OR/14/052 Appendix 3: Geochemistry of the brown and grey till deposits of Anglesey, Wales – preliminary results

Boon, D, Kirkham, M, and Scheib, A. 2014. Physical properties of till deposits from Anglesey, north west Wales. British Geological Survey Internal Report, OR/14/052.

Background

Soil geochemical baseline data have successfully been applied as proxy in the reconstruction of the flowpaths of the Middle Pleistocene British Ice Sheet in central-eastern England. In this study Scheib et al. (2011) used total element concentrations from XRFS analysis to firstly establish element associations in soils over known till deposits and secondly provenance these geochemical signatures, enabling the reconstruction of ice flow paths associated with two different Middle Pleistocene till sheets.

Analysis

Samples from the Grey Till and Brown Till units from Cemlyn Bay, North Anglesey, were analysed using a hand-held (HH) portable X-ray fluorescence spectrometry (XRFS) element analyser; a method that is an inexpensive, quick and easy way to obtain semi-quantitative results.

The analyses with the HH XRFS were carried out on two samples from each till deposit; one sample containing material below 63 μm (clay and silt fraction) and the other containing material above 63 μm (sand fraction to 2 mm). All samples were air dried and retained in a small re- sealable plastic bag.

These sample bags were laid out on a clean work bench and measured using the NITON XLt Analyser in test mode ‘Standard Bulk Mode’. Latter mode gives result in mg/kg for elements K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Rb, Sr, Ag, Cd, Sn, Sb, Hg and Pb. Individual samples were measured three times at 30 seconds each.

To assess the precision and accuracy of the HH XRFS results, the G-BASE internal secondary reference material (SRM) S22B was measured at the start and end of analysis. Tables 1 and 2 list results of those measurements. Results for some elements (Ni, Co, Cr, Hg, Cd, and Se) were excluded because concentrations were either below detectable limits or inconsistent with the lab-based XRFS results.

Tables 1 and 2 also display a column for each element that gives two standard deviations (2SD) for each of the 30 second measurement; and informs on the distribution of the concentrations. Results for the six measurements of S22B are consistent throughout, except the third measurement which seems to give lower concentrations for all determinants. Results for V should also be handled with care as measurements are not as consistent.

The average (mean) concentration of the six individual measurements (total HH) is listed with the lab-based XRFS results for SRM S22B below. The error (difference), displayed in %, between those two results is calculated as follows:

error = [(Lab–HH mean)/Lab]*100

Except for results for Ca, K, V and Sr, HH results are within 20% of the lab-based XRFS results. Positive percentages throughout the analysis indicate an underreporting by the HH XRFS method. A modification factor (Tables 1 and 2) was calculated for each element, and applied to the results obtained for the four till samples.

Table 8    Concentrations (mg/kg) for SRM S22B following analysis by handheld XRFS NITON XLt Analyser;
2SD = two standard deviations for each 30 second measurement

Sample	No	As	As 2SD	Pb	Pb 2SD	Zn	Zn 2SD	Cu	Cu 2SD	V	V 2SD	Sr	Sr 2SD	Rb	Rb 2SD
S22B	1	2864	79.78	428	40.63	1134	81.93	253	66.42	142.0	145.48	27.8	6.35	171	12.59
S22B	2	2973	82.17	422	41.13	1136	83.23	251	68.08	27.3	155.19	32.8	6.73	164	12.53
S22B	3	1870	54.00	253	27.09	700	55.13	183	46.86	58.9	123.68	21.5	4.78	120	8.93
Mean HH		2569	71.98	368	36.28	990	73.43	229	60.45	76.1	141.45	27.3	5.95	152	11.35
S22B	4	2774	76.28	378	37.85	1062	77.53	206	61.59	159.7	160.79	29.3	6.23	174	12.36
S22B	5	2821	78.12	387	38.79	1084	79.63	293	67.62	85.9	146.12	24.0	6.09	160	12.09
S22B	6	2503	69.63	336	34.41	981	71.62	174	55.79	156.7	144.65	27.8	5.84	151	11.13
Mean HH		2699	74.68	367	37.02	1042	76.26	224	61.67	134.1	150.52	27.0	6.05	162	11.86
Total HH		2634	73.33	367	36.65	1016	74.85	227	61.06	105.1	145.99	27.2	6.00	157	11.61
Lab		3405		440		1098		271		148.4		47.4		196
Difference %		22.6		16.5		7.5		16.3		29.2		42.7		20.0
Mod Factor		1.29		1.20		1.08		1.19		1.41		1.74		1.25

Table 9    Concentrations (mg/kg) for SRM S22B following analysis by handheld XRFS NITON XLt Analyser;
2SD = two standard deviations for each 30 second measurement, continued

Sample	Fe	Fe 2SD	Mn	Mn 2SD	Ti	Ti 2SD	Ca	Ca 2SD	K	K 2SD
S22B	68309	1087	7486	482	6332	722	5767	1009	14290	1647
S22B	68663	1105	7663	493	5371	768	5771	971	13636	1582
S22B	38507	687	3556	286	3077	593	2547	771	4972	1124
Mean HH	58493	959	6235	420	4926	694	4695	917	10966	1451
S22B	64557	1031	6719	448	6062	782	6282	1018	14459	1648
S22B	65774	1057	6926	461	6535	737	4733	967	14743	1650
S22B	54375	910	5472	390	4822	695	3842	891	9243	1396
Mean HH	61569	999	6372	433	5806	738	4953	959	12815	1565
Total HH	60031	979	6304	427	5366	716	4824	938	11890	1508
Lab	71049		7498		6534		7505		20581
Difference %	15.5		15.9		17.9		35.7		42.2
Mod Factor	1.18		1.19		1.22		1.56		1.73

Results

Following on from the analysis of SRM S22B, tables 8 and 9 list the results for elements As, Pb, Zn, Cu, Rb, Sr, V and majors Ca, K, Fe, Mn and Ti (as percentage), respectively. The concentrations are the average (mean) of the three measurements and have each been multiplied by the modification factor (Tables 8 and 9); hence the Mod prefix in the header of both tables. Additional to the results for both the fine and sand fraction, tables also list the concentration of the bulk sample; simply the sum of former two results.

Table 10    Modified concentrations (mg/kg) of As, Pb, Zn, Cu, Rb, Sr and V measured in the fine and coarse fraction of the Grey Till and Brown Till samples; concentration difference between the fine and coarse fraction in %

Sample	Fraction	Mod As	Mod Pb	Mod Zn	Mod Cu	Mod Rb	Mod Sr	Mod V
Grey Till fine	<63 μm	19.57	25.79	148.60	131.26	110.90	272.36	140.05
Grey Till coarse	>63 μm	8.58	6.65	52.34	21.33	57.96	182.10	104.26
difference	%	56.1	74.2	64.8	83.7	47.7	33.1	25.6
Grey Till bulk	<2 mm	28.15	32.44	200.94	152.59	168.86	454.46	244.31
Brown Till fine	<63 μm	23.92	34.76	100.42	96.20	116.73	136.62	273.46
Brown Till coarse	>63 μm	6.31	20.12	45.04	22.12	69.71	77.32	78.45
difference	%	73.6	42.1	55.2	77.0	40.3	43.4	71.3
Brown Till bulk	<2 mm	30.23	54.88	145.45	118.32	186.44	213.94	351.92
Brown Till	cobble	9.07	15.15	66.72	-18.20	11.40	2213.05	57.26

Table 11    Modified concentrations (%) of Ca, K, Fe, Mn and Ti measured in the fine and coarse fraction of the Grey Till and Brown Till samples; concentration difference between the fine and coarse fraction in %

Sample	Fraction	Mod Ca	Mod K	Mod Fe	Mod Mn	Mod Ti
Grey Till fine	<63 μm	6.24	2.68	4.44	0.16	0.59
Grey Till coarse	>63 μm	4.85	1.27	2.13	0.12	0.31
difference	%	22.3	52.5	52.0	22.8	46.8
Grey Till bulk	<2 mm	11.08	3.95	6.58	0.28	0.91
Brown Till fine	<63 μm	0.42	2.43	5.04	0.17	0.77
Brown Till coarse	>63 μm	0.19	1.52	2.65	0.15	0.34
difference	%	53.6	37.3	47.4	14.3	56.3
Brown Till bulk	<2 mm	0.61	3.95	7.69	0.32	1.11
Brown Till	cobble	31.61	0.65	0.89	0.03	0.07

Figure 23    Bar chart of concentrations (mg/kg) of As, Pb, Zn, Cu, Rb, Sr and V measured in the fine and coarse fraction of the Grey Till and Brown Till samples.

Figure 24    Bar chart of concentrations (%) of Ca, K, Fe, Mn and Ti measured in the fine and coarse fraction of the Grey Till and Brown Till samples.

Figures 23 and 24 display the concentrations of the coarse and fine fractions, for both the Grey and Brown Till as a stacked column chart. Throughout both charts, concentrations measured in the fine fraction of both till types are higher than in the sand fraction, which largely comprises quartz grains. The difference between concentrations between the fine and coarse fraction, expressed as %, are listed in Tables 8 and 9.

For the major elements, lowest concentration differences are calculated for Mn of 22.8 and 14.3% for Brown Till and Grey Till respectively (Table 24). Highest differences occur for Fe and Ti (approximately 50%) For the trace elements and base metals, concentration differences are much higher and range from 40 to 84%, with differences below 33% only calculated for Sr and V in Grey Till samples.

The most significant differences in concentration levels between the two till types can be seen for Ca and Sr (Figure 23 and 24). Whilst Sr concentrations in the Grey Till are approximately twice as high, Ca concentrations are 15 times higher in the fine fraction and 25 times higher in the coarse fraction; for the bulk sample, Ca concentrations in the Grey Till are 19 times higher. The sample from the Brown Till was obtained from a near surface, and decalcification may have occurred. Other elements that are slightly higher in the Grey Till are Cu and Zn. Concentrations of the other major elements are fairly similar across both till types.

For the Brown Till samples, results for V stand out. In particular, concentrations for the fine fraction are almost twice as high. Other elements that are slightly higher in the Brown Till are Rb and Pb.

Conclusions

• The handheld XRFS is very easy and quick to use. Results for SRM S22B showed that this method can provide consistent and statistically sound data for, in this case, 12 elements.
• Results showed that Sr and Ca are significantly higher in Grey Till samples, suggesting a calcareous signature. Most of the natural Ca relates to minerals, such as calcite and gypsum, and are subsequently particularly enriched in carbonate rocks, such as limestone, dolomite and chalk. Strontium is also often found in host minerals such as gypsum, calcite and dolomite.
• Ca and Sr levels measured in Grey Till samples are very high (bulk = 11.1% Ca and 455 mg/kg Sr) and are comparable to Ca and Sr levels measured in stream sediments from areas over Cretaceous Chalk or Jurassic Limestone of central and eastern England.
• Potential source rocks for the Grey Till are Carboniferous limestone and dolomite of the Red Wharf Bay area or/and calcareous Triassic strata of the Liverpool Bay (offshore), particularly Mercia Mudstone.
• The Brown Till is completely lacking in a calcareous signature and suggests that the Grey Till has derived (in parts) from different source material, though the sample may be decalcified.
• Results show that V is nearly twice as high in the Brown Till compared with the Grey Till samples. Vanadium is mainly associated with and enriched in basalt or gabbro with host mineral such as pyroxenes and amphiboles. The higher concentrations in Brown Till samples could maybe relate to a mafic igneous signature?

Recommendations

Geochemical data from the analysis using the handheld XRFS could be a useful additional method to help discriminate between different till deposits. Obtaining these semi-quantative results is very quick and cheap — compared with Lab XRF, and can be performed in the field.

The separate analysis of the sand and clay-silt fractions showed the same consistent trend of either elevated or low concentrations in both fractions. It is therefore appropriate and sufficient to use the dried <2 mm sample fraction. This would on the one hand reduce cost and time for sample preparation, and residual PSD material could be used, but is also in line with the size fraction used by the G-BASE project for soil samples.

Handheld XRFS should certainly be considered in future studies of tills or any other superficial deposit that may vary in its composition.

To investigate the tills of Anglesey, more samples need to be collected, prepared and analysed to firm up some of the above results and derived conclusions.