in this area is exemplified by north-south trending basins and ranges formed mainly by block faulting. Much of the material in the basins is unconsolidated or loosely con­ solidated conglomeritic gravels of Tertiary and Quaternary age.

The mountain ranges are composed of various age rocks and are generally parallel to each other. Volcanic rocks of felsic, inter­ mediate and basic compositions are found throughout the area. The basic volcanics seem to cap the igneous sequence of extrusive rocks and represent the latest igneous activity in this area.

The exact age of the volcanics is unknown but are generally believed to be late Tertiary or early Quaternary.

2.2- Stratigraphy and Late Volcanic Sequence

As mentioned previously the basic volcanics are the latest igneous extrusives present. These "basalts" are underlain in most localities by either andesite or conglomeritic sediments of probably Miocene age. When one studies the section in this area

11 12 several factors concerning the later extirusives become very apparent. Fig. 2.1 shows the volcanic sequence at several local­ ities in this area.

From Fig. 2.1 it can be seen that the basaltic rocks over- lie the andesites in most sections which were studied. There may be sediments between the "basalts" and the andesites, however. It is generally believed that the andesites were formed at an earlier stage of volcanism and subject to faulting, folding and erosion before the "basalts" were extruded over the area. Tolman (1909) discusses this. This being the case, the basaltic rocks may or may not overlie the andesites directly depending upon whether the andesites were previously being eroded away or were in slight basins which received sediments. The presence of sediments between the "basalt" and andesite marks a hiatus in extrusive activity in this area.

The first basaltic rock which was extruded over the ande­ sites or sediments, whichever the case may be, varies quite mark­ edly from the overlying basic flows. It is generally very porphy- ritic, containing plagicclase phenocrysts as much as 6 cm. long and somewhat smaller phenocrysts of augite. Tolman (1909) noticed that this particular basaltic layer was quite common within fifty miles of Tucson. He accounted for its porphyritic nature by claim­ ing the magma had a chance to crystallize in the magma chambers while the andesites were being folded and faulted. The faulting San Bernadino Area Andesite [ *•

Sediments and Rincon Mountains Andesite d mO \ 1%

Black Mesa

Tumanoc Hill

/ y y y

Sawtooth Mountains "Basalt" Andesite

\

Sasco Area Basalt " Andesite U' Kg*. 2*1 Sequence correlation of volcanics late Arizona Southern to in (not Tortilita Mountains g Andesite Lia.

15 14 cut off the extrusive pipes which brought the lava to the surface and thus allowed the phenocrysts to develop. When sufficient pressures were built up to cause the magma to seek new paths to the surface this "basalt" was the first lava to be extruded over the area.

This is a plausible mechanism for the formation of this unit. As mentioned by Tolman the unit is found in many localities in this area. Cooper (1960) also discusses the widespread occurrence of this unit. During the present work samples of it have been collected or observed in eight different localities#

Though the outcrops are not extensive in any area and the general appearance of the unit is definitely of a flow nature the samples are remarkably similar. The areas in which this basic porphyry

occurs include A-Mountain, Black Mesa, San Xavier Mission, on the

Benson Highway, near the Saguaro National Monument, Reddington

Pass area, in the Mineta Red Beds and in the Sawtooth Mountains.

Localities of this unit are indicated on Fig. 1.2. In addition,

in the San Bemadino area, a porphyry which is much more basic in

appearance is present but probably not related to the other units.

Hand specimens from the San Bernadino area do not resemble the

others. The Sawtooth unit also differs from the others in having

biotite phenocrysts present and much smaller plagioclase pheno­

crysts. Though this porphyry unit is found in several localities

it probably does not represent the same flow. Most likely these 15 units were formed at approximately the same time but were, of course, emitted from several fissures.

The basalts above the porphyry unit are either olivine containing basalts or very vesicular, amygdaloidal basalts. The thickness of the overlying "basalts" varies considerably reaching a maximum of approximately 700 feet in the Tumamoc Hill sequence and the Black Mesa area. The total thickness of these units originally may have been greater but due to mass wasting and erosion the total thickness has been lessened somewhat. With the exception of the

San Bernadino region where craters are very much in evidence, this observation is further strengthed by the general absence of any primary features associated with volcanics in this area.

However, the apparent lack of any primary features may be more superficial than one realizes. In the region around the base of A-Mountain cindery and scoriaceous material can be found scattered throughout the area. Tolman in his work discusses the removal of similar material from the Tumamoc sequence near the

turn of the century for retaining and ornamental walls within the city of Tucson. In the Black Mesa area a thin veneer of very vesicular material can be found at the top and on the flanks of

the mesa. Mass wasting is far from complete in the Black Mesa area where large vesicular boulders are found near the top of the

sequence. When viewed from a distance both the Black Mesa and

Tumamoc Hill sequence have a conical shaped appearance. In the 16

Black Mesa area a general alignment of volcanic knobs is quite apparent. It is very possible that these thick accumulations of volcanic materials may be located along the axes of fissure flows.

2.3-General Description of Sample Areas

Fig. 1.2, which shows the areas from which samples were collected, can also be used as a reference figure for the general description of the sample areas. The Tumamoc Hill sequence is located on the southwest side of the city of Tucson. It consists of three basaltic flows overlying andesite and two basic intrusions.

The total thickness of the unit is approximately 700 feet and the

total area occupied is about one square mile. The flow immediately above the andesite unit is the porphyry unit discussed in the pre­ ceding section. A tuff unit of approximately forty feet in thick­ ness separates the top flow from the underlying flow. The Tumamoc

Hill sequence is surrounded by valley fill of relatively recent age.

Directly south of this area lies the Black Mesa sequence which is about the same thickness as the Tumamoc Hill sequence.

No andesite is visible in this area but it probably is present

below the alluvium and valley fill which surround the mesa. Six

separate basaltic flows are present in this area, the bottom most

being the porphyry unit which is very similar in appearance to the

Tumamoc Hill porphyry. The total area occupied by the Black Mesa

sequence is about twice as large as the Tumamoc unit. Cindery

material is very much in evidence in this area. 17

East of Black Mesa lies the San Xavier Mission unit which consists of a small knob approximately 125 feet in total thickness.

This unit is almost entirely made up of the porphyry material found at the base of the two previous areas. The total outcrop area is about one-eighth of a square mile. No andesite is visible below this unit but again it is probably hidden under the alluvium and valley fill which surround the area.

The Sawtooth Mountain volcanic range gets its name from the sawtooth like appearance that the steep, jagged andesite faces present. The andesite section here is very thick consisting of several hundred feet. "Basic" dikes cut the andesite unit. Immedi­ ately above the andesite is a "basalt" porhyry which differs from the other porphyries in having numerous subhedral to euhedral shaped biotite crystals and much smaller plagioclase phenocrysts.

The pyroxene phenocrysts in this unit are not nearly as numerous.

Over this unit lie basaltic rocks of a later age which are not conspicuous because they grade out into the valley fill and into the Sascoarea "basalts." The total outcrop area of the Sawtooth

Mountains is several square miles. The area is surrounded by volcanics, valley fill and a quartz diorite which appears to under­ lie the whole area and may represent the basement rocks of this region.

East of the Sawtooths lies the Sasco area, so called because the ghost town of Sasco lies approximately in the middle 18 of this region of volcanic flows. Both andesites and "basalts" are present in this area. The andesites are the oldest flows.

The porphyry unit was not found in this area but could be present.

The relief is never much greater than 200 feet. The total thick­ ness of the units is difficult to determine. The total outcrop area is rather ill-defined but is approximately 15 square miles.

East and south of the Sasco region is the Tortolita

Mountains which, except for the northern end, is composed mainly of shists and granites. The northern end has thick andesite units which are overlain only in certain localities by thin basaltic

flows. Some basic dikes also occur in this region. No outcrops of the porphyry unit are found in this area.

East of Tucson on the Benson Highway is found the porphyry unit interbedded with the Rillito beds which are Miocene or younger

in age. This unit is approximately thirty feet thick in this area

and dips with the Rillito beds to the east at an angle of about

twenty-five degrees. The porphyry unit is a flow unit and not

a dike or sill. Field evidence indicates baking and alteration

of the lower contact while the upper contact shows no signs of

baking and weathering features are present. This unit resembles

the porphyry unit of Tumamoc Hill or Black Mesa.

In the Black Hills to the east of the Galiuro Mountains a

moderately thick "basalt" flow is found which overlies welded tuff 19 and bedded tuff units. It is approximately fifty to eighty feet

thick and occurs over a fairly wide area. This unit caps the

igneous sequence in this section and is very finely crystalline.

Whether more than one basaltic flow is present has not been

determined.

In the center of Cochise county near the town of Tombstone

a basaltic unit is found which appears to have the general appear­

ance of a ring dike. This unit, east of Tombstone, is finely

crystalline and outcrops only over a half square mile. The relief

within the whole basalt outcrop is not much more than thirty feet

and its contact with the surrounding alluvium and valley fill is

very indistinct. It was not possible either to determine the

thickness of this unit or whether or not it overlies andesites

or sediments.

In the extreme southeast corner of Cochise county the San

Bernadino volcanic field, composed mainly of basalts of varying

and great thickness, is present. Numerous flows of volcanic

material, mainly basaltic in composition, are present and primary

features of the volcanoes such as craters and conical shaped hills

are also present. In the southern part of this region the basalts

overlie an andesite unit of unknown thickness and some alluvial

sediments. The basalt immediately above the sediments in this area

is pprphyritic in appearance but it does not resemble the other

porphyry units in this area. This porphyry is much more basic and 20 contains only phenocrysts of plagioclase which are not nearly so numerous as in other localities. Geologically very little is known about this area but it would certainly be of interest for future

investigations.

In addition, basalt samples were collected from the White

Mountains, from the Carrizo volcanic field, eastern New Mexico,

and from Table Mountain, Colorado, to use for comparison . The

Columbia River Basalt and samples of basalts from Santa Cruz

county were available to the investigator for this purpose.

Samples of the porphyry unit were also collected in other

localities including the Reddington Pass area where it overlies

sediments and andesites and was first described by Voelger (1953).

These samples and samples further south described first by Chew

(1952) were collected by Mr. Wendel Carlson. The southernmost

units were collected in the Mlneta Red Beds. Samples were also

collected, though not thoroughly studied, from the Twin Hills

area west of the Saguaro National Monument.

2.4-Mineralogy and Petrology of the Basalts

The various basaltic flows in southern Arizona do not have

a complex mineralogical composition. They are composed chiefly of

plagioclase, pyroxene, olivine and magnetite. The plagioclase

content varies from 40-70%, pyroxene 5-20%, olivine 0-20%, and

magnetite from 5-15%. No primary quartz, orthoclase or other

minerals, with the exception of biotite in the Sawtooth flows, were 21 visible in hand specimen or thin section. In addition, usually

5-15% glassy material was present though in the Sasco and Tortolita flows this value was nearer to 30%.

The plagioclase phenocrysts in the porphyry unit are approx­ imately 50an-50ab in composition. Tolman identified the plagioclase as labradorite in his work. Taylor (1959) identified the same phenocrysts as andesine. Chew (1952) agreed with Tolman in his

study of the porphyry unit. Thin section work performed during

this investigation on the plagioclase crystals led to the con­

clusion that the feldspar is 50-50 ab-an in composition based on

the extinction angles. Chemical analysis of the feldspar placed

it at 55%an-45%ab. If one allows for the potassium present to be

in the form of orthoclase in the plagioclase phenocrysts, then this

composition is approximately 50an-50ab.

Various workers have noted that the phenocrysts of plagio­

clase in extrusive rocks are more basic than that of the groundmass.

This conclusion can be predicted from Bowen's Reaction Series. Mo

attempt was made to identify the plagioclase of the groundmass of

either the porphyry unit or the overlying flows other than noting

the plagioclase was in the andesine-labradorite region. It would

not be out of the question to assume that the feldspar would be

between 50-60% ab in the groundmass. This would lead to a lower

content of calcium in the rock and a higher content of sodium.

Chemical analysis of the "basalts," discussed in another section 22 of this report, verified this conclusion.

The pyroxene present is augite in all cases observed. The pyroxene is unusual in that it contains twice as touch magnesium as calcium and is therefore a sub-calcic variety of augite. The properties of the pyroxene all resembled the optical properties of augite with the exception of a slight pleochroism observed in almost all the thin sections. The pyroxene was definitely not pigeonite. Chemical analysis also indicated no significant amounts of sodium or potassium. This eliminated the possibility of the pyroxene being a mixture of aegerine or aegerine-augite and augite.

The pyroxene must be classified, therefore, as a sub-calcic augite.

No chemical analysis was made on the olivine because of the

difficulty of separating it, but it may safely be assumed that it

is mainly forsterite because of the high magnesium content of the various flows. However, the olivine commonly showed alteration

halos of goethite.

Secondary minerals found in the vesicles of the basalts

were chiefly either quartz or calcite. In no case - did the vesicle

fillings make up 10% of the rock. Most of the vesicles in the

samples were not filled.

On the basis of handspecimen evidence alone one would

probably classify these rocks as olivine basalts due to their dark

color, the presence of plagioclase, pyroxene, olivine and magnetite

and the absence of other primary minerals. Petrographic studies 23 would also possibly lead to the same conclusion except one might wonder about the true composition of the microlaths of plagioclase found in the groundmass. The-presence of magnetite and olivine would probably cause the petrographic investigator to classify these as basalts in the final analysis. GEOCHEMICAL STUDY OF MAJOR ELEMENTS

3.1-Chemical Analysis of Major Elements

In order to better understand both the chemical and geo­

logical nature of the "basalts" several samples were analyzed,

either partially or completely, for Ca, Mg, Fe, Na and K. The results of these analyses are shown in Table I.

Calcium, magnesium and iron were determined by the use of

EDTA by a procedure described by the author (1960). Appendix A

describes the complete procedure. Sodium was determined with a

Perkin-EImer flame photometer using lithium as an internal standard.

Potassium was determined via two methods. All the potassium results

reported in Table I were based on radiometric determinations. Ad­

vantage was taken of the fact that potassium 40 is beta active.

The procedure for determining the percent potassium present was

that described by Damon, Hedge, Taylor and Halva (1960). A complete

description of this procedure can be found in Appendix B.

In order to verify and compare the radiometric potassium de­

terminations, several samples were run by Mr. Carl Hedge on the

flame photometer using lithium as an internal standard. The results

of this comparison are shown in Table II. The standard deviation

24 25

Table I-Chemical Analysis of Major Elements in Southern Arizona . Basalts

Sample %CaO ZMgO %FeO %Na %K

White Mtn. basalt 3.97 6.38 6.10 1.71 1.92 (B-31-60) Galiuro basalt 4.04 5.69 6.88 1.76 2.38 (B-25-60) Tombstone basalt 5.85 5.27 7.58 1.78 1.25 (B-12-60) Sasco. basalt. 4.43 2.66 5.91 1.92 3.96 (B-2-60) Sawtooth basalt 4.40 6.26 3.92 1.95 1.74 (B-3-60) Black Mesa basalt 5.60 5.63 7.77 2.13 2.81 (B-39-60) San Xavier basalt 3.70 2.17 6.50 2.44 2.54 (B-42-60) Benson basalt 4.05 1.59 2.41 (B-50-60) Carrizo basalt 5.38 6.12 11.60 1.15 0.81 (B-43-60) Rincon basalt 4.12 3.06 4.98 2.47 (B-22-60) San Bernadino basalt 6.56 9.29 1.32 (B-52-60) Table Mtn. (Colo.) 3.72 3.59 10.93 2.35 4.04 (B-46-60) Tortolita dike 4.72 6.06 6.90 1.90 1.98 (B-18-60) Tumamoc Hill. Flow 5 3.50 3.90 3.87 2.22 2.87 Flow 4 4.30 4.72 1.80 Flow 3 3.11 3.41 2.94 Dike 2 3.14 3.56 2.95 Dike 1 4.13 4.60 1.60

Black Mesa basalt Top flow 3.45 4.83 4.86 2.00 3.06 Fifth flow 5.10 2.77 5.06 2.81 Third flow 6.32 4.18 8.39 1.40 Second flow 4.80 3.99 9.16 2.35 2.62 Bottom flow 3.86 1.93 5.65 .2.06 2.81

Average Southern 4.69 5.17 7.02 1.94 2.23 Arizona basalt 26

Table II-Comparison of Data for Potassium Determination

Sample Radiometric method Flame-photometer 7. difference

Tumamoc Hill basalt 1.62 1.97 9.50

Black Mesa basalt 1.51 1.46 1.99

Tombstone basalt 1.37 1.25 4.36

Sasco basalt 3.60 3.27 4.58

Tortolita dike 1.98 1.77 5.55

Sawtooth basalt 1.74 2.05 9.52

Galiuro basalt 2.38 2.47 1.89

White Mtn. basalt 1.93 1.79 3.63

Columbia River basalt 0.88 0.82 3.41

Atascosa Mtn. basalt 1.04 0.98 2.88

Table III-Variations in CaO Content within a Single Flow

Sample 7. CaO

1 3.50

2 3.55

3 3.30

4 3.52

5 3.44 27 of the radiometric method is approximately 4-7% under normal labo­ ratory procedures. The standard deviation via the flame photometric method was no greater than 4%. From Table II it can be seen-that the two methods of analysis compared favorably. In most cases the variations between the two results fell within the allowable standard deviation. This is really quite significant if one con­ siders the sample size used in making both determinations. For the flame photometric method only 0.2 gram of sample was used, while in the radiometric method several grams of sample material were used. It should be pointed out that no effort was made to bring the results into closer agreement. These results were determined by a single analysis in each case. Undoubtedly, if several deter­ minations were made both radiometricslly and flame photometrically the results would probably be in much better agreement.

Table III shows the variations in CaO content within a single flow in the Tumamoc Hills sequence. The maximum variation is approximately 10% which indicates that the basaltic flows are quite uniform at least in their calcium content. Radiometric analysis for the potassium content in a series of samples collected from a single flow also indicated that the potassium content was fairly uniform. The results of this analysis are discussed in the section on radiometric studies. It appears, therefore, that the major chemical components are distributed quite evenly within a single flow unit. 28

Upon referring to Table I it becomes obvious that the so- called basalts are in many instances not basalts but potassic basaltic andesites. This is based on the average chemical analysis of the samples compared with various types of basic rocks given in

Table IV below. ,

Table IV-Chemical Analyses of Basic Rocks

Tvoe Rock %MgO 7.FeO 7JTa20 TKgO %Ca0 7.SiOz 7A1203

Trachyandesite 2.5 8.1 8.8 5.3 1.5 56.2 17.4

Hornblende andesite 2.5 5.5 4.0 2.4 5.5 60.8 17.3

Augite andesite 3.4 6.6 3.6 1.9 6.3 59.3 16.6

Trachybasalt 5.4 11.7 3.9 2.8 8.2 48.8 16.0

Ave. basalt 6.0 11.7 3.2 1.6 8.9 48.8 15.8

Plateau basalt 6.4 13.3 2.9 1.0 9.7 49.3 14.1

Daly's andesite 2.8 6.3 3.6 2.1 5.9 60.3 17.5

Daly's diorite 4.2 7.2 3.4 2.2 6.8 57.6 16.9

Basaltic andesite 4.7 7.5 1.9 2.0 7.4 59.0 13.0

Southern Ariz. 5.2 7.0 2.6 2.7 4.7 58.0 16.4 "basalt"

The first six analyses above were taken from Tyrrel (1929),

the next two from Daly (1933) and the basaltic andesite from Faust and Callaghan (1948) in their study of the Currant Creek Magnesite.

From the above data the chemical content of the basaltic flows of 29 southern Arizona most closely resembles the chemical content of the augite andesite, Daly’s average diorite or the basaltic andesite.

In all three cases the silica and iron oxide determinations agree quite favorably with the southern Arizona basalts. The alumina content agrees quite closely with the diorite or augite andesite.

The alkali and alkaline earth elements, however, disagree in all three cases. The combined alkali or alkaline earth elements in each of the three rock types agrees very favorably with the combined contents of these elements for the "basalts” of southern Arizona.

As mentioned previously the basic flows in this area are quite rich in olivine and contain pyroxene in which the magnesium content is twice that of the calcium content. Using this infor­ mation one would expect the rock to be enriched in magnesium relative to calcium. Therefore, the relationship of magnesium to calcium would be reversed due to the presence of these high mag­ nesium containing minerals.

Since the rock is obviously not an augite andesite and if

the reaction series had continued would probably have reacted to

form different minerals, this leaves the diorite or the basaltic

andesite as the most similar chemically. The sodium oxide content

for the basaltic andesite is quite low when compared with the

sodium oxide content of the Arizona "basalts." However, it should

be pointed out that the plagioclase phenocrysts in the Currant

Creek samples were bytownite rather than the ab^Q-an^Q phenocrysts 30 from this area. This would lead to a higher sodium content for the basalts from this area. The alumina content of the basaltic andesite is lower than the alumina content of the Arizona "basaltic" flows. However, the plagioclase content of the Currant Creek basaltic andesite is much lower than the plagioclase content of the basic flows from this area. The potassium content of the

Arizona "basalt" is higher than that of the basaltic andesite, thus indicating a potassium rich rock. Therefore, the best name for the .southern Arizona "basalts" would be potassic basaltic andesites.

Though most of the basaltic rocks in southern Arizona appear to be potassic basaltic andesites a few of them can be classified as true basalts. The Atascosa basalts described by

Taylor (1959) have a very low potassium content and compare

favorably with the Columbia River Basalt in alpha and beta activity. Samples from Tombstone and San Bemadino can also be classed as true basalts on the basis of potassium content and alpha and beta activity. The Carrizo basalt of Central Arizona also appears to be a normal basalt on the basis of chemical

composition.

Though the chemical content resembles that of the diorite,

obviously this rock would not be classed petrographically as a

diorite.

The percent potassium determined by radiometric means is a very good indicator of the total felsic or basic nature of the rock. 31

Fig. 3.1 demonstrates the relationship between calcium oxide and potassium for a group of potassic basaltic andesites samples from this area for which both the calcium oxide and potassium content were known. Table I, from which this data was taken, also shows this same relationship. The relationship between the iron oxide and potassium content also demonstrates this same pattern. The relationship between magnesia and potassium does not show any sharp correlation though a general trend exists.

As a conclusion to this section of the work it can be stated that the basaltic flows in southern Arizona, when viewed from a strict chemical analysis, are much more felsic than formerly thought. The felsic nature of the "basalts” is demonstrated most clearly by the high potassium content and low calcium content of the rocks. Silica and alumina analyses were done on a sample which closely resembled the average basalt from this area and it was found to contain 58% SiOg and 16.4% AlgOg. Combining this information with the material already discussed the basic rocks in this area would best be classified as potassic basaltic andesites.

The classification of the "basalts" by chemical analysis does not agree with the classification one would predict via either petrographic or hand specimen identification. Undoubtedly, in the case of intrusive rocks or rocks of granitoid texture, the petrographic method of classification is much better than the 7

o

0 6 0

0

5

t: 0 § 0 CaO 0 0 01 4 0 0 0 1 2 0 0 5 0 0 5 0 0 0 0

2

0 ) 0

0

0 ______. ______

Sample number

PlS* ).l Calcium oxide content vs. potassium content for southern Arizona "basalts"

52 33 chemical method. However, in the case of extrusive rocks, where the groundmass is glassy or microcrystalline, the chemical classi­ fication or identification modified by petrographic description is preferable.

Since studies of the location of the potassium indicated that it was concentrated in the groundmass, and, most probably in the glassy material, this demonstrates that the basic flows in this area are the result of an incomplete reaction halted by extrusion.

If the reaction had been complete, phenocrysts of orthoclase or sanidine undoubtedly would be present within the flows and thus the petrographer or field geologist would probably classify these units as andesites or more felsic rocks. Furthermore, if the reaction had been allowed to continue, the olivine and pyroxene could very well have been converted to hornblende, thus further substantiating the more felsic classification.

3.2-Radiometric Studies

Most of the samples collected during this investigation were analyzed radiometrically to determine their beta and alpha activity as well as their potassium content. The results of this study are partially shown in Table V. During this study tuffs, andesites and some rhyolites were analyzed along with the basaltic rocks.

The alpha activity of the samples is reported in counts 34

Table V-Radiometric Analysis of Volcanic Rocks

Sample % K s.d.% «*.p.h. s.d. p.p.m. s.d.

Sasco Andesite 3.96 2.29 189.02 3.42 5.34 0.10 Sasco Basalt 1.65 4.61 227.96 4.16 3.04 0.08 Sawtooth Andesite 5.42 5.09 445.84 5.75 8.31 0.30 Sawtooth Andesite 5.38 1.75 365.33 9.53 7.84 0.09 Sawtooth Andesite 3.99 2.25 656.26 6.40 6.69 0.09 Picacho Intrusive 5.06 3.89 183.94 6.01 6.49 0.21 Tombstone Basalt 1.40 4.21 78.47 2.49 1.95 0.06 Tombstone Basalt 1.37 3.80 73.94 6.06 1.90 0.05 Sawtooth Basalt 1.63 7.09 188.43 6.71 2.79 0.12 Sawtooth Basalt 2.25 Tortolita. Andesite 5.49 4.23 212.74 7.77 7.13 0.25 Tortolita Basalt 2.00 3.61 131.58 3.51 2.88 0.08 Galiuro Basalt 2.25 4.33 77.25 3.04 2.86 0.10 Black Mesa . Top flow 3.06 4.58 226.11 10.76 4.55 0.14 Fifth flow 2.82 4.29 180.54 4.36 4.04 0.13 Fourth flow 1.40 4.83 160.82 4.07 2.40 0.07 Third flow 2.62 5.70 163.51 8.19 3.74 0.16 Second flow 2.81 5.44 150.10 8.00 3.88 0.16 Bottom flow 2.11 4.36 402.30 17.98 4.48 0.08

Tumamoc Andesite 1.34 7.43 245.67 4.47 2.79 0.10 Tumamoc tuff #1 3.40 4.46 628.28 6.30 7.10 0.16 Tuff #2 6.64 2.15 148.37 5.62 8.03 0.15 Tuff #3 5.53 322.00 7.46 Tumamoc Basalt Top flow #1 2.87 266.00 3.87 #2 2.73 286.00 3.91 n 3.02 208.00 4.23 #4 2.66 271.00 4.26 #5 2.53 273.00 4.00 #6 2.78 245.00 4.59 #7 2.49 282.00 4.00 #8 4.08 #9 4.15 #10 4.02 Flow 4 #1 1.82 6.32 216.64 3.95 3.16 0.12 #2 1.61 5.37 459.74 7.88 4.25 0.08 Flow 3 #1 2.95 2.38 192.19 6.63 4.25 0.07 #2 2.94 188.00 4.30 35

Table V-Radiometric Analysis of Volcanic Rocks, continued

Sample 7. K s.d.% p.h. s.d. 0.p.m. s.d.

Tumamoc Basalt, contd Intrusive 2 #1 3.08 4.24 246.36 8.26 4.69 0.13 n 2.95 229.00 5.34 Intrusive 1 n 1.87 5.92 121.20 6.40 2.69 0.12 White Mtns. Basalt #1 2.62 5.19 149.20 8.75 3.66 0.14 Basalt #2 2.58 2.80 157.72 4.40 3.65 0.07 Basalt #3 3.39 4.81 127.57 8.18 4.38 0.17 Basalt #4 1.35 3.34 148.27 5.78 2.28 0.04 San Xavier Basalt 2.54 2.89 382.65 5.77 4.84 0.08 Benson Hwy. Basalt 2.40 7.54 427.62 8.40 4.94 0.19 Hineta Basalt 3.30 2.41 177.08 9.44 4.55 0.08 (Chew) Reddington Basalt 2.47 2.89 328.94 10.47 4.47 0.07 (Voelger). San Bernadino Basalt #1 1.14 7.71 94.95 6.45 1.76 0.09 #2 1.33 5.80 91.40 3.48 1.94 0.08 #3 1.49 3.56 58.98 2.10 1.94 0.06 Carrizo Basalt. 0.82 7.24 51.39 2.53 1.16 0.06 Table Mtn. (Colo.) 4.04 1.72 270.93 11.45 5.86 0.06 New Mex. Basalt 2.72 4.68 136.41 4.14 3.70 0.14 Catalina Wash #1 4.37 4.22 136.31 3.92 5.44 0.20 Catalina Wash #2 7.78 1.18 138.20 7.05 9.21 0.10 Black Mesa Basalt #1 2.98 4.65 599.96 6.78 6.58 0.15 Basalt #2 2.18 3.79 206.57 6.31 3.51 0.08 Basalt #3 2.76 5.41 169.40 3.20 3.91 0.16 Basalt #4 1.70 5.80 132.95 4.91 2.57 0.06 Columbia River 0.88 5.94 40.30 1.12 Basalt

To convert alphas per hour to alphas per og. per hour multiply by 0.0122. 36 per hour and the beta activity in counts per minute. The alpha activity of the sample was measured with a five inch scintillation type counter. The samples were alternated with dunite samples which were used as background. The alpha activity is an indirect measure of the total uranium plus thorium content of the rock. Though other alpha emitters exist in nature the overwhelming abundance of uranium and thorium in igneous rocks in comparison to other alpha emitters

justifies this assumption.

Since uranium and thorium are not contained in any essential minerals, one would expect their distribution to be rather erratic.

This difficulty is partially compensated for by using a much

larger surface area for alpha counting. An attempt to check for

the erratic distribution of the alpha activity was made by sampling

the top flow of the Tumamoc Hill series, both vertically and hori­

zontally. These samples were then subjected to alpha and beta

counting. The results of this analysis are shown in Table VI.

Table VI-Radiometric Variations within a Single Flow

Sample Betas per min Alphas per hour % potassium

Tumamoc Hill basalt Flow 5-1 3.87 266 2.87 Flow 5-2 3.91 280 2.73 Flow 5-3 4.23 208 3.02 Flow 5-4 4.26 271 2.66 Flow 5-5 4.00 273 2.53 Flow 5-6 4.59 245 2.78 Flow 5-7 4.00 282 2.59 Flow 5-8 4.08 Flow 5-9 4.15 Flow 5-10 4.02 37

The statistical counting error for the alpha activity was maintained at approximately plus or minus 6.5% during the counting

period. It is interesting to observe that the standard deviation

for this group of samples is also approximately 6.5%.

The beta activity and the potassium content of this series

of samples are also quite uniform. Since potassium is both the main

source of the beta activity and is distributed much more uniformly

in igneous rocks it would not be expected to vary greatly. The

standard deviations of both the beta activity and potassium content

fall within the statistical counting error. Thus, because the

total variation is no greater than the expected variation due to

counting statistics alone, it may be concluded that this "basalt"

flow is very uniform in composition. Sabels (1960) noted this

same uniformity for the Bonito and SP flows in the San Francisco

volcanic field of northern Arizona.

Fig. 3.2 is a plot of the percent potassium versus the alpha

activity in counts per hour for most of the samples studied in this

investigation. From the figure it can be seen that, in the case of

the basaltic rocks, as the percent potassium increases the alpha

activity of the sample also increases. Although this relationship

is not followed rigorously it can be used as a rough approximation.

While this is true for the "basaltic" rocks it can also be seen

that the more acidic rocks do not exhibit this relationship. How­

ever, the andesites all contain 3.9% or more potassium which may be 7 — 6 Percent Potaseium 5 — 5 — 1 2 ol — 0 ______Fig. 5 0 2Ptsimcnetv. lh ciiyfrsuhr rzn "basalts"•2southern Potassium content Arizona for activity vs. Alpha °0

B 0 20 0 4oo 500 200 100 0

1111

a 0 0 °* °000 o 0 00 T

0 0 A £ 0 A o o o ° i °A * V Alphas per hour Alphas 9 0 o % 0 0

58 0 °A 0 A J 0 600 500 ______A-Andeslte All others areothers All R-Rhyolite KeZ. T-Tuff BP-Basalt Porphyry BP-Basalt "basalts"

1 0 T 0 T

39 used as a criteria for classifying them.

The tuff samples are the most erratic in the relationship of their beta and alpha activity. This large variation can be explained by considering the general nature of the tuffs. The porosity and permeability of the tuffs is quite high. Therefore, both ground water and runoff surface water could interact with the tuffs and precipitate, by either chemical means or evaporation, salts of uranium, thorium and potassium. Likewise, ground water could also selectively leach out uranium, thorium or potassium, thus changing the relative beta and alpha activity of the sample.

The manner in which the tuffs were laid down could also have con­ siderable influence on the radiometric nature of the samples.

By looking at Fig. 3.2 and Table II it becomes obvious that the so-called basalts are very high in potassium. While a normal basalt contains between 0.83 and 1.25% potassium, the average value of "basalts" collected during this study is greater

than 2% potassium. Flame photometric studies, as mentioned in

the preceding section, have verified this fact.

The porphyries are also quite interesting in their radio- metric makeup. Although they have the same potassium content as

the average "basalt" studied, the alpha activity is approximately

twice as great. It is also significant to note that the porphyries are found in the same general area of the graph. The increased alpha activity of this group of rocks may be due to the process by 40 which they were formed. As pointed out in the section on general geology, Tolman believes that these porphyries were extruded after having crystallized in a magma chamber in which they were stored for a considerable period of time. If this were the case one would expect the alpha activity to be lower since slow crystallization is somewhat of a purification process itself. Conversely, however, if this represented a differentiation of an already basic magma the more acid type elements would tend to be concentrated in the last formed crystals. Since uranium and thorium are more concentrated in felsic rocks the porphyries may represent a differentiation of a basic magma.

Fig. 3.3 shows the distribution of potassium in relation to the alpha activity for the "basalts" only. From this plot it can be seen that there:is a linear relationship between the percent potassium and the alpha activity of each sample. Admittedly, some deviations from this relationship are quite obvious. Also, there is an indication of grouping of "basalts" from a given area. As an example, the Tumamoc Hill "basalts" have approximately the same potassium content as the samples from Black Mesa but the alpha

activity of the Tumamoc Hill sequence is about twice as high.

The San Bemadino samples have a much lower potassium

content and likewise a much lower alpha activity than either the

Tumamoc Hill or Black Mesa sequence. The Tombstone "basalts"

resemble the San Bernadino basalts and by counting analysis alone 4

1

2 5 5 5

5 5 , 5 5

I 6 2 m03

5 5

2 2

7 8 8 7 1 7

i 1- White Mtna. "basalt" 9 2- Black Mesa "basalt" 10 5-Tumamoc "basalt" 4- Sawtooth "basalt" 5- Tortilita "basalt" 6- Galiuro "basalt" 7- San Bernadlno basalt 8- Tombstone basalt 9- Coluabla River basalt 10-Carrizo basalt

0 o loo 200 J00

Alphas per hour

Pig* 5*5 Potassium content vs. Alpha activity for "basalts"

41 42 could probably not be separated from them. The White Mountains

"basalts", while having a very different potassium content, exhibit a fairly uniform alpha activity. The Sawtooth "basalts" seen to overlap the lower range of the Tumamoc sequence in that they have a somewhat lower potassium content but approximately the same alpha

activity.

From the plot it can be seen that a sample picked at random

may not give a good indication of the locality but a series of

samples would definitely show a trend. The alpha and beta activity

is also a good indicator of the chemical nature of the sample. The

samples with a high beta and alpha activity tend to be more felsic,

while those with a low alpha and beta activity tend to be more

basic. Fig. 3.1 shows the relationship between the percent potassium

and the calcium oxide content.

3.3-Location of Potassium in "Basalts"

Since the radiometric analysis of the "basalts" showed an

abnormal amount of potassium, one of the main problems was to find

the location and chemical form of the potassium in order to interpret

its origin and relationship to the "basalt." The origin presented

a special problem since many of the basic rocks in the Basin and

Range area are potash rich.

In order to ascertain whether the potassium in the samples

was in the form of readily soluble salts, basaltic rock samples. 43 high in potassium, were powdered and heated in distilled water.

The samples were heated in distilled water for twenty-four hours at temperatures between 80 and 100° F. At the conclusion of this period of time the suspension was allowed to settle and the amount of potassium in solution was determined by flame photometric methods. Flame photometric results indicated that less than one- tenth of one percent of the potassium in the rock was leached out.

Several samples of "basalt" were heated in boiling water for a half hour. Again the potassium content of the solution was found to be approximately one-tenth of one percent. The same solution contained less than five-hundredth of one percent sodium.

From the above experiments it can be seen that potassium and sodium in the basaltic rocks are not in the form of water soluble minerals. This eliminates the possibility that the

"basalts" could have been contaminated by highly alkaline surface waters depositing alkali salts. Probably the little potassium and sodium which was detected was due to reagent contamination.

Next, a potassic basaltic andesite was placed in a very weak acid solution of HCl for one week. The rock was counted both before and after the acid leaching to determine the amount of potassium present. The "basalt" showed no loss of potassium after the exposure to the acid solution. This indicates that potassium is in minerals which are not easily destroyed or damaged by acid leaching. 44

Attempts were also made to introduce potassium into a basalt undersaturated in potassium. In this experiment a near saturated solution of potassium was placed in a flask with a powdered sample of the Columbia River Basalt and shaken vigorously.

The solution was allowed to remain in contact with the sample for a week. At the end of this period of time the solution was filtered off and washed several times. The sample was then counted to deter­ mine the change in potassium content. The result showed that no potassium had been absorbed or adsorbed by the sample.

Five gram samples of the Columbia River Basalt were placed in five different solutions composed of (a) distilled water; (b) tap water; (c) sodium solution; (d) potassium solution, and (e) sodium plus potassium solution. At the end of a week samples of the resulting solutions were determined flame photometrically for potassium and sodium and compared with the original solutions.

None of the samples showed any gain or loss of either element during this period of contact. The results of the above two

experiments indicate that potassium cannot readily be introduced under conditions of normal temperature and pressure.

Since-vesicle fillings are quite common in the "basalts"

of this area the filling material was checked for the presence of potassium minerals. The filling material was separated from the

sample and found to dissolve quite readily and completely in HCl

in most cases. The resulting solution was analyzed for potassium 45 and none was found. The crystal shape and nature of effervescence

indicated that, the filling material was calcite.

In an effort to determine the location of the potassium

two experiments were conducted. In the first experiment basaltic

andesite, which contained phenocrysts of plagioclase and pyroxene,

was chosen to locate the potassium sites. The plagioclase and

pyroxene phenocrysts were separated manually from the rock and their

potassium content was determined. It was found that the plagio­

clase, which was 50ab-50an, contained 0.657. potassium and that the

pyroxene, augite, contained only 0.05% potassium. Since the whole

sample contained 2.94% potassium, and olivine and magnetite, both

non-potassium containing minerals, were the only other phenocrysts,

the immediate conclusion was that the potassium was contained in

the ground mass.

Staining techniques also verified that the potassium was

located in the groundmass. Samples of the basaltic andesites were

etched in hydrofluoric acid and sodium cobaltinitrate solution was

added to determine the location of the potassium. It was found

that the groundmass remained a yellow color due to the presence of

potassium after rinsing in water, while the plagioclase crystals

were unaffected. 46

The data below shows the potassium content of various feld­ spars and one.pyroxene from the collected basaltic rocks.

Table Vll-Potassium Content of Plagioclase and Pyroxene in Basaltic Rocks

Sample % Potassium

Tumamoc plagioclase 0.65

Black Mesa plagioclase 0.70

Sawtooth plagioclase 1.05

Benson plagioclase 0.90

Tumamoc pyroxene 0.05

The separated plagioclase was also beta counted to determine the potassium content and the results also indicated that the po­ tassium was located in the groundmass. Approximately one-fourth of the total beta activity of the sample was found in the plagio­ clase, but the resulting separated groundmass increased in beta activity.

Portions of the groundmass were separated from the samples and were x-rayed in search for potassium minerals. Thin section work indicated the groundmass was essentially the same composition as the phenocrysts with the exception of about 5-15% glassy material.

After the groundmass was separated from the sample, pyroxene and magnetite were magnetically removed and the resulting groundmass 47 was x-rayed. The x-ray analysis showed no visible orthoclase or quartz lines which would indicate a more felsic rock. Plagioclase lines were the only visible lines. Therefore, it appears that the bulk of the potassium in the "basalts" is present in the glassy groundmass. GEOCHEMICAL STUDY OF TRACE ELEMENTS

4.1-Emission Spectrographic Studies

In order to determine the trace element content and the variance in the trace element content in volcanic rocks, samples were studied with the emission spectrograph. This work was made possible through the use of the emission spectrograph located at

Arizona State University at Tempe. Samples were prepared by mixing known volumes of sample and carbon in approximately a one to one ratio. Several duplicate samples were run and the various settings on the spectrograph were determined by a series of runs on the same sample at the beginning of the spectrographic work. It was found that a pre-bum time of five seconds gave good duplicate readings on the same sample when followed by a 15 second b u m time.

The b u m time was chosen so that comparative readings of

the same sample were good, and also, to compromise on the intensity of the cyanogen lines. The five second preburn time might volatilize

the alkali metals and some of the more volatile elements, but none

of these elements were of particular concern in this study. The

slit width was set at 25 microns and the arc width chosen was two millimeters. Only first order lines were studied and the machine was set at the standard line setting of 610.

48 49

The elements which were considered fairly reliable for first order reading and were of interest included the following:

Cr, Mn, Ti, Zn and Cu. The lines were read on a spectrophotometer and were easily calibrated by shooting an iron nail with each group of four unknowns. The following lines were read for each of the previously mentioned elements:

Cr — ------4254

M n ------— 4030

T i ------3653

Z n ------3282

C u ------3247

Possibly other first order lines could have been read, but these lines were very intense and easily separated from the very abundant iron lines which are present in the basic volcanics.

Table VI31 gives the spectrographic data obtained on the volcanic rocks. The letters under each of the elements refers to

the percent opacity of the line measured. H, L, M refer to high,

low and medium opacity respectively and are used for comparative purposes for each element. A high reading indicates that relatively

large amounts of the element are present. Conversely a low reading

indicates that relatively little of the element is present. The

numbers present under each of the elements refer to the percent

opacity in absolute terms and are equal to 100 minus the transmittancy.

From this data it can be seen that Cr, Ti and Cu appear to be 50

Table VIII-Spectrographlc Data Obtained for Various Volcanics

Sample Cr Mn Ti Zn Cu

Tombstone basalt 98.5 H 99.5 H 89.1 H

Tombstone duplicate 98.1 H 99.4 H 89.6 H

Black Mesa basalt 99.7 H 94.8 H 99.4 H 84.0 H 99.5 H

Sasco basalt 99.6 H 93.3 H 99.3 H 57.8 M 99.2 H

Sawtooth basalt 99.5 H 98.0 H 99.5 H 76.0 H 99.8 H

Sawtooth andesite 54.6 L 77.0 M 99.5 H 63.3 H 98.6 H

Sawtooth andesite 56.0 L 73.5 M 99.2 H 64.0 H 98.1 H Duplicate - •

Picacho Peak 99.8 H 94.2 H 97.7 H 59.0 M 99.1 H Intrusive

Santa Rita basalt 100 H 94.5 H 99.2 H 75.2 H 99.7 H

Pajarita lava 80.4 M 87.4 M 99.3 H 74.4 H 98.1 H

Safford tuff 98.5 H 88.3 M 99.2 H 49.0 L 99.1 H

Tumamoc Hill basalt Flow 1 99.9 H 94.5 H 75.0 M 48.0 L 99.0 H

Flow 2 99.2 H 94.6 H 99.3 H 61.0 M 99.2 H

Flow 3 96.0 H 93.5 H 99.2 H 44.0 L 95.0 H

Flow 5 99.5 H 81.5 M 98.8 H 63.0 M 98.8 H

Tumamoc tuff 39.8 L 93.2 H 91.0 M 32.0 L 98.0 H

Tumamoc andesite 99.5 H 95.5 H 99.0 H 68.0 H 92.0 H

Atascosa volcanics Older basalts 99.3 H 89.2 M 99.2 H 63.3 H 99.2 H

Upper rhyolite 99.7 H 89.5 M 99.2 H 68.6 H 98.9 H

Upper basalt 100 H 98.0 H 99.3 H 83.0 H 99.6 H

W-l standard 99.2 H 99.4 H 99.2 H 80.0 H 98.8 H

G-l standard 89.2 H 99.4 H 96.0 H 37.0 L 69.5 M 51 distributed quite evenly within the basic rock types. Mn and Zn, however, appear to vary throughout the whole volcanic suite.

Upon considering the distribution of the elements studied, it is not surprising to find that Cr and Ti do not vary widely within the various basic rock types. Titanium itself does not vary widely within the igneous rock family, therefore it should not be expected to vary widely within the volcanic rocks. Chromium is relatively enriched in basic igneous rocks and it probably sub­ stitutes for either magnesium in olivine and pyroxene or the ferrous iron in magnetite. It is interesting to note the values obtained for Cr in the more acid rocks are lower, thus indicating a lower content. Copper varies to quite an extent within the various rock families but the rocks from this area did not appear to vary greatly in copper content. This may be due to the fact that this is a copper rich province.

Manganese is another element which is not considered to vary to any great extent among the various rock families. How­ ever, in this study it did appear to vary somewhat. This presents somewhat of a problem since manganese most often substitutes for iron which is abundant in basic rocks. A possibility exists that manganese may be present as desert varnish, though the samples were thoroughly cleaned. Thin layers of desert varnish not visible to the naked eye could play havoc with the spectrographie analysis.

If one considers only the basalt samples it can be seen that the 52 manganese is distributed very evenly and in only two cases, flow 5 at Tumamoc Hill and the older basalts in the Atascosas, do the samples give a somewhat lower reading. Manganese, like chromium, gives correspondingly lower readings for the more acid rocks.

Zinc is the only element studied which really varied

erratically throughout the whole sequence. Zinc would most likely

substitute for either ferrous iron or magnesium. If this was the

case the most basic rocks should contain the most zinc. There is

no correlation between the relative basicity of the rock and the

zinc content. The distribution of the zinc in the volcanics remains

a problem.

When one considers all the elements studied with the emis­

sion spectrograph it can be shown that, with the exception of zinc,

the elements are distributed fairly evenly throughout the volcanic

suite. This could indicate that the volcanic suite of rocks were

formed from the same magma, but one must also consider that these

elements do not vary radically within the basic extrusive rock

complex.

4.2-Neutron Activation Studies

Neutron activation studies of various "basalt" samples

were made in an effort to learn more about their chemical makeup

and also to try and find a correlative property through the result­

ing spectra. The technique used in this investigation is described 53 in Appendix C. The only nuclear reaction observed or detected during this study was the neutron-gamma reaction in which an un­

stable isotope containing one more neutron than the parent isotope

is formed. These unstable isotopes usually decay by beta minus

and gamma radiation to a stable isotope of the next higher atomic

number.

Radioisotopes which were detected in the basalt samples

after the activation period included K ^ , N a ^ y Cu*’"’, Sc^, Fe*^ 51 and Cr . Since the half-life of the sodium, potassium and copper

isotopes are less than.one day no further work was done with these

isotopes other than their qualitative identification. Scandium,

iron and chromium have much longer half-lives and these isotopes

were further investigated. After these isotopes were identified,

both energy-wise and through their respective half-lives, a program

was set up so that a comparison could be made of the various

samples.

Since all the samples contained each of the various radio­

isotopes and none of the samples contained other interfering iso­

topes, the method of comparing the samples was through the relative

values of each of the gamma peaks. It was decided to read the

spectra of each set of samples exactly one month after activation.

This decay time period allowed the shorter lived isotopes to decay

away and reduced the Compton spread effect on the longer lived

isotopes. 54

One sample was chosen as a reference standard and was present in each set of activated samples. The amount of each isotope present was determined by plotting energy of the radio­ active particles versus the count rate. The background and Compton effect were subtracted from the count rate of each isotope.

Fig. 4.1 shows the relative values for some of the samples which were activated. The various curves labeled Sc, Fe and Cr represent relative amounts of each of the elements present. The scandium and iron curves resemble each other very closely in shape while the chromium content of each of the samples is fairly constant. The chromium curve resembles the emission spectrographic data in the relatively uniform distribution of the element.

Scandium is probably substituting for either magnesium or ferrous iron. Therefore, it would be present in olivine, pyroxene and magnetite. According to Goldschmidt (1954) olivine contains seven parts per million SCgOg, pyroxene 100 parts per million and basic plagioclase zero parts per million. Therefore, probably most of the scandium is concentrated in the pyroxene.

The iron which is present in the major minerals such as magnetite, pyroxene and olivine would be expected to have the same general shape as the scandium curve because the scandium substitutes

for ferrous iron. The iron curve shows the general acid to basic

increase in the extrusive samples.

Chromium, which would most likely substitute for ferric SCANDIUM 400

0 0

O o 0 200

IRON

100

0 0 0 0

CHROMIUM 200

0 0 0 0 0 0 0 0 0

100 0

Sample number

Pig. 4.1 Neutron activation curves fo r southern Arizona "basalts"

55 56 iron in magnetite or exists as a separate chromite phase in the basic extrusives, does not follow the general pattern of iron or scandium. It remains relatively constant throughout the whole suite of volcanic rocks. This same conclusion was found in the emission spectrographic work.

The results gained by this study proved very interesting and neutron activation methods of analysis certainly shows good indications of being a valuable tool for future investigations of extrusive rocks. Its relative simplicity, great inherent sensitivity, and relative ease of determination and identification renders it an ideal method of chemical analysis for geological and geochemical studies. CONCLUSIONS

The potassic basaltic andesites in this section of southern

Arizona appear to cap the igneous sequence in all localities

studied. The basaltic andesites overlie either andesites or sedi­ ments in all cases where the underlying rocks can be observed. The

first basic flow extruded after the underlying andesites is very

porphyritic and contains large phenocrysts of plagioclase and augite.

This unit has been called the "Turkey Track Porphyry" by various

people though both mineralogically and chemically it is a potassic

basaltic andesite which is slightly more felsic than the overlying

flows. This porphyry unit possibly represents a longer period of

crystallization in the magma chamber before extrusion. The strati­

graphic position, uniform mineralogical composition, and the chemical

composition points towards a common magma chamber for the entire

unit. Undoubtedly these do not represent the same flow but probably

represents a series of fissure flows which were pene-contemporaneous.

The flows above this porphyry unit appear to be much more

basic in composition. Handspecimen or petrographic identification

of these flows indicate that they are probably olivine basalts.

Olivine, pyroxene, magnetite and plagioclase are the only identifiable

57 58 minerals except in the Sawtooth sequence where biotite is also present. Secondary quartz and calcite are occasionally present in vesicle fillings.

The age of these flows is probably Plio-Pleistocene when all the evidence is considered. The fact that the basic porphyry unit is bedded within the Rillito beds in certain localities makes

this unit at most Miocene in age since portions of the Rillito beds have been tentatively assigned to this period. This is based upon

the identification of a lower Miocene rhino in the Mineta beds which are believed to be time equivalent to the Rillito beds.

There is a thick sequence above the rhino remains. Therefore, the

porphyry is probably post lower Miocene and possibly post Miocene

in age. Voelger (1953) notes the presence of porphyry pebbles in

the middle Rillito beds in the Bellota Ranch area. The presence of

cindery material, the geomorphology, lack of deformation and tilt­

ing, and the presence in some cases above unconsolidated alluvium

indicate that the uppermost basaltic andesites are Pleistocene in

age. It is possible that the entire porphyry-basaltic andesite

sequence is Plio-Pleistocene in age.

Alpha and beta counting have indicated that the potassium

content of these flows is much too high for a normal basalt. The

potassium content is also inversely proportional to the calcium

oxide content. The alpha activity of each of the various potassic

basaltic andesites seems to be related directly to the amount of 59 potassium present. This relationship is not present in the more acidic tuffs or andesites. The alpha and beta activity of the porphyry unit is.very uniform. Alpha and beta counting studies have also shown the potassium content of the andesites to be much greater than the potassium content of the basaltic andesites.

Chemical work indicates that in both the porphyry and overlying flows the potassium is concentrated in the groundoass.

Counting results have verified this. It appears that the glassy material within the groundmass is the chief source of the potass­

ium. Chemical analysis and stain tests of the feldspar indicate

that it is not the chief source of the potassium. In fact, the

feldspar is deficient in potassium when compared with the rest of

the rock.

The chemical analysis conducted during this study indicates

that the basic flows are not as basic as was formerly thought. In

fact, the composition of these flows indicate that they are nearer

to a diorite or basaltic andesite than to a plateau or olivine

basalt. This discrepancy between petrographic and chemical

analysis reflects the fact that these rocks were not completely

crystallized before extrusion. The glassy groundmass which makes

up 5-15% and as high as 30% of the total rock appears to be the

main source of the potassium and is the chief indicator of this

incomplete crystallization and reaction. Had the glassy material

been allowed to crystallize, othoclase or sanidine, hornblende and 60 possibly quartz could very well have developed leading the petro- grapher to classify these rocks as andesites or more felsic type rocks.

The results obtained by the major chemical analysis point out the fact that a complete chemical as well as a petrographies1 study should be made of extrusive rocks before classifying them.

Although petrographic work is more valuable for rocks with a holo- crystalline texture, petrographic methods alone should not be applied to the classification of extrusive rocks.

The emission spectrographic results indicated that the extrusives in this section of Arizona were quite uniform in Ti,

Cu and Cr content and to a certain extent in the Mn content. The amount of zinc in each of the various extrusives seemed to vary rather erratically without any apparent reason.

Neutron activation analysis indicated that scandium followed

the iron content in most of the samples. The chromium content which did not vary greatly according to emission spectrographic

data was found to be fairly uniform in the neutron activation

analysis.

Probably the chief conclusion that can be drawn in this

section is the fact that a geochemical investigation, coupled with

a geologic study, can lead to many valuable considerations which

would not come about by a chemical or geological study alone. APPENDIX A

81 A SEMI-MICRO ANALYSIS OF SILICATE ROCKS FOR

Ca. Mg, Fe, AND A1 EMPLOYING E. D. T. A.

By

Carroll Halva

Geochemistry Section of the Geochronology Laboratories

University of Arizona

A rapid, accurate analysis of total iron, aluminum, calcium and magnesium in silicate rocks using EDTA as the analytical reagent has been developed using sem i-micro techniques. This analysis will be a great aid to the geologist and geochemist since it is time-saving, accurate, and gives approximately 80% of the composition of the average igneous rock other than SiOg. Another advantage of this method is the fact that a relatively small sample (0. 2 g) is needed in order to make this determination. A sample can be fused, diluted to volume and analyzed for these elements in approximately 3 hours. If several samples are fused at the same time, as many as eight determinations can be made in a eight hour day.

INTRODUCTION

Since the development of Ethylene-Diamine-Tetra-Acetic acid (EDTA) as an analytical reagent by Schwarzenbach (1) in 1945, its applications and uses have spread into many fields of research. The determination of over 40 cations and numerous anions employing EDTA has been reported up to the present time, and separations based on the EDTA complex formation have become very important in the determination of various other ions. EDTA has been used extensively as an analytical reagent in the fields of chemistry, biology, agriculture and, to a certain extent, in geology. Materials analyzed by EDTA range from milk (2), through fertilizers (3), to sedimentary rocks (4) and includes many substances.

Calcium and magnesium in igneous rocks have been determined by various investigators using EDTA (5) (6) (7). However the author could find no infor­ mation in the literature on the determination of iron and aluminum in igneous rocks other than ores, and developed the method discussed in this paper for the determination of Ca, Mg, Al, and Fe in igneous rocks employing EDTA as the analytical tool following methods or modifications of methods already in the literature for the determinations of these ions in other substances. The follow­ ing report discusses the method developed by the author, and includes, in order, a section on the preparation of the sample for analysis, analysis of the sample solution, and results of the analysis of various known rocks.

PREPARATION OF SAMPLES

Samples of silicate rocks were crushed and powdered and approximately 0. 2 g of the powdered sample was weighed out into a platinum crucible using a Beckman chainomatic balance. 1.5 g of sodium carbonate was added to the crucible, thor­ oughly mixed with the igneous rock sample, and a sm all amount of sodium car­ bonate was placed upon this mixture. The crucible was covered and the mixture was slowly heated over a meeker burner until it became red hot. The heating was continued until the molten mixture became quiet and had ceased to evolve carbon dioxide. Once the molten mass had become quiet, the heat was shut off and the mixture was swirled in the crucible so that a portion of the hot mass adherred to the side of the crucible thus enabling the fused mass to dissolve faster, in the following step, because of the larger surface area. The fusion process takes 10 to 30 minutes.

The cooled crucible was placed in dilute (2:5) hydrochloric acid and heated slowly until the fused mixture dissolved. A clear solution was seldom obtained since silicic acid separated from the solution almost immediately. This did not

Contribution No. 29 of the Program in Geochronology, University of Arizona.

61 62

82 interfere with following steps and an analysis of the silicic acid which precipi­ tated from the solution showed no appreciable coprecipitation of Ca, Mg, Fe and Al. After the melt had dissolved, the solution was cooled and the platium cru­ cible was rinsed several tim es. The resulting solution was added to a volumetric flask and diluted to 200 m l.

ANALYSIS OF SOLUTION

The EDTA analysis of the final solution was accomplished by carefully con­ trolling the pH and by the utilization of masking agents to eliminate interference from interfering substances. Various size aliquots of the final solution, from one to ten ml. , were used in the analysis. The aliquot size was determined by the various concentrations of the elements present.

The solution was analyzed for calcium, magnesiu, iron and aluminum using methods similar to those in the literature. Those unfamiliar with EDTA, its nature and uses, and the principals involved in EDTA titrations, should refer to Welcher (8) or other books on the subject. Total Ca-Mg was determined at a Ph of approximately 10 using an ammonium hydroxide-ammonium chloride buffer and Erichrome Black T as the indicator. Calcium was determined at a pH of 12. 5 using potassium hydroxide as a high pH buffer and Calver II or a Murexide-Napthol Green B mixture as the indicator. Total iron was determined at a pH of approxi­ mately 2.4 using salicylic acid as the indicator. Total iron plus aluminum was determined by back titration of a hot 40% acetone solution of dithizone as the in­ dicator at a pH of 4. 5. The following flow sheet will serve as a guide to the EDTA determination of Ca, Mg, Fe, and Al in silicate rocks. The determination of each of the elements is discussed in greater detail in the following paragraphs.

0.200g

dissolve sample dilute to volume pH s c a le store in plastic bottle

12 10 4 .5 2 .4 C alciu m Ca-Mg Fe-Al Iron KOH bu ffer pH 10 buffer Na Ac buffer HCL-NaAc buffer C a lv er H Erichrome Black Dithizone Salicylic acid o r T M u rexid e

Calcium was determined using an 8N solution of KOH as a high pH buffer. The KOH eliminated interference of magnesium by precipitating it as the hydro­ xide, and from aluminum by converting it to the aluminate ion which does not in­ terfere with the EDTA titration. Potassium cyanide was added to remove inter­ ference from any heavy metals which might be present by tying up the metal as the cyanide. If the rock contained large amounts of iron, triethanolamine was added to complex the iron. Sodium tartrate was added to the solution before the high pH buffer in order to prevent coprecipitation of the calcium with the magnesium hydroxide. Hydroxylamine was added to prevent oxidation of the indicator, and if small amounts of ferric iron are present they will be reduced to the ferrous stage which will be masked by the cyanide.

The order of addition of reagents is very important in the calcium determina­ tion. Sodium tartrate must be added before the solution is made basic to prevent coprecipitation of the calcium. Hydroxyamine must be added before the potassium cyanide so that small amounts of ferric iron can be converted to ferrous iron, and obviously it must be added before the indicator to prevent oxidation of the indicator. Potassium cyanide must be added after the solution is made basic in order to pre­ vent liberation of cyanide gas which is given off when cyanides are placed in acid solutions. The indicator is always added after all the masking agents have been added and after the pH adjustment has been made. 63

83 Three different indicators were tried and the sharpest endpoints and most accurate results were obtained using Calver H (Hack Chemical Company) and a Murexide-Napthol Green B mixture (9). Murexide was tried but the end­ point was not ideal and much harder to detect than those of either of the other two indicators. Calcium forms purple complexes with these indicators and when EDTA ties up the calcium the indicators turn blue.

Total calcium plus magnesium was determined at a pH of approximately 10 using an ammonium hydroxide-ammonium chloride buffer. Interference of the heavy metals plus iron and aluminum was eliminated by potassium cyanide and triethanolamine. Triethanolamine will tie up the aluminum ion in ammonical solution by the formation of an aluminum-ammonium-triethanolanrune complex. If ferric iron is present, it will interfere by oxidizing the indicator. However, this can be overcome by the addition of hydroxylamine which reduces the iron and prevents oxidation of the indicator.

The order of adding reagents is not quite as important in the total calcium- magnesium determination as in the calcium determination but two things must be remembered. First, make sure the solution is basic before adding the cyanide, and secondly, do not add the indicator until all the adjustments on the solution have been made.

Two indicators were tried in this determination. Erichrome Black T was preferred since, in general, it appeared to be much more sensitive to small quanti­ ties of magnesium than the second indicator, Erichrome Blue Black R (as Calcon). The endpoints with both indicators were very sharp, changing from red to blue as the calcium and magnesium is tied up by the EDTA. The indicators were used in solid form in approximately 1:50 ratio using ammonium chloride for a filler as suggested by Hammon and Neuman (10). The ammonium chloride serves as a carrier and a source of the ammonium ion for the buffer. The number of mis. of EDTA required for titration of calcium was subtracted from the number of mis. required for total calcium plus magnesium. This difference represented the number of m is. required for the magnesium titration.

Total iron was determined at a pH of 2. 4-2. 6 using salicylic acid as the indicator. Amonium persulfate was added to the solution to oxidize any ferrous iron to ferric iron. The pH was adjusted to approximately 2. 5 using sodium acetate and hydrochloric acid. The endpoint was very sharp with the indicator changing from a deep violet to a yellow or colorless solution upon titration. If the concentrations of iron were high the solution usually changed to a yellow color. Thiocyanate was also tried as an indicator and performed equally well at a lower pH. Reagent grade indicators were used in the solid form.

No attempt was made to calculate ferrous and ferric iron separately although this is entirely possible providing none of the ferrous iron is oxidized daring the fusion. Kolthoff and Sandell (11) have described a carefully controlled sodium carbonate fusion whereby ferrous iron is not oxidized to ferric iron. Ferric iron in the presence of ferrous iron can be titrated under CO2 at a pH of 2. 4-2. 6 with EDTA by using salicylic acid as the indicator. Total iron can then be deter­ mined by adding ammonium persulfate which oxidizes the ferrous iron to ferric iron. Total iron is then determined and ferrous iron is found by difference. Since this determination is carried out at a low pH most divalent cations do not interfere (12). The author has not tried this method but recognizes that it would frequently be very desireable.

Total iron plus aluminum was determined by a modification of a back titra­ tion method originally proposed by Vanninen and Ringbom (13) in which a known volume of EDTA is added to a known volume of unknown solution and the excess EDTA is back titrated with zinc sulfate using diphenyl-thio-carbonozone as the indicator. In their original work the EDTA-unknown solution was made 40-50% alcoholic to serve as a solvent for the indicator. This process was modified by using acetone in place of alcohol as proposed by Banks and Bisque (14) and back titrating with zinc chloride rather than zinc sulfate. The indicator (Eastman Kodak Co. ) was prepared by dissolving 0. 75 g of the indicator in 200 ml of rea- gent grade alcohol rather than absolute alcohol as proposed in the original work.

\ 64

84 It is very important that the indicator be made up every few days.

The procedure of analysis varies from the original work (13) in that total iron and aluminum is determined by oxidizing all the iron to the ferric state (if not already in that form) and then heating the solution of EDTA and unknown for three to five minutes on a hot plate. Acetone, which has also been warmed, is then added until the solution is between 40-60% acetone. The pH is adjusted to between 4. 4-4. 6 using the hot sodium acetate and hydrochloric acid. The pH is kept in this range throughout the titration. After adjusting the pH, the indicator is added and the solution is titrated with zinc chloride from a greenish-blue to a bright pink color. The solution is heated because, in cold solutions, aluminum and EDTA complex quite slowly but, if the solution is hot, the complex forms rap id ly.

The number of m is. of zinc solution required for titration was recorded and the equivalents of EDTA used up by the unknown solution was calculated. The amount of EDTA required for the total iron determination was subtracted from this and the number of equivalents required for the aluminum titration was determined.

EDTA solutions were made up using the disodium salt of ethylene-diamine- tetra-acetic acid and standardized against a carefully prepared 0.01 M calcium chloride solution. Approximately 1 gram of magnesium chloride 6-hydrate was added for each 40 grams of the disodium salt. The magnesium salt was added to sharpen the Erichrome Black T endpoint in case only very small amounts of magnesium were present. This was added before the solution was standardized. 0. 01 and 0. 001 M solutions were made up and used in the titrations.

RESULTS OF THE ANALYSIS

Four samples from the National Bureau of Standards were selected for an­ alysis. Three of them; the Chelmsford Granite, the Graniteville Granite, and the Columbia River Basalt, we re selected as typical igneous silicate rocks and the fourth, a standard dolomite was run because it is a rock in which the calcium and magnesium predominated over the iron and aluminum. Table 1 shows the CaO results obtained by EDTA compared with the Bureau of Standards analysis. Table II shows the same for MgO, and tables III and IV illustrate this for AKO3 and Fe respectively. The EDTA results are an average of several runs. The per cent difference is calculated; it is assumed that the Bureau of Standards analysis is the correct analysis.

From the results of the analyses it can easily be seen that the EDTA deter­ mination in no case seriously disagreed with the Bureau of Standards analysis. The per cent difference in no case exceeded 5% and in only three cases did it exceed 3%. This, in itself, is very remarkable considering the small sample used and the relative speed of the determination in comparison to standard meth­ ods of determination. However it must be added, in all fairness, that the deter­ minations were very closely watched and the attention given to these samples from the Bureau of Standards would probably be slightly above that which would be given most rocks during routine laboratory analysis. Work with standard solutions, made up by dissolving the salts of Ca, Mg, Fe and A1 in water or acid and then determining the actual content with the speed and technique one normally employs in the laboratory, showed that percentage of error was about the same as in the carefully controlled fusions of the Bureau of Standards samples. Attempts to correlate beds of volcanic rock, on the basis of chemical content, in which only normal attention was given to the sample, showed that the actual variations did not exceed that expected for normal variations of the analyzed com­ ponent within volcanic flows.

The limit of sensitivity for detecting the various ions varies. Ca and Mg are very easy to detect in small amounts with a high degree of accuracy. Iron was also quite easy to detect in small amounts with a high degree of accuracy. However, in the case of the Standard Dolomite, the aliquot size had to be in­ creased and the original sample size also was larger. Aluminum was much harder to detect in small amounts. The lower limits of detection, for which the high degree of accuracy discussed in the foregoing paragraphs may be 65

85 expected, are about 0.1-0. 2% sample weight for Calcium and Magnesium and about 0. 4% for iron and aluminum providing one follows the scheme outlined in this paper and employs normal laboratory technique. However, the lim its of detection may be lowered further by increasing the size of the sample or the size of the aliquot.

T ab le I CaO % S a m p l e ______B u reau of Stan d ard s_____EDTA______% Difference

Graniteville Granite 0. 65 0. 67 3.1 Chelmsford Granite 0. 76 0 .7 7 1. 3 Columbia River Basalt 9 .0 3 9. 01 0. 2 Standard Dolomite 30. 49 30. 52 0.1

T ab le II MgO % Graniteville Granite 0. 08 N . D. Chelmsford Granite 0. 24 0. 25 4 .0 Columbia River Basalt 4. 76 4. 81 1.1 Standard Dolomite 21. 48 21. 65 0. 8

T ab le HI A1i°3 Graniteville Granite 12. 30 1 2 .2 8 0. 2 Chelmsford Granite 13. 59 13. 66 0. 5 Columbia River Basalt 13. 04 1 3 .1 0 0. 5 Standard Dolomite . 067 N.D.

T ab le IV F e % Graniteville Granite 1. 22 1.16 4 .9 Chelmsford Granite 0. 81 0. 82 1. 2 Columbia River Basalt 11 08 11 26 1. 6 Standard Dolomite 0. 058 0. 057 1. 7

Modifications in methods of fusion have been studied by the author and two significant conclusions resulted. The first of these is the fact that, for most igneous rocks with normal iron content, the method of fusion described in this paper was the most practical if the determinations were sought, fusions with sodium hydroxide in nickel crucibles were faster and just as accurate. In the case of rocks, which are high in iron, it was found that the sodium carbonate fusion was not completely successful because a certain amount of iron alloyed to the platinum crucible. The only alternative, in this case, was to dissolve the sample in hydrofluoric acid, which, of course, increased the time necessary for analysis.

As a general conclusion, it may be stated that the method of analysis out­ lined in this paper is, for most geological and geochemical work, faster and as accurate as most other published methods.

ACKNOWLEDGEMENTS

The work herein reported was supported in part by the Atomic Energy 66

86 Commission under contract AT(11-1)-689-Correlation and Chronology of Ore Deposits and Volcanic Rocks. I would like to thank the members of the Geo­ chemical Section of the Geochronology Laboratories of the University of Arizona for their helpful advice. I am especially grateful to Mr. Carl E. Hedge for his great interest in the project and for aid in the development of the fusion tech­ nique.

A special note of thanks goes to Prof. Paul E. Damon, who heads the Geo­ chemistry Laboratory, for his encouragement of this work. Prof. Damon and Prof. J. W. Anthony kindly read this report. Also, the author would like to thank Dr. Raymond Bisque for first introducing him to many of the techniques involved in the EDTA titrations.

REFERENCES

1. Schwarzenbach, G ., Schweiz. Chem.-Ztg. u. Tech. Ind. 28, 1945.

2. Jenness, R., Anal. Chem. 25, p. 966-68, 1953.

3. Berkhout, H. W. and Goosens, N. , Chem. Weekblad, (48) p. 580-82, 1952.

4. Boardman, D. C. , Proc. Iowa Acad. Sci. 60, p. 330-32, 1953.

5. Banks, J. , Analyst 77, p. 484-89, 1952.

6. Shapiro, L ., and Brannock, W. W., United States Geological Survey Circ. N o. 165, 17 p, 1952.

7. Corey, R. B ., and Jackson, M. L. , Anal. Chem. 25, p. 624-28, 1953.

8. Welcher, F. J., The Analytical Uses of Ethyl*nediamine Tetraacetic Acid, D. Van Nostrand Company Inc 366 pp 1958.

9. Knight, A. G. , Chemistry & Industry, p. 1141, 1951.

10. Neuman, F ., Das Papier 6, p. 519-22, 1952.

11. Kolthoff, I. M ., and Sandell, Textbook of Quantitative Inorganic Analysis, MacMillan Co. , 758 pp, 1956.

12. Hoshikawa, G ., Japan Analyst 4, p. 582-83, 1955.

13. Vanninen, E ., and Ring bom, A., Anal. Chem. Acta. 12, p. 308, 1955.

14. Bisque, R. E ., and Banks, C. V., Unpublished Work on the Analytical Uses of EDTA, Iowa State College, 1959. a p p e n d i x b 75 RADIOMETRIC DETERMINATION OF POTASSIUM

IN SILICATES

By

Paul E. Damon, Carl E. Hedge, Omer J. Taylor and Carroll Halva

Geochemistry Section

of

The Geochronology Laboratories

University of Arizona

INTRODUCTION

Recent developments in the dating of rocks by the potassium-argon method have resulted in a renewed interest in analytical techniques for the determina­ tion of ootassium in silicates. Replicate potassium-argon analyses yield A ^/K ratios with an average difference of about ± 3% (Baadsgaard, et. a l., 1957; Long and Kulp, in press). Surprisingly enough the deviations are about equal for both the potassium analyses in which potassium is present in amounts usually from 5% to 10%, and for the argon analyses in which argon is present in amounts of a few p. p. m. or much less.

Winchester (1959) using a radioactive potassium tracer has found potassium yields during hydrofluoric-sulfuric acid digestion of silicate samples to be as lo w a s 88% recovery. He also determined the potassium content of the standard M.I. T. biotite by neutron activation. The neutron activation results, which required no chemical treatment of the sample, were about 4% higher than the average of chemical determinations.

Experiments in this laboratory (Hedge, thesis research in progress) have confirmed Winchester's observation of occassionaly low potassium yields during silicate digestion. Furthermore, the low yields were found by Hedge to be the result of potassium loss, possibly by the physical process of transport on vapor droplets on a microscopic or submicroscopic scale. Russian workers (Levitskii, 1953) have found both potassium and sodium loss during acid digestion, and they have investigated the effect of the presence of different anions in solution on the extent of loss. The effect seems to be greater for sodium than for potassium. They conclude that the process resulting in loss is of a physical rather than ? strictly chemical nature. A double crucible technique is now used in this labora­ tory to condense volatiles on an outer cooler surface, thus entrapping any es­ caping potassium.

Because of the difficulties encountered with chemical techniques for the analy­ sis of potassium and the need for precise data, it may be of particular interest at this time to describe a radiometric method for potassium analysis which re­ quires no chemical preparation. The method to be described was developed at Columbia University (Damon and Kulp, 1958, p. 435) and further tested in this laboratory. However, the technique has not heretofore been adequately described.

Radiometric Methods for Potassium Determination

The reader may refer elsewhere (Strominger, et. a l., 1958) for the details of the decay scheme of radioactive potassium-40. Natural potassium contains about 119 micrograms of the radioactive isotope per gram of potassium. The radioactive isotope, potassium-40, decays with a half life of 1. 3x10^ years by both competitive beta em ission and electron capture. The electron capture is followed by deexcitation resulting in the em ission of a gamma ray of about 1. 46 mev. energy. Out of every 100 transmuting K40 atoms, approximately 11 decay by electron capture to A4U followed promptly by gamma em ission and approximately

Contribution No. 28 of tbe Program in Geochronology, University of Arizona. 67 68

76 89 decay by 1. 33 mev. beta emission to Ca^. Thus, for a pure potassium salt, measurement of either the beta ray activity or the gamma ray activity can be used to determine the potassium content. Gaudin and Panell (1948) used beta counting to determine potassium in sylvite. Damon (1950) has used the gamma transition for logging well cores. Wendt (1955) made use of the radiometric method for determining potassium underground in salt deposits. The radiometric method has also been applied to the rapid analysis of potassium in solution (Dresia and Beckmann, 1957).

The radiometric measurement of potassium in silicates is complicated by simultaneous beta and gamma em ission from the uranium and thorium radio­ active series. Serdyukova and Kapitanov (1958) overcame this difficulty by simultaneous radiometric determination of uranium, thorium, radium and po­ tassium. This required the measurement of four parameters: (1) total alpha activity, (2) total beta activity, (3) total gamma activity, (4) selected pulse height analysis of a predetermined portion of the gamma ray spectrum. A set of four simultaneous equations were then solved for the four components. The method is complex, the various term s in the equation have both very large and very small coefficients and, as yet, this method has not been subjected to con­ vincingly critical checks.

The method herein described is simpler and, as will be shown, it has been thoroughly checked against independent chemical analysis. The method involves alpha and beta counting of a powdered sample. The beta count may be taken with or without a 5 mil aluminum absorber to eliminate the soft beta particles from rubidium-87 (Rb® ' interference is usually negligible). The beta count is then reduced by an empirical correction for the betas resulting from the uranium and thorium series. The alpha count, which is due only to the uranium and thorium series, determines the magnitude of this correction. The resulting beta count is compared with a standard sample of potassium to determine the potassium con ten t.

Experimental Techniques for alpha-beta Method

The sample is washed in distilled water to eliminate surface contamination and then ground until the entire sample can be passed through an 80 mesh screen. It was found by trial and error that a sample ground to this fineness can be re- producibly packed to a high bulk density. Coarser samples can be used providing that they are carefully and tightly packed. An infinitely thick source is used to eliminate the dependence of alpha and beta count on sample packing and to give the highest possible count.

For beta counting, the sample is packed into a cylindrical planchette of inner dimensions 1.91 cm. in diameter and 0.47 cm. in\depth. The sample thickness for silicates is approximately 650 m g./cm . The sample is then counted in a Volchok-Kulp (1955) low level beta counter with anti-coincidence ring and lead shielding. An Anton model 1007-T pancake geiger counter is used. The counting window is 2. 86 cm. in diameter and 1.4 to 2 mg. / cm^ thick. The background of the counter averages 1.1 c .p .m .

Potassium dichromate was chosen as a counting standard (Bureau of Stan­ dards chemical standard 136-A) because its density (2.69 g/cm^) is about the same as the silicate unknowns and also because it is not deliquescent. The standard count is about 23 c.p.m . above background count shielded (5 mil Al), and about 30 c.p.m . unshielded. The standard contains 26. 57% potassium. Thus 1% potassium is sufficient to approximately double the count over background count. This high sensitivity allows potassium to be determined in amounts down to 1 part per thousand.

An alpha counter modified after the type used by Kulp, Holland and Volchok (1952) is used. The counting tube is a 5 inch Dumont 6364 photomultiplier with a thin, continuous layer of Dupont No. 1101 phosphor on the face of the tube. Any other ZnS phosphor activated with silver will do. The background of the counter is about 35 c.p.h. A planchette 12.3 cm. in diameter and 2.4 mm. deep is 69

77 employed. A sample containing 1 p. p. m. uranium will approximately double the count over background count. The planchette is filled with dunite to obtain the background count both for beta and alpha counting. Dunite has about the lowest radioactivity of any natural material.

The equation for the potassium content of the unknown is as follows:

Kx = Nx Kg - CNa ■*7 where K is the potassium content of the unknown and Kg is the potassium content of tlie standard; Nx, Ns and Na are the beta count due to the unknown and standard and the alpha count respectively; C is the empirical correction factor for uranium and thorium series beta activity. It is 0. 005% K per alpha per hour for unshielded samples and 0. 004% for shielded samples. The uncertainty in C is approximately 4- 0.0005. Ng must be determined separately for both shielded and unshielded samples. The correction for igneous rocks averages about 25% of the total equivalent potassium content. However, the correction for clean mineral separates is much less; for example, 10.97% for the M. I. T. biotite standard (B3203). This is a result of the well known fact that a large part of the uranium and thor­ ium in igneous rocks is interstitially located, whereas separated and cleaned minerals have a much lower uranium and thorium content.

The effect of the presence of rubidium in the unshielded unknown is about the same as for potassium, i. e ., 1% Rb is approximately equivalent in beta count to 1% K. The soft rubidium betas are completely eliminated by the 5 mil alum­ inum absorber.

RESULTS AND DISCUSSION

The alpha-beta radiometric method has been checked against various gravi­ metric methods as well as against flame photometry and the stable isotope dilu­ tion technique. The data for National Bureau of Standards Rock Standards and a Pb-Ba standard are shown in Table 1. Table 2 presents the results of a compari­ son of radiometric analyses and other methods for a number of silicate rocks and m in e r a ls .

The average difference of the radiometric values from other methods for samples containing greater than 1. 5% potassium is approximately 2.7%. This is about equal to the average standard deviation computed from statistical con­ siderations alone. On the other hand the error rises rapidly for samples con­ taining less than 1. 5% potassium. This is in part caused by counting errors but must also be due in large measure to the decreased accuracy of the comparison methods as well.

Taking the data as a whole, there is no evidence of a systematic difference between the different methods. However, as mentioned previously, in individual cases systematic errors do appear to be present. For example, the M. I. T. standard biotite has been analyzed six tim es in this laboratory by the flame photo­ metric method using a lithium internal standard. In the first three analyses, the digestion was carried out in a covered platinum crucible using hydrofluoric and perchloric acids and a temperature of approximately 100°C. The vapor from these digestions was passed through a polyethylene funnel and tubing into a water trap and a small, but measurable, amount of potassium was found to have collected in the cool water. The experiment was repeated without sample for a blank check and in this case, no potassium was detected. Three more replicate M. I. T. stan­ dard biotite samples were run placing the platinum crucible inside a larger, cov­ ered nickel crucible to condense the less volatile vapors. The lids of the nickel crucibles were sufficiently good radiators to keep them at a temperature below 45°C. Salts containing potassium collected on the under side of the lids of the nickel crucibles. However, potassium could no longer be detected in the water trap. The salts from the lids of the nickel crucibles were returned to the original solutions and, in this manner, three values averaging 7.80% potassium were ob­ tained. The average of the three single crucible results was 7.40%. 70

78 The alpha-beta radiometric method for potassium determination is simple and requires a minimum of labor when compared with chemical methods. The most laborious step is the sample preparation, whereas the packing of samples in the planchettes requires the most care. The method has the added advantage that it requires no wet chemistry and does not destroy the sample. One dis­ advantage is that the correction factor C assumes radioactive equilibrium within the uranium and thorium series and normal ratios of uranium and thorium (U/Th~3). It may be of interest to consider extreme hypothetical cases involving failure of these assumptions. For example, consider the very improbable case in which a granite has lost its entire content of radon and its decay products. For an otherwise typical granite with a potassium content of 4. 00% and a 25% correction for uranium and thorium betas, the error caused by this disequilib­ rium would result in an apparent potassium content of 3.88% or a -3% error. In the second, highly improbable, hypothetical case assume that the above granite contains no thorium and that the entire alpha count is from the uranium series. In this case, the apparent potassium content would be 3. 96% or a -1. 6% e r r o r . Thus the maximum uncertainty introduced by CNa is about 3% for granite and about 1% for pure potassium bearing minerals such as biotite and perthite. It should be pointed out that the alpha activity determination is a useful geochemical parameter in its own right because both uranium and thorium tend to be concen­ trated in felsic magmas and residual fluids.

In conclusion, it may be stated that alpha-beta radiometric method for deter­ mining potassium has been found to be very reliable and apparently as precise as the methods used for comparison.

ACKNOWLEDGEMENT

This work was supported in part by United States Atomic Energy Commission contract AT(ll-l)-689. The authors are grateful to Professor John W. Anthony for reading and editing this report.

T A B L E I

Radiometric K Measurements for National Bureau of Standards Rock Standards

Sample Description Radiometric Determination Gravimetric Determination w ithout with (Nat'l. Bur. of A1 a b so rb er A1 a b s o r b e r Stds. ) % (K4-Rb) % (K +R b) %K

Chelmsford granite 4. 61 4 .4 7 4 .6 1

Pb-Ba glass 6 .9 0 6 .9 1

Milford granite 3. 39 3 .3 2

Gabbro-diorite 1. 53 1 .2 9

Columbia R. basalt 0. 88 0. 88 0. 82 1. 07

Triassic diabase 0. 46 0 .4 8 71

Comparison of K Analyses by Counting and Other Methods

Sample description Radiometric method Comparison method %K %K Black Hills feldspar 11.8 11.8 (L)

McKinney feldspar 11. 7 11. 0 (G)

Brown Derby feldspar 11.1 1 0 .4 (L)

Wind River Canyon feldspar 9.65 9 .3 9 (G)

M. I. T. Standard Biotite 7 .7 1 (7. 40 (f. p. s .) (7.80 (f.p.d.)

Lake Kivu leucitite 4.83 5.20 (i.d.)

Cat Mt. rhyolite 4 .2 4 4. 41 (f. p. s .)

Wilburforce antiperthite 3. 39 3 .7 3 (L)

A ta s c o s a M ts. b a sa lt 1. 04 0.98 (f.p. s.)

Beryl Mt., N.H. beryl 0 .3 5 (0 .3 9 (L) (0 .4 2 (G)

Hawaiian Olivine basalt 0 .2 4 0. 35 (i.d.)

Stillwater plagioclase 0.1 7 0.12 ( i .d .)

L- determined by Ledoux and Co. using a wet chemical method (Kallman, 1956).

G= Analyzed in the University of Minnesota Rock Analysis Laboratory by the L. Lawrence Smith fusion method (Goldich and Qsland, 1956).

i. d. : isotope dilution analysis by P. W. Cast (1957), Lament Geological Observatory, Columbia University.

f. p. s. - flame photometric analysis in this laboratory using internal lithium standard (Hedge, 1959).

f.p.d. - analyzed as above using double crucible technique.

REFERENCES

Baadsgard, H. , Goldich, S. S., Neir, A. O. and Hoffman, J. H. (1957) The reproducibility of A^O/K^® Age determinations: Trans. Am. Geophys. U nion, ^ 8, p. 5 3 9 -5 4 2 .

Damon, P. E. (1950) Radioactivity and Mineralization in rhyolite porphyry: Geophysics 15^ p. 94-101.

Damon, P. E. and Kulp, J. L. (1958) Excess helium and argon in beryl and other minerals: Am. Min. ^ ip . 433-459.

Dresia, H. and Beckmann, R. (1957) Radiological rapid determination of po­ tassium with Geiger-M uller tubes in solution: Z. Anal. Chem. 159. p. 1- 12. ------

Gaudin, A. M. and Pannell, J. H. (1948) Radioactive determination o£ potassium in solids: Anal. Chem. 20, p. 1154. TABLE 2

Comparison of K Analyses by Counting and Other Methods

Sample description Radiometric method Comparison method %K Black Hills feldspar 11. 8 11. 8 (L)

McKinney feldspar 11.7 11. 0 (G)

Brown Derby feldspar 11.1 1 0 .4 (L)

Wind River Canyon feldspar 9. 65 9 .3 9 (G)

M.I. T. Standard Biotite 7 .7 1 (7.40 (f.p. s.) (7.80 (f.p.d.)

Lake Kivu leucitite 4.83 5.20 (i.d.)

Cat Mt. rhyolite 4 .2 4 4. 41 (f. p. s .)

Wilburforce antiperthite 3. 39 3. 73 (L)

Atascosa Mts. basalt 1 .0 4 0.98 (f.p. s.)

Beryl M t., N. H. beryl 0 .3 5 (0 .3 9 (L) (0 .4 2 (G)

Hawaiian Olivine basalt 0. 24 0. 35 (i.d.)

Stillwater plagioclase 0 .17 0.12 ( i .d .)

L= determined by Ledoux and Co. using a wet chemical method (Kallman, 1956).

G= Analyzed in the University of Minnesota Rock Analysis Laboratory by the L, Lawrence Smith fusion method (Goldich and Qsland, 1956).

i.d.= isotope dilution analysis by P. W. Cast (1957), Lament Geological Observatory, Columbia University.

f. p. s. - flame photometric analysis in this laboratory using internal lithium standard (Hedge, 1959).

f. p.d.: analyzed as above using double crucible technique.

REFERENCES

Baadsgard, H. , Goldich, S. S. , Neir, A. O. and Hoffman, J. H. (1957) The reproducibility of A40/K40 Age determinations: Trans. Am. Geophys. Union, 38. P- 539-542.

Damon, P. E. (1950) Radioactivity and Mineralization in rhyolite porphyry: Geophysics 15^ p. 94-101.

Damon, P. E. and Kulp, J. L. (1958) Excess helium and argon in beryl and other minerals: Am. Min. 4^ p. 433-459.

Dresia, H. and Beckmann, R. (1957) Radiological rapid determination of po­ tassium with Geiger-M uller tubes in solution: Z. Anal. Chem. 159, p. 1-12. ------

Gaudin, A. M. and Pannell, J. H. (1948) Radioactive determination of potassium in solids: Anal. Chem. 20, p. 1154. APPENDIX C

NEUTRON ACTIVATION STUDIES OF BASALTS

In an attempt to correlate various basalt flows and associated volcanic rocks, samples were neutron activated and their spectra were studed with a 200 channel gamma ray spectrometer. The purpose of studying the spectra was two-fold; first an effort was made to determine which elements were present and , secondly an attempt to finger print the samples by using the spectra as a correlation tool was made.

The samples were placed in the reactor for approximately three hours and then allowed to decay until they were biologically safe. The reactor in use had a neutron flux of 1 0 ^ neutrons per square centimeter per second. The cooling period before the first spectra were read was approximately 110 hours.

The technique used in studying the decay pattern was similar to that used in any nuclear laboratory with the exception of the 200 channel gamma ray spectrometer which recorded and graphed the spectrum of the elements present. The spectrometer was calibrated with standards consisting of Co®®, C o ^ , Cs^^,

Mn"^ and Na^. In addition, samples of pitchblende, monazite and

73 74 an artificial Cr*^ source were used to give additional calibration points. The spectrometer is adjustable so that all 200 channels may plot energies over a small range or conversely the channels may be spread out over a large range.

The scintillation counter used was a thallium activated sodium iodide well type crystal. The samples were contained in polyethylene vials which were sealed and cleaned before activation to keep the danger of contamination at a minimum. The background of the polyethylene vials was checked by running blank vials through the same process as the rock samples. For all practical purposes

the vials contributed nothing to the gamma ray spectrum.

Fifty milligram samples of pulverized rocks were placed in

the activation vials. The weight was chosen so the cooling off period would not be too long, thus allowing some of the shorter

lived isotopes to be studied. The fifty milligram samples could

also be weighed out quite accurately and conveniently. One dis­

advantage of this size sample is the problem of obtaining a repre­

sentative sample of the rock. This was partially eliminated by

pulverizing the rocks to a fine mesh size and quartering the

samples.

The activated samples were placed in the multi-channel

analyzer and their gamma ray spectra were studied for a preset

time of ten minutes. The ten minute time was chosen in order to

get good reproducible peaks in a minimum amount of time. Various 75 other times ranging from one to twenty minutes were tried and it was found that the lower time limits gave peaks which were difficult to distinguish and statistically inaccurate. At the higher time limits it was found that it was quite easy to separate true peaks via statistics. The ten minute time was chosen in order to obtain

the maximum amount of information in a minimum amount of time and

still maintain statistical validity.

Duplicate samples of the same rocks were analyzed as well

as making duplicate runs on the same vial, in order to test the

reproducibility of the machine and to check for any sources of

contamination which may have been introduced. The reporducibility

of the machine was found to be excellent and no contamination was

found.

The main elements which one might expect to find in basic

rocks are shown in Table IX. This table lists the stable and

longer lived isotopes of the elements which are most abundant in

basic rocks. The half-life, gamma ray energy and cross section

for neutron activation are shown in this table. Table X shows

the abundance of the elements in various basic rocks according to

different investigators.

In order to determine which of the elements present in a

fifty milligram sample were capable of producing measurable radio­

active isotopes after the decay period, the sensitivity limits of

the various elements present were calculated from the following 76

Table IX-Artificial Produced Long Lived and the Common Stable Isotopes of the Elements Commonly Found in Basalts

Isotope Half-life % Abundance Energy of Gamma Ray Cross-Section % b a m s

Na-22 2.6 yrs. 1.28 MEV

Na-23 Stable 100 0.53

Na-24 15 hrs. 1.38, 2.75 MEV

Mg-24 Stable 78.60 0.03

Mg-25 Stable 10.11 0.27

Mg-26 Stable 11.29 0.03

Mg-28 21.5 hrs. .40, .95, 1.35 MEV

Al-26 10® yrs. 1.82, .72 MEV

Al-27 Stable 100 0.23

Si-28 Stable 92.18 0.10

Si-29 Stable 4.71 0.30

Si-30 Stable 3.12 0.11

Si-32 710 yrs. No gamma

P-31 Stable 100 0.20

P-32 14.3 d No gamma

S-32 Stable 95.018 ?

S-33 Stable 0.75 0.02

S-34 Stable 4.215 0.26

S-35 85 d No gamma

S-36 Stable 0.017 0.14

K-39 Stable 93.08 1.90 77

Isotope Half-life % Abundance Energy of Gamma Ray Cross-section % b a m s K-40 1.25 b.y. 0.0119 1.46 MEV 70

K-41 Stable 6.91 • 1.10

K-42 12.5 hrs. 1.53, .32 MEV

Ca-40 Stable 96.97 0.2

Ca-41 10^ yrs. No gamma

Ca-42 Stable 0.64 40

Ca-43 Stable 0.145 ?

Ca-44 Stable 2.06 0.6

Ca-45 146 d No gamma

Ca-46 Stable 0.0033 0.3

Ca-47 5 d 1.31, .83, .48 MEV

Ca-48 Stable 0.185 1.1

Sc-45 Stable 100 23

Sc-46 84 d .88, 1.12 MEV

Ti-46 Stable 7.99 0.6

Ti-47 Stable 7.32 1.5

Ti-48 Stable 73.99 7.8

Ti-49 Stable 5.46 1.8

Ti-50 Stable 5.25 0.14

V-50 Stable 0.25 100

V-51 Stable 99.75 4.5

Cr-50 Stable 4.31 16

Cr-51 27.8 d .32, .65 MEV

Cr-52 Stable 83.76 0.8 78

Isotope Half-life % Abundance Energy of Gamma Ray Gross-section - 7. barns

Cr-53 Stable 9.55 18

Cr-54 Stable 2.38 0.37

Hn-54 300 d .84 MEV

Mn-55 Stable 100 13.3

Mn-56 2.6 hrs. .85, 1.81, 2.13 MEV

Fe-54 Stable 5.84 2.2

Fe-55 2.6 yrs. No gamma

Fe-56 Stable 91.68 2.6

Fe-57 Stable 2.17 0.9

Fe-58 Stable 0.31 2.4

Fe-59 45 d .19, 1.1, 1.28 MEV

Co-59 Stable 100 37

Co-60 5.2 yrs. 1.17, 1.33 MEV

Ni-58 Stable 67.76 4.3

Ni-59 8000 yrs. No gamma

Nt-60 Stable 26.16 2.6

Ni-61 Stable 1.35 2

Nl-62 Stable 3.66 15

Ni-63 80 yrs. No gamma

Ni-64 Stable 1.16 2

Ni-65 2.6 hrs. 1.48, 1.12, .37 MEV

Ni-66 55 hrs. No gamma

Cu-63 Stable 69.1 4.4 79

Isotope Half-life % Abundance Energy of ray Cross-section % barns

Cu-64 12.8 hrs. 1.35 MEV

...... Cu-65 Stable 30.9 2.2

Zn-64 Stable 48.89 0.5

Zn-65 245 d 1.12 MEV

Zn-66 Stable 27.81 ?

Zn-67 Stable 4.11 ?

Zn-68 Stable 18.56 1.9

Zn-69 14 hrs. 0.44 MEV

..... Zn-70 Stable 0.62 0.09

Zn-72 49 hrs. Not known

Ga-69 Stable 60.2 1.9

Ga-71 Stable 39.8 4.6

Ga-72 14.3 hrs. Several

Ge-68 280 d No gamma

Ge-70 Stable 20.55 3.4

Ge-71 11 d No gamma

Ge-72 Stable 27.37 1

Ge-73 Stable 7.67 14

Ge-74 Stable 36.74 0.7

Ge-76 Stable 7.67 0.32

As-71 60 hrs. 0.17 MEV

As-75 Stable 100 4.3

As-76 26 hrs. Several 80

Isotope Half-life % Abundance Energy of gamma ray Cro s s-section . barns - 7

Rb-85 Stable 72.15 0.85

Rb-86 18.6 days 1.08 MEV

Rb-87 4.7 b y 27.85 No gamma 0.14

Sr-84 Stable 0.55 1

Sr-85 65 d 0.51 MEV

Sr-86 Stable 9.86 ?

Sr-87 Stable 7.02 ?

Sr-88 Stable 82.56 0.005

Sr-89 50 d No gamma

Zr-90 Stable 51.46 0.1

Zr-91 Stable 11.23 1

Zr-92 Stable 17.11 0.2

Zr-94 Stable 17.40 0.1

Zr-95 65 d 0.72 MEV

Zr-96 Stable 2.80 0.1

Ba-130 Stable 0.101 6

Ba-131 12 d Several

Ba-132 Stable 0.097 3

Ba-133 7.2 yrs. 0.35, 0.276 MEV

Ba-134 Stable 2.42 4.4

Ba-135 Stable 6.59 5

Ba-136 Stable 7.81 ?

Ba-137 Stable 11.32 4

Ba-138 Stable 71.66 0.55 81

Table X-Average Compositions of Various Type Basalts

(a) (b) (c) (d) (e) (£) (s) Plateau Alkali Normal Average Basic Leucite Element Basalt Basalt Tholeiite Basalt Oceanite Rocks Basalt

sio2 49.7 46.1 51.3 49.06 45.6 48.8 46.04

Ti°2 2.2 2.6 2.0 1.36 1.7 1.53 0.64

AI2O3 14.2 14.8 14.2 15.70 8.3 16.8 12.23

3.86 Fe2°3 3.7 3.2 2.9 5.38 2.3 12.25

FeO 10.0 8.8 9.1 6.68 10.2 4.66

MnO 0.2 0.2 0.2 —— 0.1 .29 Trace

NgO 6.8 9.4 6.4 6.17 21.7 7.47 10.38

CaO 9.6 10.8 10.5 8.95 7.5 9.42 8.97

Na20 2.6 2.7 2.3 3.11 1.3 2.62 2.42

k 2o 0.7 1.0 0.9 1.52 0.4 0.99 5.77

0.3 p 2°5 0.4 0.2 .45 0.3 0.32 1.14

Sc ------2.4x 10-3 ---

V — ------2xl0-2 ---

Cr ------—- — 3xl0“2

Co ------—- — —- 4 .5xl0"3 - ——

Ni ——— " — — — — 1 .6xl0"2

Cu ------1.4xl0"2 ---

Zn ------1.3xl0”2 ---

Rb ------4 .5x 10-3 ---

Sr — ------2.7x 10-3 .20

Zr — — ■ mm mm mm IxlO"2 _ __ 82 (a) (b) (c) (d) (e) (f) (g) Plateau Alkali Normal Average Basic Leucite Element Basalt Basalt Tholeiite Basalt Oceanite Rocks Basalt

Ba — ~ — --- 2.7x10-2 .42

Ga ------“ “ “ •“ 1.8x10-3 — - —

Ge --- —- — ------1.5xl0"4 ---

As 2xl0-4

(a) Daly, 1933, p. 17; (b) Nockolds, 1954; p. 1021; (c) Nockolds, 1954, p. 1021; (d) Daly, 1933, p. 17; (e) Tyrrell, 1926, p. 131; (£) Vinogradov,.1956, Geokhimiya; (g) Clark, 1911, p. 436. . . 83 equation: A^atf'N (1-e " ^ ^ r)(e ^ ) • A equals the activity of the sample after the decay period, Tir equals the irradiation time, Td equals the decay time before the measurement of the activity, a equals the fractional isotopic abundance of the reacting isotopes,

N equals the number of target atoms of the parent element in the sample, X equals the decay constant, ^ equals the neutron flux in neutrons per square centimeter per second, and <$"* equals the cross section of the target isotope to produce the radioactive nuclide via the neutron reaction.

The sensitivity limits of the various elements were cal­

culated using a three-hour irradiation time, 110 -hour decay time

and a neutron flux of 10^ neutrons per square centimeter per

second. Five disentigrations per second was chosen as the measur­

able activity though as the shorter lived isotopes decayed away

this value would approach one for the longer lived isotopes.

Table XI shows the resulting minimum concentrations of the

elements present in basalts to give a measurable spectrum for a

three-hour activation time and a 100 -hour decay time. It can be

seen from this table that K, Na, Sc, Cr, Fe, Cu and possibly Ba

are of a great enough abundance in a fifty milligram sample to give

a measurable spectrum under these conditions.

The main reactions being studied in this investigation

were neutron-gamma reactions in which an unstable isotope contain­

ing one more neutron is formed. These isotopes usually decay by 84

Table XI-Sensitivity Limits for Elements Commonly Found in Basalts

Element g per 3-hr. irradiation g in 50 mg basalt

N a ^ — Na24 4.5 x 10-5 1 x 10-3

K61 - K42 1.23 x 10-6 5 x lO"* ca46 - Ca47 5.3 x 10”2 3.55 x 10"3

So45 - S=46 5.05 x 10”6 1.20 x 10"5

Or50 - c,51 2.34 x 10"6 1.5 x 10"5

Mn55 — Mn56 Half-life too short 1.1 x 10-4

Fa58 - Fa59 3.48 x 10”4 4.28 x 10"3

Co59 — Co60 1.80 x 10“5 2.25 x 10"6

Cu63 - Ou64 4.45 x 10“7 7 x 10-6

z„64 - Zn65 5.1 x 10“5 6.5 x lO-6

zn68 - Zn69 2.5 x 10“5 6.5 x lO"6

Ga71 - Ga72 8 x 10“6 9 x 10-7

As75 - Aa76 4 x 10"7 1 x 10"7 % in

1 1 Rb86 1.76 x 10-5 2.25 x 10"6

Zr9^ — Zr95 3 x 10”2 5 x 10"6

Sr84 - Sr85 1.8 x lO"1 2.2 x 10"5

Ba130 -- Ba131 3.12 x lO’5 1.35 x 10"5 85 beta minus and gamma to a stable isotope of the next higher atomic number. A search for radioactive isotopes resulting from other nuclear reactions was conducted. In this investigation weights of elements corresponding to those found in a fifty milligram sample of a basic rock were activated and their spectra were studied.

The resulting spectra showed no evidence of other nuclear reactions. BIBLIOGRAPHY

Bryner, Leonid, 1959, Geology of the South Comobabi Mountains and Ko Vaya Hills, Pima County, Arizona, University of Arizona unpublished doctor’s dissertation, 156 pages.

Chew, Randall T., 1952, Geology of the Mineta Ridge Area, Pima and Cochise Counties, Arizona, University of Arizona unpublished master's thesis, 53 pages.

Colby, Robert E., 1958, Stratigraphy and Structure of the Recreation Red Beds, Tucson Mountain Park, Arizona, University of Arizona unpublished master's thesis, 64 pages.

Cooper, John R., 1960, Some Geologic Features of the Pima Mining District, Pima County, Arizona, Geological Survey Bulletin 1112-C, 103 pages.

Daly, R. A., 1933, Igneous Rocks and the Depths of the Earth, McGraw-Hill.Book Co.

Damon, Paul E., Hedge, C. E., Taylor, 0. J., Halva, C. J., 1960, Radiometric Determination of Potassium in Silicates, Arizona Geological Society Digest, Vol. Ill, pp. 75-80.

Donald, Peter G., 1959, Geology of the Fresnal Peak Area, Baboquivari Mountains, Arizona, University of Arizona unpublished master's thesis, 45 pages.

Faust, G. T., and Callaghan, Eugene, 1948, Mineralogy and Petrology of the Currant Creek Magnesite Area, Bulletin of the Geological Society of America, vol. 59, number 1, January 1948.

Goldschmidt, V. M . , 1954, Geochemistry. Clarendon Press, Oxford, England.

Halva, Carroll J., 1960, A Semi-Micro Analysis of Silicate Rocks for Ca, Mg, Fe, and A1 Employing E.D.T.A., Arizona Geo­ logical Society Digest, Vol. Ill, pp. 81-86,

86 87

Harshman, Elbert N., 1939, Geology of the Belmont-Queen Creek Area, Superior, Arizona, University of Arizona unpublished doctor's dissertation, 167 pages.

Higdon, Charles E., 1935, Geology and Ore Deposits of the Sunshine Area, Pima County, Arizona, University of Arizona unpublished master's thesis, 25 pages.

Hillebrand, James R., 1953, Geology and Ore Deposits in the Vicinity of Putnam Wash, Pinal County, Arizona, University of Arizona unpublished master's thesis, 94 pages.

Johnson, Vard H., 1941, The Geology of the Helvetia Mining District, Arizona, University of Arizona unpublished doctor's, dissertation, 111 pages.

Jones, William R., 1941, The Geology of the Sycamore Ridge Area, Pima County, Arizona, University of Arizona unpublished master's thesis, 89 pages.

Lovering, Tom G., 1953, The Geology of a Western Portion of the Santa Rita Quadrangle, Grant County, New Mexico, University of Arizona unpublished master's thesis, 44 pages.

Popoff, Constantine, 1940, The Geology of the Rosemont Mining Camp. Pima County, Arizona, University of Arizona unpublished master's thesis, 100 pages.

Price, William E., Jr., 1948, Rim Rocks of Sycamore Canyon, Arizona, University of Arizona unpublished master's thesis, 93 pages.

Rasor, Charles A . , 1937, Mineralogy and Petrography of the Tombstone Mining District, Arizona, University of Arizona unpublished doctor's dissertation, 115 pages.

Sabels, Bruno E., 1960, Late Genozoic Volcanism in the San Francisco Volcanic Field and Adjacent Areas in North-Central Arizona, University of Arizona unpublished doctor's dissertation, 135 pages.

Smith, Lewis A . , 1927, The Geology of the Commonwealth Mine, .University of Arizona unpublished master's thesis, 73 pages.

Sopp, George P ., 1940, Geology of the Montana Mine Area, Empire Mountains, Arizona, University of Arizona unpublished master's thesis, 63 pages. 88

St. Clair, Charles S., 1957, Geologic Reconnaissance of the Aqua Fria River Area, Central Arizona, University of Arizona unpublished master's thesis, 76 pages.

Taylor, Omar J., 1960, Correlation of Volcanic Rocks in Santa Cruz County, Arizona, University of Arizona unpublished master's thesis, 59 pages.

Tolman, C., 1909, Geology of Tumamoc Hill, Carnegie Institute of Washington Publication no. 113, p. 67-82.

Tyrrell, G. W.,.1926, The Principles of Petrology, E. P. Dutton & Co., Inc., New York, New York, chapter VI, page.101-131.

Voelger, Klaus, 1953, Cenozoic Deposits in the Southern Foothills of the Santa Catalina Mountains near Tucson, Arizona, University of Arizona unpublished master's thesis, 101 pages.

Wargo, Joseph G., 1959, The Geology of the Schoolhouse Mountain .Quadrangle, Grant County, New Mexico, University of Arizona unpublished doctor's dissertation, 187 pages.